WO2024143573A1 - Apparatus and method for supporting parallel neural network architecture for cumulative csi feedback in wireless communication system - Google Patents

Abstract

According to various embodiments of the present disclosure, provided is a method for operating a user equipment (UE) in a wireless communication system, the method comprising the steps of: receiving one or more synchronization signals from a base station (BS); receiving system information from the base station; receiving a reference signal from the base station; generating channel state information (CSI) on the basis of the reference signal; generating multiple signals for the CSI, on the basis of an encoder neural network (NN) model including multiple encoder components arranged in parallel; and transmitting the multiple signals to the base station.

Description

Apparatus and method for supporting parallel neural network architecture for cumulative CSI feedback in a wireless communication system

This disclosure relates to wireless communication systems. Specifically, the present disclosure relates to an apparatus and method for supporting a parallel neural network architecture for cumulative CSI feedback in a wireless communication system (AI for communication, AI4C) incorporating artificial intelligence (AI).

Wireless communication systems are being widely deployed to provide various types of communication services such as voice and data. In addition, attempts to incorporate artificial intelligence (AI) into communication systems are rapidly increasing. Methods being attempted can be largely divided into AI for communications (AI4C), which utilizes AI to improve communication performance, and communications for AI (C4AI), which develops communication technology to support AI. In the AI4C area, there are attempts to design the role of a channel encoder/decoder, modulator/demodulator, or channel equalizer by replacing it with an end-to-end autoencoder or neural network. In the C4AI area, there is a method of updating a common prediction model while protecting personal information by sharing only the weight or gradient of the model with the server without sharing device raw data through federated learning, a technique of distributed learning.

In order to solve the above-described problems, the present disclosure provides a method and apparatus for supporting a parallel neural network architecture for cumulative CSI feedback in a wireless communication system (AI for communication, AI4C) incorporating artificial intelligence (AI).

The present disclosure provides an apparatus and method for performing adaptive feedback rate operation according to downlink channel environment and uplink resource conditions.

The present disclosure provides an apparatus and method for saving required storage space and computing resources by reducing the complexity of a neural network (NN) model (parameter set).

This disclosure provides an apparatus and method for securing flexibility for the NN model in training and inference operations.

The present disclosure provides an apparatus and method for saving learning time of a neural network (NN).

The technical problems to be achieved by this disclosure are not limited to the technical problems mentioned above, and other technical problems not mentioned can be clearly understood by those skilled in the art from the description below. You will be able to.

According to various embodiments of the present disclosure, in a method of operating a user equipment (UE) in a wireless communication system, the steps include receiving one or more synchronization signals from a base station (BS); Receiving system information, receiving a reference signal from the base station, generating channel state information (CSI) based on the reference signal, an encoder including a plurality of encoder components arranged in parallel A method is provided including generating a plurality of signals for the CSI based on a neural network (NN) model and transmitting the plurality of signals to the base station.

According to various embodiments of the present disclosure, a method of operating a base station (BS) in a wireless communication system includes transmitting one or more synchronization signals to a user equipment (UE), the terminal Transmitting system information, transmitting a reference signal to the terminal, and receiving a plurality of signals from the terminal, wherein the plurality of signals are channel state information (CSI) based on the reference signal. ), and a method is provided in which the plurality of signals are generated based on an encoder NN (neural network) model including a plurality of encoder components arranged in parallel.

According to various embodiments of the present disclosure, a terminal (user equipment, UE) in a wireless communication system includes a transceiver and at least one processor, and the at least one processor receives one or more signals from a base station (BS). Receive a synchronization signal, receive system information from the base station, receive a reference signal from the base station, and generate channel state information (CSI) based on the reference signal, in parallel. A terminal configured to generate a plurality of signals for the CSI based on an encoder neural network (NN) model including a plurality of encoder components arranged and transmit the plurality of signals to the base station is provided.

According to various embodiments of the present disclosure, a base station (BS) in a wireless communication system includes a transceiver and at least one processor, and the at least one processor is a user equipment (UE). ), transmitting one or more synchronization signals to the terminal, transmitting system information to the terminal, transmitting a reference signal to the terminal, and receiving a plurality of signals from the terminal, and the plurality of The signals are related to channel state information (CSI) based on the reference signal, and the plurality of signals are generated based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel. A base station is provided. .

According to various embodiments of the present disclosure, in one or more non-transitory computer-readable media storing one or more instructions, the one or more instructions, based on execution by one or more processors, perform operations. Performing the operations, the steps include receiving one or more synchronization signals from a base station (BS), receiving system information from the base station, and receiving a reference signal from the base station. , generating channel state information (CSI) based on the reference signal, generating a plurality of signals for the CSI based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel. A computer-readable medium is provided, including transmitting the plurality of signals to the base station.

According to various embodiments of the present disclosure, in one or more non-transitory computer-readable media storing one or more instructions, the one or more instructions, based on execution by one or more processors, perform operations. The operations include transmitting one or more synchronization signals to a user equipment (UE), transmitting system information to the terminal, and transmitting a reference signal to the terminal. , comprising receiving a plurality of signals from the terminal, wherein the plurality of signals are related to channel state information (CSI) based on the reference signal, and the plurality of signals include a plurality of encoder components arranged in parallel. A computer-readable medium generated based on an encoder NN (neural network) model is provided.

In order to solve the above-mentioned problems, the present disclosure can provide a method and device for supporting a parallel neural network architecture for cumulative CSI feedback in a wireless communication system (AI for communication, AI4C) incorporating artificial intelligence (AI). there is.

The present disclosure can provide an apparatus and method for performing adaptive feedback rate operation according to downlink channel environment and uplink resource conditions.

The present disclosure can provide an apparatus and method for saving required storage space and computing resources by reducing the complexity of a neural network (NN) model (parameter set).

The present disclosure can provide an apparatus and method for securing flexibility for the NN model in training and inference operations.

The present disclosure can provide an apparatus and method for saving learning time of a neural network (NN).

The drawings attached below are intended to aid understanding of the present disclosure and may provide embodiments of the present disclosure along with a detailed description. However, the technical features of the present disclosure are not limited to specific drawings, and the features disclosed in each drawing may be combined to form a new embodiment. Reference numerals in each drawing may refer to structural elements.

Figure 1 is a diagram showing an example of physical channels and general signal transmission used in a 3GPP system.

Figure 2 is a diagram showing the system structure of a next-generation radio access network (NG-RAN).

Figure 3 is a diagram showing the functional division between NG-RAN and 5GC.

Figure 4 is a diagram showing an example of a 5G usage scenario.

Figure 5 is a diagram showing an example of a communication structure that can be provided in a 6G system.

Figure 6 is a diagram schematically showing an example of a perceptron structure.

Figure 7 is a diagram schematically showing an example of a multi-layer perceptron structure.

Figure 8 is a diagram schematically showing an example deep neural network.

Figure 9 is a diagram schematically showing an example of a convolutional neural network.

Figure 10 is a diagram schematically showing an example of a filter operation in a convolutional neural network.

Figure 11 is a diagram schematically showing an example of a neural network structure in which a cyclic loop exists.

Figure 12 is a diagram schematically showing an example of the operating structure of a recurrent neural network.

Figure 13 is a diagram showing an example of an electromagnetic spectrum.

Figure 14 is a diagram showing an example of THz communication application.

Figure 15 is a diagram showing an example of an electronic device-based THz wireless communication transceiver.

FIG. 16 is a diagram illustrating an example of a method for generating an optical device-based THz signal.

Figure 17 is a diagram showing an example of an optical device-based THz wireless communication transceiver.

Figure 18 is a diagram showing the structure of a photon source-based transmitter.

Figure 19 is a diagram showing the structure of an optical modulator.

FIG. 20 is a diagram illustrating an example of a CSI network structure in a system applicable to the present disclosure.

FIG. 21 is a diagram illustrating an example of a CSI matrix preprocessing process for a CSI network in a system applicable to the present disclosure.

FIG. 22 is a diagram illustrating an example of comparing various CSI network architectures in a system applicable to the present disclosure.

FIG. 23 is a diagram illustrating an example of source information for various CSI network architectures in a system applicable to the present disclosure.

FIG. 24 is a diagram illustrating an example of a building block of Residual Block: ResNet in a system applicable to the present disclosure.

FIG. 25 is a diagram illustrating an example of the effect of skip connection in residual block in a system applicable to the present disclosure.

FIG. 26 is a diagram illustrating an example of a CSI network architecture: ACRNet in a system applicable to the present disclosure.

FIG. 27 is a diagram illustrating an example of CSI feedback bit streams accumulable in decoder in a system applicable to the present disclosure.

FIG. 28 is a diagram illustrating an example of feature extraction before skip connection in encoder in a system applicable to the present disclosure.

FIG. 29 is a diagram illustrating an example of a proposed parallel encoder NN architecture in a system applicable to the present disclosure.

FIG. 30 is a diagram illustrating an example where the total number of bit streams is 2 in a system applicable to the present disclosure.

FIG. 31 is a diagram illustrating an example of a case where a convolution operation in CNN is applied in a system applicable to the present disclosure.

FIG. 32 is a diagram illustrating an example of a depth-wise convolution operation applied in a system applicable to the present disclosure.

FIG. 33 is a diagram illustrating an example of a group convolution operation applied in a system applicable to the present disclosure.

FIG. 34 is a diagram illustrating an example of a group convolution operation applied in a system applicable to the present disclosure.

FIG. 35A is a diagram illustrating an example of a neural network structure in a system applicable to the present disclosure.

Figure 35b is a diagram showing an example of a neural network structure in a system applicable to the present disclosure.

FIG. 35C is a diagram illustrating an example of a neural network structure in a system applicable to the present disclosure.

FIG. 35D is a diagram illustrating an example of a neural network structure in a system applicable to the present disclosure.

FIG. 36 is a diagram illustrating an example of various NN architectures for accumulable CSI feedback in a system applicable to the present disclosure.

FIG. 37 is a diagram illustrating an example of the performance and complexity of the proposed parallel architecture in a system applicable to the present disclosure.

FIG. 38 is a diagram illustrating an example of an operation process of a terminal (user equipment, UE) in a system applicable to the present disclosure.

FIG. 39 is a diagram illustrating an example of the operation process of a base station (BS) in a system applicable to the present disclosure.

40 illustrates a communication system 1 applied to various embodiments of the present disclosure.

Figure 41 illustrates a wireless device that can be applied to various embodiments of the present disclosure.

Figure 42 shows another example of a wireless device that can be applied to various embodiments of the present disclosure.

Figure 43 illustrates a signal processing circuit for a transmission signal.

Figure 44 shows another example of a wireless device applied to various embodiments of the present disclosure.

Figure 45 illustrates a portable device applied to various embodiments of the present disclosure.

46 illustrates a vehicle or autonomous vehicle applied to various embodiments of the present disclosure.

47 illustrates a vehicle applied to various embodiments of the present disclosure.

Figure 48 illustrates an XR device applied to various embodiments of the present disclosure.

Figure 49 illustrates a robot applied to various embodiments of the present disclosure.

Figure 50 illustrates an AI device applied to various embodiments of the present disclosure.

In various embodiments of the present disclosure, “A or B” may mean “only A,” “only B,” or “both A and B.” In other words, “A or B” may be interpreted as “A and/or B” in various embodiments of the present disclosure. For example, in various embodiments of the present disclosure, “A, B or C” can be replaced with “only A,” “only B,” “only C,” or “any of A, B, and C.” It can mean "any combination of A, B and C".

A slash (/) or a comma used in various embodiments of the present disclosure may mean “and/or.” For example, “A/B” can mean “A and/or B.” Accordingly, “A/B” can mean “only A,” “only B,” or “both A and B.” For example, “A, B, C” can mean “A, B, or C.”

In various embodiments of the present disclosure, “at least one of A and B” may mean “only A,” “only B,” or “both A and B.” Additionally, in various embodiments of the present disclosure, the expression “at least one of A or B” or “at least one of A and/or B” can be interpreted the same as “at least one of A and B.”

Additionally, in various embodiments of the present disclosure, “at least one of A, B and C” may be referred to as “only A,” “only B,” “only C,” or “A.” , any combination of A, B and C.” Also, “at least one of A, B or C” or “at least one of A, B and/or C” means It may mean “at least one of A, B and C.”

Additionally, parentheses used in various embodiments of the present disclosure may mean “for example.” Specifically, when “control information (PDCCH)” is indicated, “PDCCH” may be proposed as an example of “control information.” In other words, “control information” in various embodiments of the present disclosure is not limited to “PDCCH,” and “PDDCH” may be proposed as an example of “control information.” Additionally, even when “control information (i.e., PDCCH)” is indicated, “PDCCH” may be proposed as an example of “control information.”

In various embodiments of the present disclosure, technical features individually described in one drawing may be implemented individually or simultaneously.

The following technologies can be used in various wireless access systems such as CDMA, FDMA, TDMA, OFDMA, SC-FDMA, etc. CDMA can be implemented with wireless technologies such as Universal Terrestrial Radio Access (UTRA) or CDMA2000. TDMA can be implemented with wireless technologies such as Global System for Mobile communications (GSM)/General Packet Radio Service (GPRS)/Enhanced Data Rates for GSM Evolution (EDGE). OFDMA can be implemented with wireless technologies such as IEEE 802.11 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802-20, Evolved UTRA (E-UTRA), etc. UTRA is part of the Universal Mobile Telecommunications System (UMTS). 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) is part of Evolved UMTS (E-UMTS) using E-UTRA, and LTE-A (Advanced)/LTE-A pro is an evolved version of 3GPP LTE. 3GPP NR (New Radio or New Radio Access Technology) is an evolved version of 3GPP LTE/LTE-A/LTE-A pro. 3GPP 6G may be an evolved version of 3GPP NR.

For clarity of explanation, the description is based on a 3GPP communication system (eg, LTE, NR, etc.), but the technical idea of the present invention is not limited thereto. LTE refers to technology after 3GPP TS 36.xxx Release 8. In detail, LTE technology after 3GPP TS 36.xxx Release 10 is referred to as LTE-A, and LTE technology after 3GPP TS 36.xxx Release 13 is referred to as LTE-A pro. 3GPP NR refers to technology after TS 38.xxx Release 15. 3GPP 6G may refer to technologies after TS Release 17 and/or Release 18. “xxx” refers to the standard document detail number. LTE/NR/6G can be collectively referred to as a 3GPP system. Regarding background technology, terms, abbreviations, etc. used in the description of the present invention, matters described in standard documents published before the present invention may be referred to. For example, you can refer to the following document:

3GPP LTE

- 36.211: Physical channels and modulation

- 36.212: Multiplexing and channel coding

- 36.213: Physical layer procedures

- 36.300: Overall description

- 36.331: Radio Resource Control (RRC)

3GPP NR

- 38.211: Physical channels and modulation

- 38.212: Multiplexing and channel coding

- 38.213: Physical layer procedures for control

- 38.214: Physical layer procedures for data

- 38.300: NR and NG-RAN Overall Description

- 38.331: Radio Resource Control (RRC) protocol specification

Physical Channel and Frame Structure

Physical channels and typical signal transmission

Figure 1 is a diagram showing an example of physical channels and general signal transmission used in a 3GPP system.

In a wireless communication system, a terminal receives information from a base station through downlink (DL), and the terminal transmits information to the base station through uplink (UL). The information transmitted and received between the base station and the terminal includes data and various control information, and various physical channels exist depending on the type/purpose of the information they transmit and receive.

When the terminal is turned on or enters a new cell, it performs an initial cell search task such as synchronizing with the base station (S11). To this end, the terminal can receive a primary synchronization signal (PSS) and a secondary synchronization signal (SSS) from the base station to synchronize with the base station and obtain information such as cell ID. Afterwards, the terminal can receive broadcast information within the cell by receiving a physical broadcast channel (PBCH) from the base station. Meanwhile, the terminal can check the downlink channel status by receiving a downlink reference signal (DL RS) in the initial cell search stage.

After completing the initial cell search, the terminal acquires more specific system information by receiving a physical downlink control channel (PDCCH) and a physical downlink shared channel (PDSCH) according to the information carried in the PDCCH. You can do it (S12).

Meanwhile, when accessing the base station for the first time or when there are no radio resources for signal transmission, the terminal can perform a random access procedure (RACH) to the base station (S13 to S16). To this end, the terminal transmits a specific sequence as a preamble through a physical random access channel (PRACH) (S13 and S15), and a response message (RAR (Random Access In the case of contention-based RACH, a contention resolution procedure can be additionally performed (S16).

The terminal that has performed the above-described procedure can then perform PDCCH/PDSCH reception (S17) and Physical Uplink Shared Channel (PUSCH)/Physical Uplink Control Channel (Physical Uplink) as a general uplink/downlink signal transmission procedure. Control Channel; PUCCH) transmission (S18) can be performed. In particular, the terminal can receive downlink control information (DCI) through PDCCH. Here, DCI includes control information such as resource allocation information for the terminal, and different formats may be applied depending on the purpose of use.

Meanwhile, the control information that the terminal transmits to the base station through uplink or that the terminal receives from the base station includes downlink/uplink ACK/NACK signals, CQI (Channel Quality Indicator), PMI (Precoding Matrix Index), and RI (Rank Indicator). ), etc. may be included. The terminal can transmit control information such as the above-described CQI/PMI/RI through PUSCH and/or PUCCH.

Structure of uplink and downlink channels

Downlink channel structure

The base station transmits related signals to the terminal through a downlink channel described later, and the terminal receives related signals from the base station through a downlink channel described later.

(1) Physical downlink shared channel (PDSCH)

PDSCH carries downlink data (e.g., DL-shared channel transport block, DL-SCH TB), and modulation methods such as QPSK (Quadrature Phase Shift Keying), 16 QAM (Quadrature Amplitude Modulation), 64 QAM, and 256 QAM are used. Applies. A codeword is generated by encoding TB. PDSCH can carry multiple codewords. Scrambling and modulation mapping are performed for each codeword, and modulation symbols generated from each codeword are mapped to one or more layers (Layer mapping). Each layer is mapped to resources along with DMRS (Demodulation Reference Signal), generated as an OFDM symbol signal, and transmitted through the corresponding antenna port.

(2) Physical downlink control channel (PDCCH)

PDCCH carries downlink control information (DCI) and QPSK modulation method is applied. One PDCCH consists of 1, 2, 4, 8, or 16 CCEs (Control Channel Elements) depending on the AL (Aggregation Level). One CCE consists of six REGs (Resource Element Group). One REG is defined by one OFDM symbol and one (P)RB.

The terminal obtains DCI transmitted through the PDCCH by performing decoding (aka blind decoding) on a set of PDCCH candidates. The set of PDCCH candidates that the terminal decodes is defined as the PDCCH search space set. The search space set may be a common search space or a UE-specific search space. The UE can obtain DCI by monitoring PDCCH candidates within one or more search space sets set by MIB or higher layer signaling.

Uplink channel structure

The terminal transmits related signals to the base station through an uplink channel, which will be described later, and the base station will receive the related signals from the terminal through an uplink channel, which will be described later.

(1) Physical uplink shared channel (PUSCH)

PUSCH carries uplink data (e.g., UL-shared channel transport block, UL-SCH TB) and/or uplink control information (UCI), and CP-OFDM (Cyclic Prefix ? Orthogonal Frequency Division Multiplexing) waveform. , DFT-s-OFDM (Discrete Fourier Transform ? spread - Orthogonal Frequency Division Multiplexing) waveform, etc. are transmitted. When the PUSCH is transmitted based on the DFT-s-OFDM waveform, the terminal transmits the PUSCH by applying transform precoding. For example, if transform precoding is not possible (e.g., transform precoding is disabled), the terminal transmits PUSCH based on the CP-OFDM waveform, and if transform precoding is possible (e.g., transform precoding is enabled), the terminal transmits CP-OFDM. PUSCH can be transmitted based on the waveform or DFT-s-OFDM waveform. PUSCH transmission is scheduled dynamically by UL grant within DCI, or semi-statically based on upper layer (e.g., RRC) signaling (and/or Layer 1 (L1) signaling (e.g., PDCCH)). Can be scheduled (configured grant). PUSCH transmission can be performed based on codebook or non-codebook.

(2) Physical Uplink Control Channel (PUCCH)

PUCCH carries uplink control information, HARQ-ACK, and/or scheduling request (SR), and can be divided into multiple PUCCHs depending on the PUCCH transmission length.

Hereinafter, a new radio access technology (new RAT, NR) will be described.

As more communication devices require greater communication capacity, there is a need for improved mobile broadband communication compared to existing radio access technology (RAT). Additionally, Massive Machine Type Communications (MTC), which provides various services anytime, anywhere by connecting multiple devices and objects, is also one of the major issues to be considered in next-generation communications. In addition, communication system design considering services/terminals sensitive to reliability and latency is being discussed. In this way, the introduction of next-generation wireless access technology considering expanded mobile broadband communication, massive MTC, URLLC (Ultra-Reliable and Low Latency Communication), etc. is being discussed, and various embodiments of the present disclosure are used for convenience. The technology is called new RAT or NR.

Figure 2 is a diagram showing the system structure of a next-generation radio access network (NG-RAN).

Referring to FIG. 2, NG-RAN may include a gNB and/or eNB that provide user plane and control plane protocol termination to the UE. Figure 1 illustrates a case including only gNB. gNB and eNB are connected to each other through the Xn interface. gNB and eNB are connected through the 5G Core Network (5GC) and NG interface. More specifically, it is connected to the access and mobility management function (AMF) through the NG-C interface, and to the user plane function (UPF) through the NG-U interface.

Figure 3 is a diagram showing the functional division between NG-RAN and 5GC.

Referring to FIG. 3, gNB performs inter-cell radio resource management (Inter Cell RRM), radio bearer management (RB control), connection mobility control, radio admission control, and measurement configuration and provision. Functions such as (Measurement configuration & Provision) and dynamic resource allocation can be provided. AMF can provide functions such as NAS security and idle state mobility handling. UPF can provide functions such as mobility anchoring and PDU processing. SMF (Session Management Function) can provide functions such as terminal IP address allocation and PDU session control.

Figure 4 is a diagram showing an example of a 5G usage scenario.

The 5G usage scenario shown in FIG. 4 is merely illustrative, and technical features of various embodiments of the present disclosure may also be applied to other 5G usage scenarios not shown in FIG. 4.

Referring to Figure 4, the three main requirements areas for 5G are (1) enhanced mobile broadband (eMBB) area, (2) massive machine type communication (mMTC) area, and ( 3) Includes ultra-reliable and low latency communications (URLLC) areas. Some use cases may require multiple areas for optimization, while others may focus on just one key performance indicator (KPI). 5G supports these diverse use cases in a flexible and reliable way.

eMBB focuses on overall improvements in data speeds, latency, user density, capacity and coverage of mobile broadband connections. eMBB aims for a throughput of around 10Gbps. eMBB goes far beyond basic mobile Internet access and covers rich interactive tasks, media and entertainment applications in the cloud or augmented reality. Data is one of the key drivers of 5G, and we may not see dedicated voice services for the first time in the 5G era. In 5G, voice is expected to be processed simply as an application using the data connection provided by the communication system. The main reasons for the increased traffic volume are the increase in content size and the number of applications requiring high data transfer rates. Streaming services (audio and video), interactive video and mobile Internet connections will become more prevalent as more devices are connected to the Internet. Many of these applications require always-on connectivity to push real-time information and notifications to users. Cloud storage and applications are rapidly increasing mobile communication platforms, and this can apply to both work and entertainment. Cloud storage is a particular use case driving growth in uplink data rates. 5G will also be used for remote work in the cloud and will require much lower end-to-end latency to maintain a good user experience when tactile interfaces are used. In entertainment, for example, cloud gaming and video streaming are other key factors driving increased demand for mobile broadband capabilities. Entertainment is essential on smartphones and tablets everywhere, including high mobility environments such as trains, cars and planes. Another use case is augmented reality for entertainment and information retrieval. Here, augmented reality requires very low latency and instantaneous amounts of data.

mMTC is designed to enable communication between large numbers of low-cost devices powered by batteries, and is intended to support applications such as smart metering, logistics, and field and body sensors. mMTC targets batteries with a lifespan of around 10 years and/or around 1 million devices per km2. mMTC enables seamless connectivity of embedded sensors in all fields and is one of the most anticipated 5G use cases. Potentially, IoT devices are expected to reach 20.4 billion by 2020. Industrial IoT is one area where 5G will play a key role in enabling smart cities, asset tracking, smart utilities, agriculture and security infrastructure.

URLLC enables devices and machines to communicate highly reliably, with very low latency and high availability, making it ideal for vehicle communications, industrial control, factory automation, remote surgery, smart grid and public safety applications. URLLC aims for a delay of around 1ms. URLLC includes new services that will transform industries through ultra-reliable/low-latency links, such as remote control of critical infrastructure and autonomous vehicles. Levels of reliability and latency are essential for smart grid control, industrial automation, robotics, and drone control and coordination.

Next, we will look in more detail at the multiple use examples included in the triangle in FIG. 4.

5G can complement fiber-to-the-home (FTTH) and cable-based broadband (or DOCSIS) as a means of delivering streams rated at hundreds of megabits per second to gigabits per second. Such high speeds may be required to deliver TV at resolutions of 4K and higher (6K, 8K and beyond) as well as virtual reality (VR) and augmented reality (AR). VR and AR applications include nearly immersive sporting events. Certain applications may require special network settings. For example, for VR games, gaming companies may need to integrate their core servers with a network operator's edge network servers to minimize latency.

Automotive is expected to be an important new driver for 5G, with many use cases for mobile communications for vehicles. For example, entertainment for passengers requires both high capacity and high mobile broadband. That's because future users will continue to expect high-quality connections regardless of their location and speed. Another use case in the automotive sector is augmented reality dashboards. An augmented reality contrast board allows the driver to identify objects in the dark above what he or she is seeing through the front window. The augmented reality dashboard displays information that informs the driver about the distance and movement of objects. In the future, wireless modules will enable communication between vehicles, information exchange between vehicles and supporting infrastructure, and information exchange between vehicles and other connected devices (eg, devices accompanied by pedestrians). Safety systems can reduce the risk of accidents by guiding drivers to alternative courses of action to help them drive safer. The next step will be remotely controlled or autonomous vehicles. This requires highly reliable and very fast communication between different autonomous vehicles and/or between vehicles and infrastructure. In the future, autonomous vehicles will perform all driving activities, leaving drivers to focus only on traffic abnormalities that the vehicles themselves cannot discern. The technical requirements of autonomous vehicles call for ultra-low latency and ultra-high reliability, increasing traffic safety to levels unachievable by humans.

Smart cities and smart homes, referred to as smart societies, will be embedded with high-density wireless sensor networks. A distributed network of intelligent sensors will identify conditions for maintaining a city or home cost-effectively and energy-efficiently. A similar setup can be done for each household. Temperature sensors, window and heating controllers, burglar alarms and home appliances are all connected wirelessly. Many of these sensors typically require low data rates, low power, and low cost. However, real-time HD video may be required in certain types of devices for surveillance, for example.

Consumption and distribution of energy, including heat or gas, is highly decentralized, requiring automated control of distributed sensor networks. A smart grid interconnects these sensors using digital information and communications technologies to collect and act on information. This information can include the behavior of suppliers and consumers, allowing smart grids to improve the efficiency, reliability, economics, sustainability of production and distribution of fuels such as electricity in an automated manner. Smart grid can also be viewed as another low-latency sensor network.

The health sector has many applications that can benefit from mobile communications. Communications systems can support telemedicine, providing clinical care in remote locations. This can help reduce the barrier of distance and improve access to health services that are consistently unavailable in remote rural areas. It is also used to save lives in critical care and emergency situations. Mobile communication-based wireless sensor networks can provide remote monitoring and sensors for parameters such as heart rate and blood pressure.

Wireless and mobile communications are becoming increasingly important in industrial applications. Wiring is expensive to install and maintain. Therefore, the possibility of replacing cables with reconfigurable wireless links is an attractive opportunity for many industries. However, achieving this requires that wireless connections operate with similar latency, reliability and capacity as cables, and that their management be simplified. Low latency and very low error probability are new requirements needed for 5G connectivity.

Logistics and cargo tracking are important use cases for mobile communications, enabling tracking of inventory and packages from anywhere using location-based information systems. Use cases in logistics and cargo tracking typically require low data rates but may require large ranges and reliable location information.

Hereinafter, examples of next-generation communication (eg, 6G) that can be applied to various embodiments of the present disclosure will be described.

6G system general

6G (wireless communications) systems require (i) very high data rates per device, (ii) very large number of connected devices, (iii) global connectivity, (iv) very low latency, (v) battery- The goals are to reduce the energy consumption of battery-free IoT devices, (vi) ultra-reliable connectivity, and (vii) connected intelligence with machine learning capabilities. The vision of the 6G system can be four aspects such as intelligent connectivity, deep connectivity, holographic connectivity, and ubiquitous connectivity, and the 6G system can satisfy the requirements as shown in Table 1 below. That is, Table 1 is a table showing an example of the requirements of a 6G system.

Per device peak data rate 1 Tbps E2E latency 1ms Maximum spectral efficiency 100bps/Hz Mobility support Up to 1000km/hr Satellite integration Fully A.I. Fully Autonomous vehicle Fully XR Fully Haptic Communication Fully

The 6G system includes Enhanced mobile broadband (eMBB), Ultra-reliable low latency communications (URLLC), massive machine-type communication (mMTC), AI integrated communication, Tactile internet, High throughput, High network capacity, High energy efficiency, Low backhaul and It can have key factors such as access network congestion and enhanced data security.

Figure 5 is a diagram showing an example of a communication structure that can be provided in a 6G system.

The 6G system is expected to have simultaneous wireless communication connectivity that is 50 times higher than that of the 5G wireless communication system. URLLC, a key feature of 5G, will become an even more important technology in 6G communications by providing end-to-end delay of less than 1ms. The 6G system will have much better volumetric spectral efficiency, unlike the frequently used area spectral efficiency. 6G systems can provide ultra-long battery life and advanced battery technologies for energy harvesting, so mobile devices in 6G systems will not need to be separately charged. New network characteristics in 6G may include:

- Satellites integrated network: 6G is expected to be integrated with satellites to serve the global mobile constellation. Integration of terrestrial, satellite and aerial networks into one wireless communication system is very important for 6G.

- Connected intelligence: Unlike previous generations of wireless communication systems, 6G is revolutionary and will update the evolution of wireless from “connected things” to “connected intelligence.” AI will be used to control each step of the communication process (or signal, as described later). can be applied in each procedure of processing).

- Seamless integration wireless information and energy transfer: 6G wireless networks will deliver power to charge the batteries of devices such as smartphones and sensors. Therefore, wireless information and energy transfer (WIET) will be integrated.

- Ubiquitous super 3D connectivity: Connectivity of drones and very low Earth orbit satellites to networks and core network functions will create super 3D connectivity in 6G ubiquitous.

In the above new network characteristics of 6G, some general requirements may be as follows.

- Small cell networks: The idea of small cell networks was introduced to improve received signal quality resulting in improved throughput, energy efficiency and spectral efficiency in cellular systems. As a result, small cell networks are an essential feature for 5G and beyond 5G (5GB) communications systems. Therefore, the 6G communication system also adopts the characteristics of a small cell network.

- Ultra-dense heterogeneous network: Ultra-dense heterogeneous networks will be another important characteristic of the 6G communication system. Multi-tier networks comprised of heterogeneous networks improve overall QoS and reduce costs.

- High-capacity backhaul: Backhaul connections are characterized by high-capacity backhaul networks to support high-capacity traffic. High-speed optical fiber and free space optics (FSO) systems may be possible solutions to this problem.

- Radar technology integrated with mobile technology: High-precision localization (or location-based services) through communication is one of the functions of the 6G wireless communication system. Therefore, radar systems will be integrated with 6G networks.

- Softwarization and virtualization: Softwarization and virtualization are two important features that are fundamental to the design process in 5GB networks to ensure flexibility, reconfigurability, and programmability. Additionally, billions of devices may be shared on a shared physical infrastructure.

Core implementation technology of 6G system

Artificial Intelligence

The most important and newly introduced technology in the 6G system is AI. AI was not involved in the 4G system. 5G systems will support partial or very limited AI. However, 6G systems will be AI-enabled for full automation. Advances in machine learning will create more intelligent networks for real-time communications in 6G. Introducing AI in communications can simplify and improve real-time data transmission. AI can use numerous analytics to determine how complex target tasks are performed. In other words, AI can increase efficiency and reduce processing delays.

Time-consuming tasks such as handover, network selection, and resource scheduling can be performed instantly by using AI. AI can also play an important role in M2M, machine-to-human and human-to-machine communications. Additionally, AI can enable rapid communication in BCI (Brain Computer Interface). AI-based communication systems can be supported by metamaterials, intelligent structures, intelligent networks, intelligent devices, intelligent cognitive radios, self-sustaining wireless networks, and machine learning.

Recently, attempts have been made to integrate AI with wireless communication systems, but these have been focused on the application layer and network layer, especially deep learning in the field of wireless resource management and allocation. However, this research is gradually progressing to the MAC layer and Physical layer, and in particular, attempts are being made to combine deep learning with wireless transmission in the Physical layer. AI-based physical layer transmission means applying signal processing and communication mechanisms based on AI drivers, rather than traditional communication frameworks, in terms of fundamental signal processing and communication mechanisms. For example, deep learning-based channel coding and decoding, deep learning-based signal estimation and detection, deep learning-based MIMO mechanism, AI-based resource scheduling, and May include allocation, etc.

Machine learning can be used for channel estimation and channel tracking, and can be used for power allocation, interference cancellation, etc. in the physical layer of the DL (downlink). Machine learning can also be used for antenna selection, power control, and symbol detection in MIMO systems.

However, application of DNN for transmission in the physical layer may have the following problems.

Deep learning-based AI algorithms require a large amount of training data to optimize training parameters. However, due to limitations in acquiring data from a specific channel environment as training data, a lot of training data is used offline. This means that static training on training data in a specific channel environment may result in a contradiction between the dynamic characteristics and diversity of the wireless channel.

Additionally, current deep learning mainly targets real signals. However, signals of the physical layer of wireless communication are complex signals. More research is needed on neural networks that detect complex domain signals to match the characteristics of wireless communication signals.

Below, we will look at machine learning in more detail.

Machine learning refers to a series of operations that train machines to create machines that can perform tasks that are difficult or difficult for humans to perform. Machine learning requires data and a learning model. In machine learning, data learning methods can be broadly divided into three types: supervised learning, unsupervised learning, and reinforcement learning.

Neural network learning is intended to minimize errors in output. Neural network learning repeatedly inputs learning data into the neural network, calculates the output of the neural network and the error of the target for the learning data, and backpropagates the error of the neural network from the output layer of the neural network to the input layer to reduce the error. ) is the process of updating the weight of each node in the neural network.

Supervised learning uses training data in which the correct answer is labeled, while unsupervised learning may not have the correct answer labeled in the learning data. That is, for example, in the case of supervised learning on data classification, the learning data may be data in which each training data is labeled with a category. Labeled learning data is input to a neural network, and error can be calculated by comparing the output (category) of the neural network with the label of the learning data. The calculated error is backpropagated in the reverse direction (i.e., from the output layer to the input layer) in the neural network, and the connection weight of each node in each layer of the neural network can be updated according to backpropagation. The amount of change in the connection weight of each updated node may be determined according to the learning rate. The neural network's calculation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently depending on the number of repetitions of the learning cycle of the neural network. For example, in the early stages of neural network training, a high learning rate can be used to ensure that the neural network quickly achieves a certain level of performance to increase efficiency, and in the later stages of training, a low learning rate can be used to increase accuracy.

Learning methods may vary depending on the characteristics of the data. For example, in a communication system, when the goal is to accurately predict data transmitted from a transmitter at a receiver, it is preferable to perform learning using supervised learning rather than unsupervised learning or reinforcement learning.

The learning model corresponds to the human brain, and can be considered the most basic linear model. However, deep learning is a machine learning paradigm that uses a highly complex neural network structure, such as artificial neural networks, as a learning model. ).

Neural network cores used as learning methods are broadly divided into deep neural networks (DNN), convolutional deep neural networks (CNN), and Recurrent Boltzmann Machine (RNN). there is.

An artificial neural network is an example of connecting multiple perceptrons.

Figure 6 is a diagram schematically showing an example of a perceptron structure.

Referring to FIG. 6, when the input vector x=(x1,x2,...,xd) is input, each component is multiplied by weights (W1,W2,...,Wd), and the results are added together, The entire process of applying the activation function σ(·) is called a perceptron. A large artificial neural network structure can extend the simplified perceptron structure shown in FIG. 6 and apply input vectors to different multi-dimensional perceptrons. For convenience of explanation, input or output values are referred to as nodes.

Meanwhile, the perceptron structure shown in FIG. 6 can be described as consisting of a total of three layers based on input and output values. An artificial neural network in which there are H (d+1)-dimensional perceptrons between the 1st layer and the 2nd layer, and K (H+1)-dimensional perceptrons between the 2nd layer and the 3rd layer can be expressed as shown in Figure 7.

Figure 7 is a diagram schematically showing an example of a multi-layer perceptron structure.

The layer where the input vector is located is called the input layer, the layer where the final output value is located is called the output layer, and all layers located between the input layer and the output layer are called hidden layers. In the example of Figure 7, three layers are disclosed, but when counting the actual number of artificial neural network layers, the input layer is counted excluding the input layer, so it can be viewed as a total of two layers. An artificial neural network is constructed by two-dimensionally connecting perceptrons of basic blocks.

The above-described input layer, hidden layer, and output layer can be jointly applied not only to the multi-layer perceptron, but also to various artificial neural network structures such as CNN and RNN, which will be described later. As the number of hidden layers increases, the artificial neural network becomes deeper, and the machine learning paradigm that uses a sufficiently deep artificial neural network as a learning model is called deep learning. Additionally, the artificial neural network used for deep learning is called a deep neural network (DNN).

Figure 8 is a diagram schematically showing an example deep neural network.

The deep neural network shown in Figure 8 is a multi-layer perceptron consisting of 8 hidden layers and 8 output layers. The multi-layer perceptron structure is expressed as a fully-connected neural network. In a fully connected neural network, no connection exists between nodes located on the same layer, and only between nodes located on adjacent layers. DNN has a fully connected neural network structure and is composed of a combination of multiple hidden layers and activation functions, so it can be usefully applied to identify correlation characteristics between input and output. Here, the correlation characteristic may mean the joint probability of input and output.

Meanwhile, depending on how multiple perceptrons are connected to each other, various artificial neural network structures different from the above-mentioned DNN can be formed.

Figure 9 is a diagram schematically showing an example of a convolutional neural network.

In DNN, nodes located inside one layer are arranged in a one-dimensional vertical direction. However, in Figure 9, it can be assumed that the nodes are arranged two-dimensionally with w nodes horizontally and h nodes vertically (convolutional neural network structure in Figure 9). In this case, a weight is added to each connection during the connection process from one input node to the hidden layer, so a total of hХw weights must be considered. Since there are hХw nodes in the input layer, a total of h2w2 weights are needed between two adjacent layers.

The convolutional neural network of Figure 9 has a problem in that the number of weights increases exponentially according to the number of connections, so instead of considering all mode connections between adjacent layers, it is assumed that a small filter exists, and the number of weights increases exponentially according to the number of connections. As shown, weighted sum and activation function calculations are performed on areas where filters overlap.

Figure 10 is a diagram schematically showing an example of a filter operation in a convolutional neural network.

One filter has a weight corresponding to its size, and the weight can be learned so that a specific feature in the image can be extracted and output as a factor. In Figure 10, a filter of size 3Х3 is applied to the upper leftmost 3Х3 area of the input layer, and the output value as a result of performing weighted sum and activation function calculations for the corresponding node is stored in z22.

The filter scans the input layer and moves at regular intervals horizontally and vertically, performs weighted sum and activation function calculations, and places the output value at the current filter position. This operation method is similar to the convolution operation on images in the field of computer vision, so a deep neural network with this structure is called a convolutional neural network (CNN), and the hidden layer generated as a result of the convolution operation is is called a convolutional layer. Additionally, a neural network with multiple convolutional layers is called a deep convolutional neural network (DCNN).

In the convolutional layer, the number of weights can be reduced by calculating a weighted sum by including only nodes located in the area covered by the filter from the node where the current filter is located. Because of this, one filter can be used to focus on features for a local area. Accordingly, CNN can be effectively applied to image data processing where physical distance in a two-dimensional area is an important decision criterion. Meanwhile, CNN may have multiple filters applied immediately before the convolution layer, and may generate multiple output results through the convolution operation of each filter.

Meanwhile, depending on the data properties, there may be data for which sequence characteristics are important. Considering the length variability and precedence relationship of these sequence data, elements in the data sequence are input one by one at each time step, and the output vector (hidden vector) of the hidden layer output at a specific time point is input together with the next element in the sequence. The structure that applies this method to an artificial neural network is called a recurrent neural network structure.

Figure 11 is a diagram schematically showing an example of a neural network structure in which a cyclic loop exists.

Referring to FIG. 11, a recurrent neural network (RNN) connects the elements (x1(t), During the input process, at the previous point t-1, the hidden vector (z1(t-1), z2(t-1),..., zH(t-1)) is input together to perform the weighted sum and activation function. This is the structure that is applied. The reason for passing the hidden vector to the next time point like this is because the information in the input vector from previous time points is considered to be accumulated in the hidden vector at the current time point.

Figure 12 is a diagram schematically showing an example of the operating structure of a recurrent neural network.

Referring to FIG. 12, the recurrent neural network operates in a predetermined time point order with respect to the input data sequence.

Hidden vector (z1(1),z2(1),.. when the input vector (x1(t), .,zH(1)) is input together with the input vector (x1(2),x2(2),...,xd(2)) of viewpoint 2, and the vector of the hidden layer (z1( 2) Determine z2(2) ,...,zH(2)). This process is performed repeatedly until time point 2, time point 3, ,,, time point T.

Meanwhile, when multiple hidden layers are placed within a recurrent neural network, it is called a deep recurrent neural network (DRNN). Recurrent neural networks are designed to be usefully applied to sequence data (for example, natural language processing).

As a neural network core used as a learning method, in addition to DNN, CNN, and RNN, it includes Restricted Boltzmann Machine (RBM), deep belief networks (DBN), Deep Q-Network, and It includes various deep learning techniques, and can be applied to fields such as computer vision, speech recognition, natural language processing, and voice/signal processing.

Recently, attempts have been made to integrate AI with wireless communication systems, but these have been focused on the application layer and network layer, especially deep learning in the field of wireless resource management and allocation. However, this research is gradually progressing to the MAC layer and Physical layer, and in particular, attempts are being made to combine deep learning with wireless transmission in the Physical layer. AI-based physical layer transmission means applying signal processing and communication mechanisms based on AI drivers, rather than traditional communication frameworks, in terms of fundamental signal processing and communication mechanisms. For example, deep learning-based channel coding and decoding, deep learning-based signal estimation and detection, deep learning-based MIMO mechanism, AI-based resource scheduling, and May include allocation, etc.

Terahertz (THz) communications

The data transfer rate can be increased by increasing the bandwidth. This can be accomplished by using sub-THz communications with wide bandwidth and applying advanced massive MIMO technology. THz waves, also known as submillimeter radiation, typically represent a frequency band between 0.1 THz and 10 THz with a corresponding wavelength in the range 0.03 mm-3 mm. The 100GHz-300GHz band range (Sub THz band) is considered the main part of the THz band for cellular communications. Adding the sub-THz band to the mmWave band increases 6G cellular communication capacity. Among the defined THz bands, 300GHz-3THz is in the far infrared (IR) frequency band. The 300GHz-3THz band is part of the wideband, but it is at the border of the wideband and immediately behind the RF band. Therefore, this 300 GHz-3 THz band shows similarities to RF.

Figure 13 is a diagram showing an example of an electromagnetic spectrum.

Key characteristics of THz communications include (i) widely available bandwidth to support very high data rates, (ii) high path loss occurring at high frequencies (highly directional antennas are indispensable). The narrow beamwidth produced by a highly directional antenna reduces interference. The small wavelength of THz signals allows a much larger number of antenna elements to be integrated into devices and BSs operating in this band. This enables the use of advanced adaptive array techniques that can overcome range limitations.

Optical wireless technology

OWC technology is planned for 6G communications in addition to RF-based communications for all possible device-to-access networks. These networks connect to network-to-backhaul/fronthaul network connections. OWC technology has already been used since 4G communication systems, but will be more widely used to meet the needs of 6G communication systems. OWC technologies such as light fidelity, visible light communication, optical camera communication, and optical bandwidth-based FSO communication are already well-known technologies. Communications based on optical wireless technology can provide very high data rates, low latency, and secure communications. LiDAR can also be used for ultra-high-resolution 4D mapping in 6G communications based on wide bandwidth.

FSO backhaul network

The transmitter and receiver characteristics of an FSO system are similar to those of a fiber optic network. Therefore, data transmission in FSO systems is similar to fiber optic systems. Therefore, FSO can be a good technology to provide backhaul connectivity in 6G systems along with fiber optic networks. Using FSO, very long-distance communication is possible, even over distances of 10,000 km. FSO supports high-capacity backhaul connections for remote and non-remote areas such as oceans, space, underwater, and isolated islands. FSO also supports cellular BS connections.

Massive MIMO technology

One of the key technologies to improve spectral efficiency is applying MIMO technology. As MIMO technology improves, spectral efficiency also improves. Therefore, massive MIMO technology will be important in 6G systems. Because MIMO technology uses multiple paths, multiplexing technology and beam generation and operation technology suitable for the THz band must be carefully considered so that data signals can be transmitted through more than one path.

blockchain

Blockchain will become an important technology for managing large amounts of data in future communication systems. Blockchain is a form of distributed ledger technology, where a distributed ledger is a database distributed across numerous nodes or computing devices. Each node replicates and stores a copy of the same ledger. Blockchain is managed as a P2P network. It can exist without being managed by a centralized authority or server. Data in a blockchain is collected together and organized into blocks. Blocks are linked together and protected using encryption. Blockchain is a perfect complement to large-scale IoT through its inherently improved interoperability, security, privacy, reliability, and scalability. Therefore, blockchain technology provides several features such as interoperability between devices, large-scale data traceability, autonomous interaction of other IoT systems, and large-scale connection stability in 6G communication systems.

3D networking

The 6G system integrates terrestrial and aerial networks to support vertically expanded user communications. 3D BS will be provided via low-orbit satellites and UAVs. Adding new dimensions in terms of altitude and associated degrees of freedom makes 3D connections significantly different from traditional 2D networks.

quantum communication

In the context of 6G networks, unsupervised reinforcement learning of networks is promising. Supervised learning methods cannot label the massive amounts of data generated by 6G. Unsupervised learning does not require labeling. Therefore, this technique can be used to autonomously build representations of complex networks. Combining reinforcement learning and unsupervised learning allows the network to operate in a truly autonomous manner.

drone

Unmanned Aerial Vehicles (UAVs), or drones, will become an important element in 6G wireless communications. In most cases, high-speed data wireless connectivity is provided using UAV technology. The BS entity is installed on the UAV to provide cellular connectivity. UAVs have certain features not found in fixed BS infrastructure, such as easy deployment, strong line-of-sight links, and controlled degrees of freedom for mobility. During emergency situations such as natural disasters, the deployment of terrestrial communications infrastructure is not economically feasible and sometimes cannot provide services in volatile environments. UAVs can easily handle these situations. UAV will become a new paradigm in the wireless communication field. This technology facilitates three basic requirements of wireless networks: eMBB, URLLC, and mMTC. UAVs can also support several purposes, such as improving network connectivity, fire detection, disaster emergency services, security and surveillance, pollution monitoring, parking monitoring, accident monitoring, etc. Therefore, UAV technology is recognized as one of the most important technologies for 6G communications.

Cell-free Communication

Tight integration of multiple frequencies and heterogeneous communication technologies is very important in 6G systems. As a result, users can seamlessly move from one network to another without having to make any manual configuration on their devices. The best network is automatically selected from the available communication technologies. This will break the limitations of the cell concept in wireless communications. Currently, user movement from one cell to another causes too many handovers in high-density networks, causing handover failures, handover delays, data loss, and ping-pong effects. 6G cell-free communication will overcome all of this and provide better QoS. Cell-free communications will be achieved through multi-connectivity and multi-tier hybrid technologies and different heterogeneous radios in devices.

Integration of wireless information and energy transmission

WIET uses the same fields and waves as wireless communication systems. In particular, sensors and smartphones will be charged using wireless power transfer during communication. WIET is a promising technology for extending the life of battery-charged wireless systems. Therefore, devices without batteries will be supported in 6G communications.

Integration of sensing and communication

An autonomous wireless network is the ability to continuously sense dynamically changing environmental conditions and exchange information between different nodes. In 6G, sensing will be tightly integrated with communications to support autonomous systems.

Consolidation of access backhaul networks

The density of access networks in 6G will be enormous. Each access network is connected by backhaul connections such as fiber optics and FSO networks. To cope with the very large number of access networks, there will be tight integration between access and backhaul networks.

Holographic Beam Forming

Beamforming is a signal processing procedure that adjusts antenna arrays to transmit wireless signals in a specific direction. It is a subset of smart antennas or advanced antenna systems. Beamforming technology has several advantages, such as high call-to-noise ratio, interference prevention and rejection, and high network efficiency. Holographic beamforming (HBF) is a new beamforming method that differs significantly from MIMO systems because it uses software-defined antennas. HBF will be a very effective approach for efficient and flexible transmission and reception of signals in multi-antenna communication devices in 6G.

big data analytics

Big data analytics is a complex process for analyzing various large data sets or big data. This process ensures complete data management by uncovering information such as hidden data, unknown correlations, and customer preferences. Big data is collected from various sources such as videos, social networks, images, and sensors. This technology is widely used to process massive amounts of data in 6G systems.

Large Intelligent Surface (LIS)

In the case of THz band signals, the linearity is strong, so there may be many shadow areas due to obstructions. By installing LIS near these shadow areas, LIS technology that expands the communication area, strengthens communication stability, and enables additional value-added services becomes important. LIS is an artificial surface made of electromagnetic materials and can change the propagation of incoming and outgoing radio waves. LIS can be seen as an extension of massive MIMO, but has a different array structure and operating mechanism from massive MIMO. Additionally, LIS operates as a reconfigurable reflector with passive elements, i.e., it only passively reflects signals without using an active RF chain, resulting in low power consumption. There are advantages to having one. Additionally, because each passive reflector of LIS must independently adjust the phase shift of the incident signal, this can be advantageous for wireless communication channels. By appropriately adjusting the phase shift through the LIS controller, the reflected signal can be collected at the target receiver to boost the received signal power.

Terahertz (THz) wireless communication general

THz wireless communication uses wireless communication using THz waves with a frequency of approximately 0.1 to 10 THz (1 THz = 1012 Hz), and can refer to terahertz (THz) band wireless communication using a very high carrier frequency of 100 GHz or more. . THz waves are located between RF (Radio Frequency)/millimeter (mm) and infrared bands. (i) Compared to visible light/infrared, they penetrate non-metal/non-polarized materials better and have a shorter wavelength than RF/millimeter waves, so they have high straightness. Beam focusing may be possible. In addition, the photon energy of THz waves is only a few meV, so it is harmless to the human body. The frequency band expected to be used for THz wireless communication may be the D-band (110GHz to 170GHz) or H-band (220GHz to 325GHz) bands, which have small propagation loss due to absorption of molecules in the air. In addition to 3GPP, standardization discussions on THz wireless communication are being discussed centered around the IEEE 802.15 THz working group, and standard documents issued by the IEEE 802.15 Task Group (TG3d, TG3e) specify the content described in various embodiments of the present disclosure. Or it can be supplemented. THz wireless communication can be applied to wireless cognition, sensing, imaging, wireless communication, THz navigation, etc.

Figure 14 is a diagram showing an example of THz communication application.

As shown in FIG. 14, THz wireless communication scenarios can be classified into macro network, micro network, and nanoscale network. In macro networks, THz wireless communication can be applied to vehicle-to-vehicle connections and backhaul/fronthaul connections. In micronetworks, THz wireless communication has applications in indoor small cells, fixed point-to-point or multi-point connections such as wireless connections in data centers, and near-field communication such as kiosk downloading. It can be.

Table 2 below shows an example of technology that can be used in THz waves.

Transceiver Device Available immatures: UTC-PD, RTD and SBD Modulation and Coding Low order modulation techniques (OOK, QPSK), LDPC, Reed Soloman, Hamming, Polar, Turbo Antenna Omni and Directional, phased array with low number of antenna elements Bandwidth 69GHz (or 23 GHz) at 300GHz Channel models Partially Data rate 100Gbps Outdoor deployment No Free space loss High Coverage Low Radio Measurements 300GHz indoor Device size Few micrometers

THz wireless communication can be classified based on the method for generating and receiving THz. THz generation methods can be classified as optical or electronic device-based technologies.

Figure 15 is a diagram showing an example of an electronic device-based THz wireless communication transceiver.

Methods for generating THz using electronic devices include methods using semiconductor devices such as resonant tunneling diodes (RTDs), methods using local oscillators and multipliers, and integrated circuits based on compound semiconductor HEMT (High Electron Mobility Transistor). There are methods using MMIC (Monolithic Microwave Integrated Circuits) and methods using Si-CMOS-based integrated circuits. In the case of Figure 15, a doubler, tripler, multiplier is applied to increase the frequency, and it passes through the subharmonic mixer and is radiated by the antenna. Since the THz band produces high frequencies, a multiplier is essential. Here, the multiplier is a circuit that has an output frequency N times that of the input, matches it to the desired harmonic frequency, and filters out all remaining frequencies. Additionally, beamforming may be implemented by applying an array antenna to the antenna of FIG. 15. In Figure 15, IF represents intermediate frequency, tripler and multipler represent multiplier, PA power amplifier, LNA is low noise amplifier, and PLL is phase locking circuit. -Locked Loop).

FIG. 16 is a diagram illustrating an example of a method for generating an optical device-based THz signal.

Figure 17 is a diagram showing an example of an optical device-based THz wireless communication transceiver.

Optical device-based THz wireless communication technology refers to a method of generating and modulating THz signals using optical devices. Optical device-based THz signal generation technology is a technology that generates ultra-fast optical signals using lasers and optical modulators, and converts them into THz signals using ultra-fast photo detectors. Compared to technologies using only electronic devices, this technology makes it easier to increase the frequency, enables the generation of high-power signals, and achieves flat response characteristics over a wide frequency band. To generate a THz signal based on an optical device, a laser diode, a broadband optical modulator, and an ultra-fast photodetector are required, as shown in FIG. 16. In the case of Figure 16, light signals from two lasers with different wavelengths are combined to generate a THz signal corresponding to the wavelength difference between the lasers. In Figure 16, an optical coupler refers to a semiconductor device that transmits electrical signals using light waves to provide electrical insulation and coupling between circuits or systems, and is referred to as UTC-PD (Uni-Travelling Carrier Photo-PD). Detector is a type of photodetector that uses electrons as active carriers and is a device that reduces the travel time of electrons through bandgap grading. UTC-PD is capable of photodetection above 150GHz. In Figure 17, EDFA (Erbium-Doped Fiber Amplifier) represents an erbium-doped optical fiber amplifier, PD (Photo Detector) represents a semiconductor device capable of converting optical signals into electrical signals, and OSA represents various optical communication functions (photoelectric). It represents an optical module (Optical Sub Assembly) that modularizes (conversion, electro-optical conversion, etc.) into one component, and DSO stands for digital storage oscilloscope.

The structure of the photoelectric converter (or photoelectric converter) will be described with reference to FIGS. 18 and 19.

Figure 18 is a diagram showing the structure of a photon source-based transmitter.

Figure 19 is a diagram showing the structure of an optical modulator.

In general, the phase of a signal can be changed by passing the optical source of a laser through an optical wave guide. At this time, data is loaded by changing the electrical characteristics through microwave contact, etc. Accordingly, the optical modulator output is formed as a modulated waveform. A photoelectric modulator (O/E converter) operates in optical rectification using a nonlinear crystal, photoelectric conversion using a photoconductive antenna, and a bunch of electrons in light. THz pulses can be generated according to emission from relativistic electrons. A terahertz pulse generated in the above manner may have a length ranging from femto second to pico second. An photoelectric converter (O/E converter) uses the non-linearity of the device to perform down conversion.

Considering the THz spectrum usage, several contiguous GHz bands are needed for terahertz systems, either fixed or mobile service. There is a high possibility of using it. According to the outdoor scenario standard, the available bandwidth can be classified based on oxygen attenuation of 10^2 dB/km in the spectrum up to 1 THz. Accordingly, a framework in which the available bandwidth consists of several band chunks may be considered. As an example of the above framework, if the length of a terahertz pulse for one carrier is set to 50 ps, the bandwidth (BW) is about 20 GHz.

Effective down conversion from the IR band to the THz band depends on how to utilize the nonlinearity of the photoelectric converter (O/E converter). In other words, in order to down-convert to the desired terahertz band (THz band), an photoelectric converter (O/E converter) with the most ideal non-linearity for transfer to the relevant terahertz band (THz band) is required. Design is required. If an photoelectric converter (O/E converter) that does not fit the target frequency band is used, there is a high possibility that errors will occur in the amplitude and phase of the corresponding pulse.

In a single carrier system, a terahertz transmission/reception system can be implemented using one photoelectric converter. Although it varies depending on the channel environment, a multi-carrier system may require as many photoelectric converters as the number of carriers. This phenomenon will be particularly noticeable in the case of multi-carrier systems that utilize multiple bandwidths according to the above-described spectrum usage plans. In this regard, a frame structure for the multi-carrier system may be considered. A signal that has been down-frequently converted based on a photoelectric converter may be transmitted in a specific resource area (e.g., a specific frame). The frequency domain of the specific resource region may include a plurality of chunks. Each chunk may consist of at least one component carrier (CC).

Detailed description of various embodiments of the present disclosure

Hereinafter, various embodiments of the present disclosure will be described in more detail.

This disclosure relates to a method and device used in a wireless communication system incorporating AI (AI radio - AI4C).

Specifically, the present disclosure relates to an apparatus and method for supporting a parallel neural network architecture for cumulative CSI feedback in a wireless communication system incorporating artificial intelligence (AI) (AI for communication, AI4C).

The symbols/abbreviations/terms used in this disclosure are as follows.

- CSI: Channel State Information

- NN: (artificial) Neural Network

- DNN: Deep Neural Network

- CNN: Convolutional (deep) Neural Network

- ResNet: Residual (neural) network

- FDD: Frequency Division Duplex

- UE: User Equipment (Terminal)

- BS: Base Station

- CR: compression ratio

-FC: fully-connected (layer)

- NMSE: normalized mean squared error

- FLOPs: floating point operations

- ABC-Net: “Accumulable feature extraction Before skip Connection” Network

그리고

는 scalar, vector, matrix 그리고 집합을 의미한다.

는 복소수 집합을 나타내고,

은 m by n dimensional complex space를 나타낸다.

는 적절한 차원을 갖는 identity matrix를 의미한다. Superscript

는 Hermitian transpose를 나타낸다.

와

는 각각 trace와 expectation operator를 나타낸다.

은 vector의 Euclidean norm을 나타낸다.

은

을 covariance matrix로 갖는 zero-mean circularly symmetric complex Gaussian distribution을 나타낸다. 임의의 집합

와

에 대하여,

는 집합

와

의 Cartesian product를 나타내며,

은 집합

의 n-ary Cartesian power를 나타낸다.Notations: Lower-case (or upper-case) italic letters represent scalars. Lower-case bold-face letter and upper-case bold-face letter represent vector and matrix, respectively. Calligraphic letter means set. For example,

and

stands for scalar, vector, matrix and set.

represents a set of complex numbers,

represents m by n dimensional complex space.

means an identity matrix with appropriate dimensions. Superscript

represents Hermitian transpose.

and

represents trace and expectation operator, respectively.

represents the Euclidean norm of the vector.

silver

It represents a zero-mean circularly symmetric complex Gaussian distribution with as the covariance matrix. random set

and

about,

is a set

and

It represents the Cartesian product of

silver set

It represents the n-ary Cartesian power of .

Background on various embodiments of the present disclosure

FIG. 20 is a diagram illustrating an example of a CSI network structure in a system applicable to the present disclosure.

Specifically, Figure 20 shows an example of CSI network architecture: CsiNet. (Source: Wen, Chao-Kai, Wan-Ting Shih, and Shi Jin. “Deep learning for massive MIMO CSI feedback.” IEEE Wireless Communications Letters 7.5 (2018): 748-751.)

In this disclosure, an artificial neural network that compresses and reconstructs channel state information (CSI) based on deep learning (DL) is defined as a “CSI network.” Recently, various evolutions have been made regarding the architecture of the CSI network. Figure 20 shows CsiNet [Wen, Chao-Kai, Wan-Ting Shih, and Shi Jin. “Deep learning for massive MIMO CSI feedback.” IEEE Wireless Communications Letters 7.5 (2018): 748-751.)].

A CSI network can generally be viewed as consisting of a CSI encoder and a CSI decoder, as shown in Figure 20. For example, in a downlink system where data transmission is from BS to UE, BS can be a transmitter (Tx) and UE can be a receiver (Rx). In a downlink system, the CSI encoder may exist in the UE, which is Rx, and the CSI decoder may exist in the BS, which is Tx. In this disclosure, a downlink system is assumed for convenience of explanation, but the content of the invention is not limited to the downlink system. For example, the same can be applied to the uplink system.

The CSI encoder present in the UE can compress information about the channel state. The compressed information that is the output of the CSI encoder is delivered to the BS through uplink feedback, and the BS inputs the received compressed information into the CSI decoder, and the CSI decoder can restore information about the UE's channel status. In this disclosure, for convenience of explanation, the compressed information that is the output of the CSI encoder and the input of the CSI decoder will be called a “CSI feedback signal.” In this disclosure, the case where the CSI feedback signal is in the form of a bit stream is considered. Bit stream refers to a sequence composed of binary digits (bits) of 0 or 1 rather than a vector composed of floating point numbers.

In this disclosure, it is assumed that the number of transmit antennas of the BS is N _t , and for convenience, the number of receive antennas of the UE is assumed to be one. The content of the present invention is not limited to the single receive antenna case, but can be simply expanded and applied to the multi-antenna case. Additionally, an OFDM system using N _c orthogonal subcarriers is assumed.

The signal received by the UE through the nth subcarrier can be expressed as Equation 1 below.

,

,

, 그리고

은 각각 frequency domain에서의 instantaneous channel vector, precoding vector, downlink로 전송되는 data symbol, 그리고 AWGN을 나타내며, 모두 n번째 subcarrier에 해당한다. 따라서, n번째 subcarrier에 대한 channel vector인 h_n은 UE에서 추정되어 BS로 feedback되어야 할 수 있다. 전체적으로 모든 subcarriers를 고려하면

로 표현될 수 있는 CSI matrix가 UE로부터 BS로 적절히 feedback되어야만 BS가 precoding vectors를 올바르게 결정할 수 있다. here,

,

,

, and

represents the instantaneous channel vector, precoding vector, data symbol transmitted downlink, and AWGN in the frequency domain, respectively, and all correspond to the nth subcarrier. Therefore, h _n , the channel vector for the nth subcarrier, may need to be estimated at the UE and fed back to the BS. Considering all subcarriers as a whole,

The CSI matrix, which can be expressed as , must be properly fed back from the UE to the BS so that the BS can correctly determine the precoding vectors.

FIG. 21 is a diagram illustrating an example of a CSI matrix preprocessing process for a CSI network in a system applicable to the present disclosure.

Specifically, Figure 21 shows CSI matrix preprocessing for CSI network. (Source: Cao, Zheng, et al. “Lightweight convolutional neural networks for CSI feedback in massive MIMO.” IEEE Communications Letters 25.8 (2021): 2624-2628.)

As shown in Figure 21, H, the CSI matrix in the spatial-frequency domain, is composed of two-dimensional discrete Fourier transform (2D-DFT), truncation to the delay axis in the angular-delay domain, and separation into real and imaginary parts. Preprocessing may be necessary.

In other words, in order to utilize the CSI network, the following three steps of preprocessing can be performed as shown in FIG. 21. It should be noted that in Figure 21, the length and width are reversed from the way a matrix is generally represented.

(1) 2D-DFT

와

는 두 가지 DFT 행렬이 된다. CSI matrix H' in the angular-delay domain can be obtained from CSI matrix H in the spatial-frequency domain, and the relational expression is H'=F _d HF _a . here

and

becomes two DFT matrices.

(2) Truncation with respect to delay-axis

은 오직 첫

rows(행들)에서만 큰 값을 갖고, 나머지 부분에서는 0에 가까운 값을 갖는다. 따라서, angular-delay domain에서의 CSI matrix

의 처음

개의 행들(rows)만을 취하여

을 얻는다.Since the time delay between multipath arrivals exists within a limited period, the time delays for all subcarriers are placed within a certain period. Therefore, the CSI matrix in the angular-delay domain

is only the first

It has large values only in rows, and has values close to 0 in the rest. Therefore, the CSI matrix in the angular-delay domain

the beginning of

By taking only rows

get

(3) Split into real & imaginary part

는 행렬의 각 원소가 complex number로 이루어져 있으나, 일반적인 NN에서는 complex number를 취급하기 어렵다. 따라서, NN 처리의 편의를 위하여, 각 원소의 실수부와 허수부를 나누는 방식으로 두 행렬들을 만들고, 두 행렬들을 3번째 차원으로 쌓아

의 size를 갖는 tensor를 구성한다.Truncated CSI matrix

Each element of the matrix consists of a complex number, but it is difficult to handle complex numbers in a general NN. Therefore, for the convenience of NN processing, two matrices are created by dividing the real and imaginary parts of each element, and the two matrices are stacked in the third dimension.

Construct a tensor with a size of

FIG. 22 is a diagram illustrating an example of comparing various CSI network architectures in a system applicable to the present disclosure.

FIG. 23 is a diagram illustrating an example of source information for various CSI network architectures in a system applicable to the present disclosure.

Specifically, Figure 22 summarizes and compares the characteristics of major CSI network architectures. The source of the CSI network architectures included in FIG. 22 (documents in which each structure is proposed) can be confirmed through FIG. 23. All CSI network architectures in Figure 22 can be seen as utilizing the ResNet structure (ResNet-like architecture).

FIG. 24 is a diagram illustrating an example of a building block of Residual Block: ResNet in a system applicable to the present disclosure.

(Source: He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.)

Specifically, Figure 24 shows a residual block, which is a building block that constitutes the ResNet structure. The fact that a certain NN architecture utilizes the ResNet structure means that the residual block shown in Figure 24 is included in the entire NN structure. A characteristic of residual blocks is a data flow called skip connection or identity shortcut connection. Skip connection refers to a form of skipping without going through several layers, where the so-called identity signal of a path directly connected after the layers is added to the signal that passed through the layers.

FIG. 25 is a diagram illustrating an example of the effect of skip connection in residual block in a system applicable to the present disclosure.

(Source: https://github.com/ndb796/Deep-Learning-Paper-Review-and-Practice/blob/master/lecture_notes/ResNet.pdf)

라는 신호를 입력으로 받아

라는 신호를 출력하도록 학습해야 했다면, skip connection이 존재할 경우

라는 신호가 출력에 그대로 더해지므로 상기 layers이

로 표현될 수 있는 residual 신호를 출력할 수 있도록 학습하기만 하면 동등한 효과를 얻을 수 있다.

전부가 아닌 residual에 해당하는

만큼만을 학습하면 되기 때문에 학습이 수월해지는 것으로 생각해볼 수 있다. ResNet 구조(ResNet-like architecture)에서는, backpropagation (역전파) 과정에서 skip connection을 통해 gradient가 전파될 수 있기 때문에, 다중으로 쌓인 layers에서 생길 수 있는 vanishing gradient problem을 피할 수 있다. 즉, ResNet 구조가 vanishing gradient problem을 극복할 수 있기 때문에 학습이 수월해지는 것으로 볼 수 있다.Specifically, Figure 25 shows the effect of skip connection on residual blocks. When a skip connection does not exist, the layers corresponding to the residual block are

Receive the signal as input

If a skip connection exists, it would have to learn to output a signal like

Since the signal is added to the output as is, the above layers

You can achieve the same effect by simply learning to output a residual signal that can be expressed as .

Corresponds to residual, not all

You can think of learning as becoming easier because you only need to learn as much. In the ResNet-like architecture, the gradient can be propagated through skip connections during the backpropagation process, so the vanishing gradient problem that can occur in multiple stacked layers can be avoided. In other words, it can be seen that learning becomes easier because the ResNet structure can overcome the vanishing gradient problem.

All CSI network architectures in Figure 22 can be seen as utilizing the ResNet structure (ResNet-like architecture), and the decoder of all CSI networks in Figure 22 includes a type of residual block. Encoders may also include blocks such as residual blocks. In fact, it can be seen in Figure 22 that many CSI network architectures utilize the ResNet structure (ResNet-like architecture) in the encoder.

FIG. 26 is a diagram illustrating an example of a CSI network architecture: ACRNet in a system applicable to the present disclosure.

(Source: Lu, Zhilin, et al. "Binarized aggregated network with quantization: Flexible deep learning deployment for CSI feedback in massive MIMO system." IEEE Transactions on Wireless Communications (2022).)

Specifically, Figure 26 shows ACRNet [Lu, Zhilin, et al. “Binarized aggregated network with quantization: Flexible deep learning deployment for CSI feedback in massive MIMO system.” IEEE Transactions on Wireless Communications (2022).]. ACRNet is a current state-of-the-art CSI network architecture. (a) in Figure 26 corresponds to the encoder, and (b) corresponds to the decoder. Looking at Figure 26, it can be seen that ACRNet also includes a structure called ACREnBlock, a type of residual block, in the encoder as well as the decoder.

As shown in Figure 22, existing CSI network architectures are generally unable to support variable CR (compression ratio) as a single NN model, so when the feedback rate (or feedback overhead) changes depending on the system environment, it must be adjusted according to the feedback rate. The NN model for the encoder and decoder had to be changed and used.

의 역수로 정의된다.In literature related to existing CSI network architectures, CR is generally expressed in Equation 2 below:

It is defined as the reciprocal of .

은 특정한 CSI network의 encoder가 출력한 feature vector인 CSI feedback signal의 dimension이다따라서, 상기 CSI network의 encoder는

개의 real numbers (

) 를 M개의 real numbers (CSI feedback signal)로 압축하는 것으로 볼 수 있다. 본 개시에서는 1개의 real number를 32-bit floating-point number로 취급하여 설명한다.here,

is the dimension of the CSI feedback signal, which is a feature vector output by the encoder of a specific CSI network. Therefore, the encoder of the CSI network is

real numbers (

) can be seen as compressing M real numbers (CSI feedback signal). In this disclosure, one real number is treated as a 32-bit floating-point number.

In this disclosure, the operation of the part where the feature vector is output from the encoder NN of the CSI network is defined as “feature extraction.”

를 feedback overhead의 performance indicator로 간주하는 것은 적절하지 못하다.However, in digital communication systems, uplink feedback is generally digital feedback that can only be transmitted in the form of a bit stream. Therefore, for practical deployment of a CSI network, an additional feature that can convert the feature vector composed of real numbers into the form of a bit stream is required. A procedure is needed. Basically, quantization can be considered as a way to change a feature vector composed of real numbers into a bit stream. If the M-dimensional feature vector is transmitted as is from the UE to the BS in 32-bit floating-point format, the feedback overhead is too large to be tolerated by the system. Therefore, in comparing multiple CSI network architectures, simply CR or

It is not appropriate to consider as a performance indicator of feedback overhead.

는 아래 수학식 3과 같이 계산될 수 있다.For example, when B-bit uniform quantization is applied to each real-number element of a feature vector, feedback overhead is reduced by 32?B times. The number of feedback bits

Can be calculated as in Equation 3 below.

를 사용한다. 도 22에 언급된 architectures를 비롯한 기존의 CSI networks 중 대부분은 CR 또는

에 따라 다른 NN model을 사용해야만 했기 때문에, 시스템 환경에 따라 feedback bits의 수

를 바꿔야 하는 상황에서는 다수의 NN models이 필요하다.In this disclosure, as a performance indicator of feedback overhead, the number of feedback bits

Use . Most of the existing CSI networks, including the architectures mentioned in Figure 22, are CR or

Because different NN models had to be used depending on the system environment, the number of feedback bits

In situations where NN models need to be changed, multiple NN models are needed.

The CsiNet+ (CsiNetPlus) paper proposed CSI network architectures that can support Variable CR [Guo, Jiajia, et al. “Convolutional neural network-based multiple-rate compressive sensing for massive MIMO CSI feedback: Design, simulation, and analysis.” [IEEE Transactions on Wireless Communications 19.4 (2020): 2827-2840.], the proposed CSI network architectures, SM-CsiNet+ and PM-CsiNet+, can use the same encoder NN model for different CRs, but still have different encoder NN models for each CR. You must use the decoder NN model. As such, existing CSI network architectures must use different NN models for different CRs, so if the feedback rate needs to change depending on the environment, the model (parameter set) of the CSI network must be changed and used to suit the feedback rate.

For example, the feedback rate may change depending on the coherence time of the channel. The feedback rate may need to be adjusted differently depending on the environment. In existing CSI network architectures, the model of the CSI network must be changed depending on the feedback rate, so the UE and BS must store multiple models (parameter sets). However, because storage space in UE and BS is a finite resource, a CSI network architecture that can support variable feedback rate through a single model (parameter set) is needed.

FIG. 27 is a diagram illustrating an example of CSI feedback bit streams accumulable in decoder in a system applicable to the present disclosure.

In this disclosure, for convenience of explanation, the compressed information that is the output of the CSI encoder and the input of the CSI decoder is called a “CSI feedback signal.” In this disclosure, the case where the CSI feedback signal is in the form of a bit stream is considered. Bit stream refers to a sequence composed of binary digits (bits) of 0 or 1 rather than a vector composed of floating point numbers. Therefore, in the present invention, “CSI feedback bit stream” is considered as the output of the encoder and the input of the decoder. However, the content of the present invention is not limited to signals in the form of bit streams. To-Be shown on the right side of Figure 27 shows the situation in which different CSI feedback bit streams are added before being input to the decoder.

In the present disclosure, “added” in expressions for CSI feedback signals such as “added before being input to the decoder NN,” “added and input to the decoder NN,” or “added to and input to the decoder NN.” The technique encompasses not only summation, but also weighted sum and weighted average.

As-Is shown on the left side of Figure 27 is a simplified diagram of existing general CSI network architectures. The blue trapezoid represents the encoder NN, the green trapezoid represents the decoder NN, and the red rectangle between the encoder NN and decoder NN represents the CSI feedback signal (bit stream). In the As-Is shown on the left side of Figure 27, three cases are expressed according to the number of feedback bits. In Figure 27, not only is the vertical length of the red rectangle representing the CSI feedback bit stream expressed differently depending on the number of transmitted feedback bits, but also the length of the right side of the blue trapezoid representing the encoder NN and the left side of the green trapezoid representing the decoder NN. It is also expressed differently. The reason is that, in general, in existing CSI network architectures, the dimensions of the output of the encoder NN and the input of the decoder NN vary depending on the feedback rate, so the structures of the encoder NN and decoder NN themselves may vary.

On the other hand, To-Be shown on the right side of Figure 27 is a previously proposed CSI feedback method. In the previously proposed CSI feedback method, different CSI feedback bit streams can be added to each other before being input to the decoder NN of the CSI network. Therefore, the input dimension of the decoder NN can be maintained as is, the structure of the decoder NN can be maintained as is, and furthermore, the model (parameter set) of the decoder NN can also be maintained as is.

To-Be expressed on the right side of Figure 27 indicates that the same decoder NN model can always be used regardless of the number of CSI feedback bit streams added before being input to the decoder NN. When added to the Decoder NN, the number of feedback bits increases in proportion to the number of input CSI feedback bit streams. For example, if the length of the CSI feedback bit stream that can be individually input to the decoder NN is 256 bits, as the number of CSI feedback bit streams becomes 2, 3, and 4, the number of feedback bits is 512, 768, and It increases to 1024.

Meanwhile, even if the same decoder NN model is used, CSI reconstruction performance can improve as the number of CSI feedback bit streams added to the decoder NN and input increases. In Figure 27, the improvement in CSI reconstruction performance as the number of CSI feedback bit streams increases is metaphorically expressed as the image of Lenna (Lena Forsen) being restored with higher resolution. However, what is expressed as an image of Lenna is strictly a metaphor, and in reality, information such as the CSI matrix that can be restored from the CSI network is difficult to recognize with the human eye like an image. Expressing the increase in CSI reconstruction performance as an increase in image resolution is also nothing more than a metaphor, and it is difficult to confirm that CSI reconstruction performance is improving in terms of image resolution. Due to the nature of deep learning, it is difficult to explain exactly what the different CSI feedback bit streams added to the decoder NN mean and what role they play.

In the To-Be shown on the right side of Figure 27, CSI feedback bit streams playing different roles added before being input to the decoder NN are expressed in different colors. The red CSI feedback bit stream may be a signal that allows CSI reconstruction even if it is independently input to the decoder NN. The orange CSI feedback bit stream must be added to the red CSI feedback bit stream and input to the decoder NN to be a signal for CSI reconstruction. Considering that CSI feedback bit streams of different roles can be added and input to the decoder NN as shown in To-Be of Figure 25, CSI feedback bit streams of different roles can also be expressed as “different level” CSI feedback bit streams.

FIG. 28 is a diagram illustrating an example of feature extraction before skip connection in encoder in a system applicable to the present disclosure.

As shown in the right To-Be of Figure 27, the previously proposed ABC-Net (“Accumulable feature extraction Before skip Connection” Network). On the right side of Figure 28, the structural characteristics of the encoder NN that constitutes ABC-Net are shown.

In the As-Is shown on the left of FIG. 28, only one CSI feedback bit stream is output, and the CSI feedback bit stream appears in red. In the encoder of a general CSI network, as shown in As-Is of Figure 28, when the feedback rate changes, the output dimension of the encoder NN may change, so the structure of the encoder NN itself may change. Even if the structure of the encoder NN does not change depending on the feedback rate, it is common that at least the model (parameter set) of the encoder NN inevitably changes depending on the feedback rate. The reason is that in general CSI networks, a specific encoder NN model outputs only a CSI feedback signal for a specific fixed feedback rate corresponding to the model.

ABC-Net, shown on the right side of Figure 28, represents the encoder NN structure that outputs CSI feedback bit streams of different roles proposed previously. Considering that CSI feedback bit streams of different roles can be added and input to the decoder NN as shown in To-Be of Figure 27, CSI feedback bit streams of different roles can also be expressed as “different level” CSI feedback bit streams.

As shown in Figure 28, different levels of CSI feedback signals can be obtained from ResNet-like architecture blocks at different positions.

In Figures 27 and 28, different levels of CSI feedback signals are expressed in different colors. In the proposed encoder NN structure of ABC-Net, in order to output different levels of CSI feedback bit streams, the signal immediately before the skip connection of the ResNet structure is input to the fully-connected (FC) layer and the CSI feedback bit stream is output. In other words, it can be said that a feature of the ABC-Net structure is that feature extraction is performed using the identity signal and the residual signal immediately before being added. In other words, ABC-Net has the characteristic of “feature extraction before skip connection”. However, the feature vector extracted before skip connection is an accumulable signal in the decoder. Therefore, it can be seen that accumulable features are extracted. In other words, ABC-Net has the characteristic of “Accumulable feature extraction before skip connection”.

One of the features of the previously proposed ABC-Net is that the encoder NN performs feature extraction using the residual signal before skip connection. The above feature can be seen as creating different levels of CSI feedback bit streams that can be added before being input to the decoder NN. In this disclosure, different levels of CSI feedback signals that can be added before being input to the decoder NN will be referred to as “accumulable feedback signals.”

As an example of the encoder NN structure of a CSI network that generates accumulable feedback signals, a NN structure different from the previously proposed method is possible. As shown in Figure 28, in the encoder NN structure in which ResNet blocks are connected in series, the previous layers of the encoder NN must be operated in order to generate feedback signals output later among CSI feedback signals of different levels, so complexity (e.g. There is a problem that computing resources, memory) are not flexible. In training the Encoder NN, in order to learn the layer related to the CSI feedback signal output earlier, the layer related to the feedback signal output after the CSI feedback signal must be trained together. In order to secure flexibility for the NN model in both training and inference aspects, an encoder NN structure that generates accumulable CSI feedback signals and is different from the method proposed in the existing patent [LG-REF: 22ASL818PC01] is needed.

Therefore, in this disclosure, we propose a new encoder NN structure for accumulable CSI feedback.

Configuration of various embodiments of the present disclosure

In this disclosure, we propose a structure of a parallel encoder NN that generates accumulable CSI feedback signals and is parallel, unlike the previously proposed serial method. Specifically, the present disclosure proposes a parallel encoder NN structure that outputs different levels of CSI feedback signals.

FIG. 29 is a diagram illustrating an example of a proposed parallel encoder NN architecture in a system applicable to the present disclosure.

In this disclosure, we propose a structure of an encoder NN that has parallel characteristics while outputting CSI feedback bit streams of different roles that can be input in addition to the decoder NN.

As-Is shown on the left of Figure 29 briefly shows the encoder structure of a general CSI network. For example, ACRNet [Lu, Zhilin, et al. “Binarized aggregated network with quantization: Flexible deep learning deployment for CSI feedback in massive MIMO system.” An example is the encoder structure of [IEEE Transactions on Wireless Communications (2022).]. The block indicated by a dotted line in FIG. 29 represents the ResNet structure (ResNet-like architecture) explained through FIG. 24 or FIG. 25. Examples of the block indicated by the dotted line in FIG. 30 include the ACREnBlock of FIG. 26, the JC-ResNet Block in the encoder of JC-ResNet listed in FIG. 22, the Encoder Head variant C in BCsiNet, and the encoder structure of CRNet. Some examples include: In this disclosure, the ResNet-like architecture of the same type as exemplified above is collectively referred to as a ResNet block for convenience.

The light blue (light blue) block present in the ResNet block of FIG. 29 may be one layer, but may generally be multiple layers, and the layers may generally be convolutional layers. The dark blue (dark blue) block that exists outside the ResNet block in Figure 29 may represent a fully-connected (FC) layer.

를 출력할 수 있도록, FC layer의 activation function으로

을 적용하는 방법이 제시되어 있다. Sign function은 signum function이라고도 불리며, 아래 수학식 4와 같이 정의된다.In general, the output of the FC layer can be viewed as a vector made up of real numbers. Existing literature [Sohrabi, Foad, Kareem M. Attiah, and Wei Yu. “Deep learning for distributed channel feedback and multiuser precoding in FDD massive MIMO.” [IEEE Transactions on Wireless Communications 20.7 (2021): 4044-4057.] shows a bipolar vector equivalent to the bit stream of B bits as the output of the FC layer.

With the activation function of the FC layer, to output

A method of applying is presented. Sign function is also called signum function and is defined as Equation 4 below.

In order for a bit stream to be output from the encoder NN and input to the decoder NN in the form of a CSI feedback signal, existing literature [Sohrabi, Foad, Kareem M. Attiah, and Wei Yu. “Deep learning for distributed channel feedback and multiuser precoding in FDD massive MIMO.” The method presented in [IEEE Transactions on Wireless Communications 20.7 (2021): 4044-4057.] can be used.

In As-Is shown on the left side of Figure 29, only one CSI feedback bit stream is output. In the encoder of a general CSI network, as shown in As-Is of Figure 29, when the feedback rate changes, the output dimension of the encoder NN may change, so the structure of the encoder NN itself may change. Even if the structure of the encoder NN does not change depending on the feedback rate, it is common that at least the model (parameter set) of the encoder NN inevitably changes depending on the feedback rate. The reason is that in general CSI networks, a specific encoder NN model outputs only a CSI feedback signal for a specific fixed feedback rate corresponding to the model.

To-Be (proposed) shown on the right side of FIG. 29 represents the encoder NN structure that outputs CSI feedback bit streams of different roles proposed in this disclosure. Considering that CSI feedback bit streams of different roles can be added and input to the decoder NN as shown in To-Be of Figure 27, CSI feedback bit streams of different roles can also be expressed as “different level” CSI feedback bit streams.

As in To-Be of Figure 29, each of the different levels of CSI feedback signals can be obtained from each of the ResNet blocks at different positions. In Figure 29, different levels of CSI feedback signals are expressed in different colors. In the proposed encoder NN structure, multiple ResNet blocks can be located in parallel with each other to output different levels of CSI feedback bit streams. The ResNet blocks located in parallel can commonly receive signals output from the previous layer of the encoder NN. For convenience of explanation, the ResNet block is used as an example as a unit structure located in parallel, but unit structures other than ResNet block can be located in parallel. In other words, the scope of the present invention is not limited to ResNet block. For example, single or multiple convolution layer(s) may be located in parallel as a unit structure.

The characteristic of the encoder NN proposed in this disclosure is that the unit structure (e.g., ResNet block) constituting the encoder NN receives common signals as input and is located in parallel. The above feature can be seen as creating different levels of CSI feedback bit streams that can be added before being input to the decoder NN.

It should be noted that there is a part omitted as “...” in To-Be of Figure 29. That is, depending on the embodiment, the encoder NN may include one or more additional ResNet blocks in parallel.

Hereinafter, various embodiments having the above-described characteristics will be presented.

FIG. 30 is a diagram illustrating an example where the total number of bit streams is 2 in a system applicable to the present disclosure.

Specifically, Figure 30 shows a possible embodiment for a case where there are a total of two CSI feedback bit streams transmitted from the UE to the BS through uplink feedback. The CSI feedback bit stream expressed in red is a signal that allows CSI restoration even if it is input alone to the decoder NN. On the other hand, the CSI feedback bit stream expressed in orange must be added to the red signal and input to the decoder NN for CSI restoration to be possible.

In Figure 30, as in Figure 27, to aid understanding, the improvement in CSI reconstruction performance as the number of CSI feedback bit streams increases is metaphorically expressed as the image of Lenna (Lena Forsen) being restored with higher resolution. However, as in Figure 29, it is strictly a metaphor. According to various embodiments of the present disclosure, different types of images may be applied in various ways.

When only one CSI feedback bit stream is transmitted alone from the UE to the BS, the number of feedback bits may be 512. When both CSI feedback bit streams are transmitted from the UE to the BS, the number of feedback bits may be 1024. When two CSI feedback bit streams are added and input to the decoder NN, CSI reconstruction performance may be better than when only one CSI feedback bit stream is input to the decoder NN alone.

CSI feedback bit streams of different levels can all be output from the same encoder NN model (parameter set). The same decoder NN model (parameter set) can be used when only one CSI feedback bit stream is input to the decoder NN and when two different CSI feedback bit streams are added and input to the decoder. That is, the same encoder NN model and decoder NN model can always be used regardless of the number of CSI feedback bit streams transmitted from the UE to the BS.

FIG. 31 is a diagram illustrating an example of a case where a convolution operation in CNN is applied in a system applicable to the present disclosure.

Figure 31 shows the convolution operation in a general CNN. In a convolutional layer, three axes can be considered: height, width, and depth. The most common convolution operation in CNN moves weights (parameters), which can be considered filters, to the height and width axes and multiplies each element of the input, adding one value as output. Since CNN is mainly used in the image processing field, there may be input corresponding to three RGB channels. In this disclosure, to avoid confusion with wireless channels, the term channel corresponding to RGB is replaced with the term depth channel in convolution operations. In other words, the number of depth channels is depth. In general convolution, no matter how large the depth of the input is, an output with only one depth channel is obtained for all depth channels of the input. Therefore, in order to increase the depth of the output, the number of filters must be increased by the depth of the output.

FIG. 32 is a diagram illustrating an example of a depth-wise convolution operation applied in a system applicable to the present disclosure.

Unlike the embodiment of FIG. 31, in depth-wise convolution, an output is obtained separately for each depth channel of the input, so the number of parameters required to obtain the same number of depth channels for the output may be smaller than in a general convolution. Figure 32 shows depth-wise (separable) convolution.

FIG. 33 is a diagram illustrating an example of a group convolution operation applied in a system applicable to the present disclosure.

Figure 33 shows group convolution. Group convolution refers to dividing the depth channels of a given input into multiple groups and performing convolution separately for each group.

FIG. 34 is a diagram illustrating an example of a group convolution operation applied in a system applicable to the present disclosure.

Figure 34 shows an example where the unit structure located in parallel in the encoder NN is a convolutional layer from the perspective of the depth axis. Placing convolutional layers in parallel from the perspective of the depth axis can be seen as applying group convolution. As-Is in Figure 34 shows the encoder NN structure of a typical existing CSI network. In the existing general CSI network, group convolution or depth-wise convolution on the encoder side was not considered, and therefore, a CSI feedback stream for a fixed feedback rate is generally output through a single NN model, and a variable feedback rate is generally output. Multiple NN models are required to support it. On the other hand, in the embodiment of the proposed invention, such as To-Be in FIG. 34, CSI feedback streams of different levels are output through a convolution operation only between depth channels of each group, and some or all of the CSI feedback streams of the different levels are output. Since can be input by adding weighted sum on the decoder side, variable feedback rate can be supported through the same NN model (parameter set).

FIG. 35A is a diagram illustrating an example of a neural network structure in a system applicable to the present disclosure.

Figure 35b is a diagram showing an example of a neural network structure in a system applicable to the present disclosure.

FIG. 35C is a diagram illustrating an example of a neural network structure in a system applicable to the present disclosure.

FIG. 35D is a diagram illustrating an example of a neural network structure in a system applicable to the present disclosure.

Figures 35A to 35D show detailed examples of neural network structures for the structure proposed in this disclosure.

Specifically, Figure 35a shows the existing literature [Lu, Zhilin, et al. “Binarized aggregated network with quantization: Flexible deep learning deployment for CSI feedback in massive MIMO system.” shows ACRNet, a state-of-the-art CSI network architecture proposed in [IEEE Transactions on Wireless Communications (2022).]

Specifically, Figure 35b shows the structure of ACRNet-bipolar, a CSI network architecture that is a partial modification of ACRNet.

Specifically, Figure 35c shows the structure of the previously proposed ABC-Net.

Specifically, Figure 35d shows the structure of the parallel encoder NN architecture proposed in this disclosure.

=1/4인 ACRNet-1×이며, B=2인 uniform quantization이 적용된다. 2-bit uniform quantization이 적용된

=1/4의 ACRNet-1×이 비교 대상으로 표현된 이유는 기존 문헌 [Lu, Zhilin, et al. "Binarized aggregated network with quantization: Flexible deep learning deployment for CSI feedback in massive MIMO system." IEEE Transactions on Wireless Communications (2022).]에 성능이 보고된 ACRNet 구조를 이용한 feedback 방식들 중 유일하게 feedback bits 수

가 1024인 방법이기 때문이다.The ACRNet structure expressed in Figure 35a is

=1/4 is ACRNet-1×, and uniform quantization with B=2 is applied. 2-bit uniform quantization applied

The reason why ACRNet-1× of =1/4 was expressed as the comparison target is because of the existing literature [Lu, Zhilin, et al. “Binarized aggregated network with quantization: Flexible deep learning deployment for CSI feedback in massive MIMO system.” Among the feedback methods using the ACRNet structure whose performance was reported in [IEEE Transactions on Wireless Communications (2022).], it is the only one with a high number of feedback bits.

This is because it is a method where is 1024.

In the proposed structure of Figure 35d, a CSI feedback signal in the form of a bit stream can be output directly from the encoder NN. However, in the state-of-the-art ACRNet, the output of the encoder NN is a feature vector composed of real numbers, so for comparison, you can look at the performance of applying uniform quantization to ACRNet. However, because applying uniform quantization is known to be sub-optimal, it cannot be said to be an equal comparison from the perspective of CSI reconstruction performance, and the index of complexity (amount of computation and storage space) for quantization is the complexity (amount of computation and storage) for NN. Since it may be somewhat different from the indicator of (amount of space), it is difficult to calculate the complexity (amount of computation and storage space) expressed by different indicators by integrating them.

을 적용한 구조이다. FC layer에 상기 변화를 적용한 것을 본 개시에서는 bipolar layer라 표현한다. FC layer를 bipolar layer로 변경한 것 외의 다른 모든 구조는 ACRNet과 동일하다. 성능 비교를 위한 실험에 사용된 ACRNet-bipolar는 실험에 사용된 ACRNet-1×을 기반으로 변형된다.Therefore, we present ACRNet-bipolar, a CSI network architecture that modifies ACRNet so that the bit stream is directly output from the encoder NN and the bit stream output from the encoder NN can be directly input to the decoder NN. ACRNet-bipolar is a structure designed to compare the performance of the present invention. In the existing ACRNet, the output dimension of the FC layer that exists at the end of the encoder and the input dimension of the FC layer that exists at the beginning of the decoder are matched to the number of feedback bits, and the existing ACRNet Sohrabi, Foad, Kareem M. Attiah, and Wei Yu. “Deep learning for distributed channel feedback and multiuser precoding in FDD massive MIMO.” As an activation function of the FC layer that exists at the end of the encoder, as shown in [IEEE Transactions on Wireless Communications 20.7 (2021): 4044-4057.]

This is a structure that applies . The application of the above changes to the FC layer is referred to as a bipolar layer in this disclosure. Other than changing the FC layer to a bipolar layer, all other structures are the same as ACRNet. ACRNet-bipolar used in the experiment for performance comparison is modified based on ACRNet-1× used in the experiment.

In the embodiment of Figure 35d, there are a total of two CSI feedback bit streams transmitted through the proposed parallel architecture, and one CSI feedback bit stream consists of 512 bits. Therefore, the proposed structure of Figure 35d can support two numbers of feedback bits: 512 or 1024. Therefore, it can be compared with ACRNet with uniform quantization applied so that the number of feedback bits is 1024, ACRNet-bipolar with 512 feedback bits, and ACRNet-bipolar with 1024 feedback bits. Since the performance of the ACRNet structure with uniform quantization applied so that the number of feedback bits is 512 is not known, it is excluded from the comparison.

In order to support both 512 and 1024 bits of feedback rates using ACRNet-bipolar in Figure 35b, 2.102M+4.200M=6.302M parameters are required, but when using the proposed parallel architecture, 3.151M parameters are required. Therefore, storage space can be saved by 50%.

When the proposed parallel architecture of Figure 35d operates at a feedback rate of 512 bits, only one of the two parallel paths needs to operate, so it can operate with less computational effort than the ACRNet-bipolar of Figure 35b, where the number of feedback bits is 512. possible. When the proposed parallel architecture in Figure 35d operates at a feedback rate of 1024 bits, 5.792M FLOPs are required, which is 84.7% of the 6.840M FLOPs required by ACRNet-bipolar in Figure 35b, or more than 15.3%. It can be seen that there is a computational saving effect. It can be noted that in the case of ACRNet, additional complexity is required for quantization and dequantization operations.

Effects of various embodiments of the present disclosure

Expected effects of various embodiments of the present disclosure are as follows.

(1) Adaptive feedback rate operation according to downlink channel environment and uplink resource situation

Using the proposed parallel architecture, the feedback rate can be operated adaptively by considering downlink channel environment and uplink resource conditions such as coherence time. For example, in a situation where uplink resources are limited, the UE transmits only a CSI feedback bit stream at a level capable of independently decoding, and immediately after receiving the CSI feedback bit stream from the BS (within the coherence time for the downlink channel), the uplink resource If is additionally given, it is possible to operate by sending only the CSI feedback bit stream that will be added, excluding the CSI feedback bit stream already received from the BS, without sending all different CSI feedback bit streams.

(2) Reduction in complexity of NN model (parameter set)

Save required storage space and computing resources

(3) Flexibility for NN model in training and inference operations

In serial architecture, the previous layers of the encoder NN must operate in order to generate feedback signals that are output later among CSI feedback signals of different levels, but in the proposed parallel architecture, the parallel unit structure can operate independently, saving computing resources and memory. Flexible in terms of Therefore, there is an advantage in terms of power consumption.

In serial architecture, in order to train the layer related to the CSI feedback signal output earlier, the layer related to the feedback signal output after the CSI feedback signal must be trained together, but in the proposed parallel architecture, the parallel unit structure is independent. It can be learned.

(4) Save learning time

FIG. 36 is a diagram illustrating an example of various NN architectures for accumulable CSI feedback in a system applicable to the present disclosure.

Specifically, Figure 36 is a diagram showing examples of three different structures capable of accumulable CSI feedback with the same number of parameters and FLOPs.

FIG. 37 is a diagram illustrating an example of the performance and complexity of the proposed parallel architecture in a system applicable to the present disclosure.

Specifically, Figure 36 shows three different structures capable of accumulable CSI feedback with the same number of parameters and FLOPs. Among the three structures, two structures are serial structures and one structure is a parallel structure. The detailed NN structure of ABC-Net (before skip connection) shown in Figure 36 and the proposed parallel architecture are shown in Figures 35c and 35d. The experimental results for the three NN architectures shown in Figure 36 are shown in Figure 37. The experimental environment and settings are the same as before when ABC-Net was proposed. Using the parallel structure, it is possible to show channel reconstruction performance at a level comparable to the serial structure, and it can be seen from Figure 37 that the learning time can be greatly reduced to 1/5 compared to the serial structure.

Characteristic configurations of various embodiments of the present disclosure are as follows.

(1) In a wireless communication system, when a transmitting device (terminal) transmits channel state information (CSI) to a base station using a neural network (NN), the CSI is composed of a plurality of sub-CSI signals of different levels. , The terminal obtains each of the sub-CSI signals from each part of the NN, and the base station restores or estimates the CSI through a signal obtained by adding all or part of the plurality of sub-CSI signals.

(2) A method characterized in that at least some of the sub-CSI signals are obtained in parallel (within the corresponding NN part).

[Description regarding terminal claims]

Hereinafter, the above-described embodiments will be described in detail with reference to FIG. 38 in terms of terminal operation. The methods described below are separated for convenience of explanation, and it goes without saying that, unless mutually exclusive, some components of one method may be replaced with some components of another method, or may be applied in combination with each other.

FIG. 38 is a diagram illustrating an example of an operation process of a terminal (user equipment, UE) in a system applicable to the present disclosure.

According to various embodiments of the present disclosure, a method performed by a terminal in a wireless communication system is provided.

The embodiment of FIG. 38 includes, before step S3801, a terminal receiving one or more synchronization signals from a base station (BS); The terminal may further include receiving system information from the base station. In addition, the embodiment of FIG. 38 includes, before step S3801, the terminal transmitting a random access preamble to the base station; The terminal may further include receiving a random access response from the base station.

In step S3801, the terminal receives a reference signal from a base station (BS).

In step S3802, the terminal generates channel state information (CSI) based on the reference signal.

In step S3803, the terminal generates a plurality of signals for the CSI based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel.

In step S3804, the terminal transmits the plurality of signals to the base station.

According to various embodiments of the present disclosure, the CSI may be restored based on a first signal among the plurality of signals. The CSI may be restored based on the first signal and one or more other signals among the plurality of signals. Restoration of the CSI may not be possible with only one or more signals other than the first signal.

According to various embodiments of the present disclosure, when based on the first signal and at least one other signal, restoration performance of the CSI may be higher than when based on only the first signal.

According to various embodiments of the present disclosure, each of the first signal and the other one or more signals may correspond to each of the plurality of encoder components.

According to various embodiments of the present disclosure, when restoration of the CSI is performed based only on the first signal and when restoration of the CSI is performed based on the first signal and the other one or more signals, all are the same. A decoder NN model may be used.

According to various embodiments of the present disclosure, when restoration of the CSI is performed based on the first signal and the one or more other signals, the first signal and the one or more other signals are obtained through a weighted sum. After being combined into one signal, it can be input to the decoder NN model.

According to various embodiments of the present disclosure, the decoder NN model provides information on the first signal and the other one or more signals according to a CSI feedback rate that varies based on at least one of a channel state and an uplink resource state. It may be configured to adaptively perform the weighted sum.

According to various embodiments of the present disclosure, each of the plurality of encoder components may include one or more convolutional layers. The encoder NN model may be composed of a parameter set for convolutional layers included in the plurality of encoder components.

According to various embodiments of the present disclosure, a terminal is provided in a wireless communication system. The terminal includes a transceiver and at least one processor, and the at least one processor may be configured to perform the terminal operation method according to FIG. 38.

According to various embodiments of the present disclosure, an apparatus for controlling a terminal in a communication system is provided. The device includes at least one processor and at least one memory operably connected to the at least one processor. The at least one memory may be configured to store instructions for performing the operation method of the terminal according to FIG. 38, based on execution by the at least one processor.

According to various embodiments of the present disclosure, one or more non-transitory computer readable medium (CRM) storing one or more instructions is provided. The one or more instructions perform operations based on execution by one or more processors, and the operations may include the method of operating a terminal according to FIG. 38.

[Explanation regarding base station claims]

Hereinafter, the above-described embodiments will be described in detail with reference to FIG. 39 in terms of operation of the base station. The methods described below are separated for convenience of explanation, and it goes without saying that, unless mutually exclusive, some components of one method may be replaced with some components of another method, or may be applied in combination with each other.

FIG. 39 is a diagram illustrating an example of the operation process of a base station (BS) in a system applicable to the present disclosure.

According to various embodiments of the present disclosure, a method performed by a base station (BS) in a wireless communication system is provided.

The embodiment of FIG. 39 includes, before step S3901, a base station transmitting one or more synchronization signals to a user equipment (UE); It may further include the step of the base station transmitting system information to the terminal. In addition, the embodiment of FIG. 39 includes, before step S3901, a base station receiving a random access preamble from the terminal; It may further include the base station transmitting a random access response to the terminal.

In step S3901, the base station transmits a reference signal to the user equipment (UE).

In step S3902, the base station receives a plurality of signals from the terminal. The plurality of signals are related to channel state information (CSI) based on the reference signal. The plurality of signals are generated based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel.

According to various embodiments of the present disclosure, the CSI may be restored based on a first signal among the plurality of signals. The CSI may be restored based on the first signal and one or more other signals among the plurality of signals. Restoration of the CSI may not be possible with only one or more signals other than the first signal.

According to various embodiments of the present disclosure, when based on the first signal and at least one other signal, restoration performance of the CSI may be higher than when based on only the first signal.

According to various embodiments of the present disclosure, each of the first signal and the other one or more signals may correspond to each of the plurality of encoder components.

According to various embodiments of the present disclosure, when restoration of the CSI is performed based only on the first signal and when restoration of the CSI is performed based on the first signal and the other one or more signals, all are the same. A decoder NN model may be used.

According to various embodiments of the present disclosure, when restoration of the CSI is performed based on the first signal and the one or more other signals, the first signal and the one or more other signals are obtained through a weighted sum. After being combined into one signal, it can be input to the decoder NN model.

According to various embodiments of the present disclosure, the decoder NN model provides information on the first signal and the other one or more signals according to a CSI feedback rate that varies based on at least one of a channel state and an uplink resource state. It may be configured to adaptively perform the weighted sum.

According to various embodiments of the present disclosure, each of the plurality of encoder components may include one or more convolutional layers. The encoder NN model may be composed of a parameter set for convolutional layers included in the plurality of encoder components.

According to various embodiments of the present disclosure, a base station (BS) is provided in a wireless communication system. The base station includes a transceiver and at least one processor, and the at least one processor may be configured to perform the method of operating the base station according to FIG. 39.

According to various embodiments of the present disclosure, an apparatus for controlling a base station in a wireless communication system is provided. The device includes at least one processor and at least one memory operably connected to the at least one processor. The at least one memory may be configured to store instructions for performing the operating method of the base station according to FIG. 39, based on execution by the at least one processor.

According to various embodiments of the present disclosure, one or more non-transitory computer readable medium (CRM) storing one or more instructions is provided. The one or more instructions perform operations based on execution by one or more processors, and the operations may include the method of operating a base station according to FIG. 39.

Communication systems applicable to this disclosure

40 illustrates a communication system 1 applied to various embodiments of the present disclosure.

Referring to FIG. 40, the communication system 1 applied to various embodiments of the present disclosure includes a wireless device, a base station, and a network. Here, a wireless device refers to a device that performs communication using wireless access technology (e.g., 5G NR (New RAT), LTE (Long Term Evolution), 6G wireless communication), and includes communication/wireless/5G device/6G device. It may be referred to as . Although not limited thereto, wireless devices include robots (100a), vehicles (100b-1, 100b-2), XR (eXtended Reality) devices (100c), hand-held devices (100d), and home appliances (100e). ), IoT (Internet of Thing) device (100f), and AI device/server (400). For example, vehicles may include vehicles equipped with wireless communication functions, autonomous vehicles, vehicles capable of inter-vehicle communication, etc. Here, the vehicle may include an Unmanned Aerial Vehicle (UAV) (eg, a drone). XR devices include AR (Augmented Reality)/VR (Virtual Reality)/MR (Mixed Reality) devices, HMD (Head-Mounted Device), HUD (Head-Up Display) installed in vehicles, televisions, smartphones, It can be implemented in the form of computers, wearable devices, home appliances, digital signage, vehicles, robots, etc. Portable devices may include smartphones, smart pads, wearable devices (e.g., smartwatches, smart glasses), and computers (e.g., laptops, etc.). Home appliances may include TVs, refrigerators, washing machines, etc. IoT devices may include sensors, smart meters, etc. For example, a base station and network may also be implemented as wireless devices, and a specific wireless device 200a may operate as a base station/network node for other wireless devices.

Wireless devices 100a to 100f may be connected to the network 300 through the base station 200. AI (Artificial Intelligence) technology may be applied to wireless devices (100a to 100f), and the wireless devices (100a to 100f) may be connected to the AI server 400 through the network 300. The network 300 may be configured using a 3G network, 4G (eg, LTE) network, 5G (eg, NR) network, or 6G network. Wireless devices 100a to 100f may communicate with each other through the base station 200/network 300, but may also communicate directly (e.g. sidelink communication) without going through the base station/network. For example, vehicles 100b-1 and 100b-2 may communicate directly (e.g. V2V (Vehicle to Vehicle)/V2X (Vehicle to everything) communication). Additionally, an IoT device (eg, sensor) may communicate directly with another IoT device (eg, sensor) or another wireless device (100a to 100f).

Wireless communication/connection (150a, 150b, 150c) may be established between the wireless devices (100a to 100f)/base station (200) and the base station (200)/base station (200). Here, wireless communication/connection includes various wireless connections such as uplink/downlink communication (150a), sidelink communication (150b) (or D2D communication), and inter-base station communication (150c) (e.g. relay, IAB (Integrated Access Backhaul)). This can be achieved through technology (e.g., 5G NR) through wireless communication/connection (150a, 150b, 150c), where a wireless device and a base station/wireless device, and a base station and a base station can transmit/receive wireless signals to each other. For example, the wireless communication/connection 150a, 150b, and 150c may transmit/receive signals through various physical channels, based on various proposals of various embodiments of the present disclosure. /At least some of various configuration information setting processes for reception, various signal processing processes (e.g., channel encoding/decoding, modulation/demodulation, resource mapping/demapping, etc.), resource allocation processes, etc. may be performed.

Meanwhile, NR supports multiple numerologies (or subcarrier spacing (SCS)) to support various 5G services. For example, if SCS is 15kHz, it supports wide area in traditional cellular bands, and if SCS is 30kHz/60kHz, it supports dense-urban, lower latency. and a wider carrier bandwidth, and when SCS is 60kHz or higher, it supports a bandwidth greater than 24.25GHz to overcome phase noise.

The NR frequency band can be defined as two types of frequency ranges (FR1, FR2). The values of the frequency range may be changed. For example, the frequency ranges of the two types (FR1, FR2) may be as shown in Table 3 below. For convenience of explanation, among the frequency ranges used in the NR system, FR1 may mean "sub 6GHz range", and FR2 may mean "above 6GHz range" and may be called millimeter wave (mmW). .

Frequency Range designation Corresponding frequency range Subcarrier Spacing FR1 450MHz-6000MHz 15, 30, 60kHz FR2 24250MHz-52600MHz 60, 120, 240 kHz

As mentioned above, the numerical value of the frequency range of the NR system can be changed. For example, FR1 may include a band of 410 MHz to 7125 MHz as shown in Table 4 below. That is, FR1 may include a frequency band of 6GHz (or 5850, 5900, 5925 MHz, etc.). For example, the frequency band above 6 GHz (or 5850, 5900, 5925 MHz, etc.) included within FR1 may include an unlicensed band. Unlicensed bands can be used for a variety of purposes, for example, for communications for vehicles (e.g., autonomous driving).

Frequency Range designation Corresponding frequency range Subcarrier Spacing FR1 41MHz-7125MHz 15, 30, 60kHz FR2 24250MHz-52600MHz 60, 120, 240 kHz

Wireless devices applicable to this disclosure

Below, examples of wireless devices to which various embodiments of the present disclosure are applied will be described.

Figure 41 illustrates a wireless device that can be applied to various embodiments of the present disclosure.

Referring to FIG. 41, the first wireless device 100 and the second wireless device 200 can transmit and receive wireless signals through various wireless access technologies (eg, LTE, NR). Here, {first wireless device 100, second wireless device 200} refers to {wireless device 100x, base station 200} and/or {wireless device 100x, wireless device 100x) in FIG. 40. } can be responded to.

The first wireless device 100 includes one or more processors 102 and one or more memories 104, and may additionally include one or more transceivers 106 and/or one or more antennas 108. Processor 102 controls memory 104 and/or transceiver 106 and may be configured to implement the descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed herein. For example, the processor 102 may process information in the memory 104 to generate first information/signal and then transmit a wireless signal including the first information/signal through the transceiver 106. Additionally, the processor 102 may receive a wireless signal including the second information/signal through the transceiver 106 and then store information obtained from signal processing of the second information/signal in the memory 104. The memory 104 may be connected to the processor 102 and may store various information related to the operation of the processor 102. For example, memory 104 may perform some or all of the processes controlled by processor 102 or instructions for performing the descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed herein. Software code containing them can be stored. Here, the processor 102 and memory 104 may be part of a communication modem/circuit/chip designed to implement wireless communication technology (eg, LTE, NR). Transceiver 106 may be coupled to processor 102 and may transmit and/or receive wireless signals via one or more antennas 108. Transceiver 106 may include a transmitter and/or receiver. The transceiver 106 can be used interchangeably with an RF (Radio Frequency) unit. In various embodiments of the present disclosure, a wireless device may mean a communication modem/circuit/chip.

The second wireless device 200 includes one or more processors 202, one or more memories 204, and may further include one or more transceivers 206 and/or one or more antennas 208. Processor 202 controls memory 204 and/or transceiver 206 and may be configured to implement the descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed herein. For example, the processor 202 may process the information in the memory 204 to generate third information/signal and then transmit a wireless signal including the third information/signal through the transceiver 206. Additionally, the processor 202 may receive a wireless signal including the fourth information/signal through the transceiver 206 and then store information obtained from signal processing of the fourth information/signal in the memory 204. The memory 204 may be connected to the processor 202 and may store various information related to the operation of the processor 202. For example, memory 204 may perform some or all of the processes controlled by processor 202 or instructions for performing the descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed herein. Software code containing them can be stored. Here, the processor 202 and memory 204 may be part of a communication modem/circuit/chip designed to implement wireless communication technology (eg, LTE, NR). Transceiver 206 may be coupled to processor 202 and may transmit and/or receive wireless signals via one or more antennas 208. Transceiver 206 may include a transmitter and/or receiver. Transceiver 206 may be used interchangeably with an RF unit. In various embodiments of the present disclosure, a wireless device may mean a communication modem/circuit/chip.

Hereinafter, the hardware elements of the wireless devices 100 and 200 will be described in more detail. Although not limited thereto, one or more protocol layers may be implemented by one or more processors 102, 202. For example, one or more processors 102, 202 may implement one or more layers (e.g., functional layers such as PHY, MAC, RLC, PDCP, RRC, SDAP). One or more processors 102, 202 may generate one or more Protocol Data Units (PDUs) and/or one or more Service Data Units (SDUs) according to the descriptions, functions, procedures, suggestions, methods and/or operational flow charts disclosed herein. can be created. One or more processors 102, 202 may generate messages, control information, data or information according to the descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed herein. One or more processors 102, 202 generate signals (e.g., baseband signals) containing PDUs, SDUs, messages, control information, data or information according to the functions, procedures, suggestions and/or methods disclosed herein. , can be provided to one or more transceivers (106, 206). One or more processors 102, 202 may receive signals (e.g., baseband signals) from one or more transceivers 106, 206, and the descriptions, functions, procedures, suggestions, methods, and/or operational flowcharts disclosed herein. Depending on the device, PDU, SDU, message, control information, data or information can be obtained.

One or more processors 102, 202 may be referred to as a controller, microcontroller, microprocessor, or microcomputer. One or more processors 102, 202 may be implemented by hardware, firmware, software, or a combination thereof. As an example, one or more Application Specific Integrated Circuits (ASICs), one or more Digital Signal Processors (DSPs), one or more Digital Signal Processing Devices (DSPDs), one or more Programmable Logic Devices (PLDs), or one or more Field Programmable Gate Arrays (FPGAs) May be included in one or more processors 102 and 202. The descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed in this document may be implemented using firmware or software, and the firmware or software may be implemented to include modules, procedures, functions, etc. Firmware or software configured to perform the descriptions, functions, procedures, suggestions, methods, and/or operational flowcharts disclosed in this document may be included in one or more processors (102, 202) or stored in one or more memories (104, 204). It may be driven by the above processors 102 and 202. The descriptions, functions, procedures, suggestions, methods and/or operational flowcharts disclosed in this document may be implemented using firmware or software in the form of codes, instructions and/or sets of instructions.

One or more memories 104, 204 may be connected to one or more processors 102, 202 and may store various types of data, signals, messages, information, programs, codes, instructions, and/or instructions. One or more memories 104, 204 may consist of ROM, RAM, EPROM, flash memory, hard drives, registers, cache memory, computer readable storage media, and/or combinations thereof. One or more memories 104, 204 may be located internal to and/or external to one or more processors 102, 202. Additionally, one or more memories 104, 204 may be connected to one or more processors 102, 202 through various technologies, such as wired or wireless connections.

One or more transceivers 106, 206 may transmit user data, control information, wireless signals/channels, etc. mentioned in the methods and/or operation flowcharts of this document to one or more other devices. One or more transceivers 106, 206 may receive user data, control information, wireless signals/channels, etc. referred to in the descriptions, functions, procedures, suggestions, methods and/or operational flow charts, etc. disclosed herein from one or more other devices. there is. For example, one or more transceivers 106 and 206 may be connected to one or more processors 102 and 202 and may transmit and receive wireless signals. For example, one or more processors 102, 202 may control one or more transceivers 106, 206 to transmit user data, control information, or wireless signals to one or more other devices. Additionally, one or more processors 102, 202 may control one or more transceivers 106, 206 to receive user data, control information, or wireless signals from one or more other devices. In addition, one or more transceivers (106, 206) may be connected to one or more antennas (108, 208), and one or more transceivers (106, 206) may be connected to the description and functions disclosed in this document through one or more antennas (108, 208). , may be set to transmit and receive user data, control information, wireless signals/channels, etc. mentioned in procedures, proposals, methods and/or operation flow charts, etc. In this document, one or more antennas may be multiple physical antennas or multiple logical antennas (eg, antenna ports). One or more transceivers (106, 206) process the received user data, control information, wireless signals/channels, etc. using one or more processors (102, 202), and convert the received wireless signals/channels, etc. from the RF band signal. It can be converted to a baseband signal. One or more transceivers (106, 206) may convert user data, control information, wireless signals/channels, etc. processed using one or more processors (102, 202) from baseband signals to RF band signals. For this purpose, one or more transceivers 106, 206 may comprise (analog) oscillators and/or filters.

Figure 42 shows another example of a wireless device that can be applied to various embodiments of the present disclosure.

According to FIG. 42, the wireless device may include at least one processor (102, 202), at least one memory (104, 204), at least one transceiver (106, 206), and one or more antennas (108, 208). there is.

As a difference between the example of the wireless device previously described in FIG. 41 and the example of the wireless device in FIG. 42, in FIG. 41 the processors 102 and 202 and the memories 104 and 204 are separated, but in the example of FIG. 42, the processor The point is that memories (104, 204) are included in (102, 202).

Here, specific descriptions of the processors 102 and 202, memories 104 and 204, transceivers 106 and 206, and one or more antennas 108 and 208 are as described above, so to avoid unnecessary repetition of description, Repeated descriptions should be omitted.

Below, an example of a signal processing circuit to which various embodiments of the present disclosure are applied will be described.

Figure 43 illustrates a signal processing circuit for a transmission signal.

Referring to FIG. 43, the signal processing circuit 1000 may include a scrambler 1010, a modulator 1020, a layer mapper 1030, a precoder 1040, a resource mapper 1050, and a signal generator 1060. there is. Although not limited thereto, the operations/functions of Figure 43 may be performed in the processors 102, 202 and/or transceivers 106, 206 of Figure 41. The hardware elements of Figure 43 may be implemented in the processors 102, 202 and/or transceivers 106, 206 of Figure 41. For example, blocks 1010 to 1060 may be implemented in processors 102 and 202 of FIG. 41. Additionally, blocks 1010 to 1050 may be implemented in the processors 102 and 202 of FIG. 41, and block 1060 may be implemented in the transceivers 106 and 206 of FIG. 41.

The codeword can be converted into a wireless signal through the signal processing circuit 1000 of FIG. 43. Here, a codeword is an encoded bit sequence of an information block. The information block may include a transport block (eg, UL-SCH transport block, DL-SCH transport block). Wireless signals may be transmitted through various physical channels (eg, PUSCH, PDSCH).

Specifically, the codeword may be converted into a scrambled bit sequence by the scrambler 1010. The scramble sequence used for scrambling is generated based on an initialization value, and the initialization value may include ID information of the wireless device. The scrambled bit sequence may be modulated into a modulation symbol sequence by the modulator 1020. Modulation methods may include pi/2-BPSK (pi/2-Binary Phase Shift Keying), m-PSK (m-Phase Shift Keying), m-QAM (m-Quadrature Amplitude Modulation), etc. The complex modulation symbol sequence may be mapped to one or more transport layers by the layer mapper 1030. The modulation symbols of each transport layer may be mapped to the corresponding antenna port(s) by the precoder 1040 (precoding). The output z of the precoder 1040 can be obtained by multiplying the output y of the layer mapper 1030 with the precoding matrix W of N*M. Here, N is the number of antenna ports and M is the number of transport layers. Here, the precoder 1040 may perform precoding after performing transform precoding (eg, DFT transformation) on complex modulation symbols. Additionally, the precoder 1040 may perform precoding without performing transform precoding.

The resource mapper 1050 can map the modulation symbols of each antenna port to time-frequency resources. A time-frequency resource may include a plurality of symbols (e.g., CP-OFDMA symbol, DFT-s-OFDMA symbol) in the time domain and a plurality of subcarriers in the frequency domain. The signal generator 1060 generates a wireless signal from the mapped modulation symbols, and the generated wireless signal can be transmitted to another device through each antenna. For this purpose, the signal generator 1060 may include an Inverse Fast Fourier Transform (IFFT) module, a Cyclic Prefix (CP) inserter, a Digital-to-Analog Converter (DAC), a frequency uplink converter, etc. .

The signal processing process for the received signal in the wireless device may be configured as the reverse of the signal processing process (1010 to 1060) of FIG. 43. For example, a wireless device (eg, 100 and 200 in FIG. 41) may receive a wireless signal from the outside through an antenna port/transceiver. The received wireless signal can be converted into a baseband signal through a signal restorer. To this end, the signal restorer may include a frequency downlink converter, an analog-to-digital converter (ADC), a CP remover, and a Fast Fourier Transform (FFT) module. Afterwards, the baseband signal can be restored to a codeword through a resource de-mapper process, postcoding process, demodulation process, and de-scramble process. The codeword can be restored to the original information block through decoding. Accordingly, a signal processing circuit (not shown) for a received signal may include a signal restorer, resource de-mapper, postcoder, demodulator, de-scrambler, and decoder.

Below, examples of wireless device utilization to which various embodiments of the present disclosure are applied will be described.

Figure 44 shows another example of a wireless device applied to various embodiments of the present disclosure. Wireless devices can be implemented in various forms depending on usage-examples/services (see FIG. 40).

Referring to FIG. 44, the wireless devices 100 and 200 correspond to the wireless devices 100 and 200 of FIG. 41 and include various elements, components, units/units, and/or modules. ) can be composed of. For example, the wireless devices 100 and 200 may include a communication unit 110, a control unit 120, a memory unit 130, and an additional element 140. The communication unit may include communication circuitry 112 and transceiver(s) 114. For example, communication circuitry 112 may include one or more processors 102, 202 and/or one or more memories 104, 204 of FIG. 41. For example, transceiver(s) 114 may include one or more transceivers 106, 206 and/or one or more antennas 108, 208 of FIG. 41. The control unit 120 is electrically connected to the communication unit 110, the memory unit 130, and the additional element 140 and controls overall operations of the wireless device. For example, the control unit 120 may control the electrical/mechanical operation of the wireless device based on the program/code/command/information stored in the memory unit 130. In addition, the control unit 120 transmits the information stored in the memory unit 130 to the outside (e.g., another communication device) through the communication unit 110 through a wireless/wired interface, or to the outside (e.g., to another communication device) through the communication unit 110. Information received through a wireless/wired interface from another communication device may be stored in the memory unit 130.

The additional element 140 may be configured in various ways depending on the type of wireless device. For example, the additional element 140 may include at least one of a power unit/battery, an input/output unit (I/O unit), a driving unit, and a computing unit. Although not limited thereto, wireless devices include robots (FIGS. 40, 100a), vehicles (FIGS. 40, 100b-1, 100b-2), XR devices (FIGS. 40, 100c), portable devices (FIGS. 40, 100d), and home appliances. (FIGS. 40, 100e), IoT devices (FIGS. 40, 100f), digital broadcasting terminals, hologram devices, public safety devices, MTC devices, medical devices, fintech devices (or financial devices), security devices, climate/environment devices, It can be implemented in the form of an AI server/device (FIG. 40, 400), base station (FIG. 40, 200), network node, etc. Wireless devices can be mobile or used in fixed locations depending on the usage/service.

In FIG. 44 , various elements, components, units/parts, and/or modules within the wireless devices 100 and 200 may be entirely interconnected through a wired interface, or at least some of them may be wirelessly connected through the communication unit 110. For example, within the wireless devices 100 and 200, the control unit 120 and the communication unit 110 are connected by wire, and the control unit 120 and the first unit (e.g., 130 and 140) are connected through the communication unit 110. Can be connected wirelessly. Additionally, each element, component, unit/part, and/or module within the wireless devices 100 and 200 may further include one or more elements. For example, the control unit 120 may be comprised of one or more processor sets. For example, the control unit 120 may be comprised of a communication control processor, an application processor, an electronic control unit (ECU), a graphics processing processor, and a memory control processor. As another example, the memory unit 130 includes random access memory (RAM), dynamic RAM (DRAM), read only memory (ROM), flash memory, volatile memory, and non-volatile memory. volatile memory) and/or a combination thereof.

Hereinafter, the implementation example of FIG. 44 will be described in more detail with reference to the drawings.

Figure 45 illustrates a portable device applied to various embodiments of the present disclosure. Portable devices may include smartphones, smart pads, wearable devices (e.g., smartwatches, smartglasses), and portable computers (e.g., laptops, etc.). A mobile device may be referred to as a Mobile Station (MS), user terminal (UT), Mobile Subscriber Station (MSS), Subscriber Station (SS), Advanced Mobile Station (AMS), or Wireless terminal (WT).

Referring to FIG. 45, the portable device 100 includes an antenna unit 108, a communication unit 110, a control unit 120, a memory unit 130, a power supply unit 140a, an interface unit 140b, and an input/output unit 140c. ) may include. The antenna unit 108 may be configured as part of the communication unit 110. Blocks 110 to 130/140a to 140c correspond to blocks 110 to 130/140 in FIG. 44, respectively.

The communication unit 110 can transmit and receive signals (eg, data, control signals, etc.) with other wireless devices and base stations. The control unit 120 can control the components of the portable device 100 to perform various operations. The control unit 120 may include an application processor (AP). The memory unit 130 may store data/parameters/programs/codes/commands necessary for driving the portable device 100. Additionally, the memory unit 130 can store input/output data/information, etc. The power supply unit 140a supplies power to the portable device 100 and may include a wired/wireless charging circuit, a battery, etc. The interface unit 140b may support connection between the mobile device 100 and other external devices. The interface unit 140b may include various ports (eg, audio input/output ports, video input/output ports) for connection to external devices. The input/output unit 140c may input or output image information/signals, audio information/signals, data, and/or information input from the user. The input/output unit 140c may include a camera, a microphone, a user input unit, a display unit 140d, a speaker, and/or a haptic module.

For example, in the case of data communication, the input/output unit 140c acquires information/signals (e.g., touch, text, voice, image, video) input from the user, and the obtained information/signals are stored in the memory unit 130. It can be saved. The communication unit 110 may convert the information/signal stored in the memory into a wireless signal and transmit the converted wireless signal directly to another wireless device or to a base station. Additionally, the communication unit 110 may receive a wireless signal from another wireless device or a base station and then restore the received wireless signal to the original information/signal. The restored information/signal may be stored in the memory unit 130 and then output in various forms (eg, text, voice, image, video, haptics) through the input/output unit 140c.

46 illustrates a vehicle or autonomous vehicle applied to various embodiments of the present disclosure.

A vehicle or autonomous vehicle can be implemented as a mobile robot, vehicle, train, manned/unmanned aerial vehicle (AV), ship, etc.

Referring to Figure 46, the vehicle or autonomous vehicle 100 includes an antenna unit 108, a communication unit 110, a control unit 120, a drive unit 140a, a power supply unit 140b, a sensor unit 140c, and an autonomous driving unit. It may include a portion 140d. The antenna unit 108 may be configured as part of the communication unit 110. Blocks 110/130/140a to 140d respectively correspond to blocks 110/130/140 in FIG. 44.

The communication unit 110 may transmit and receive signals (e.g., data, control signals, etc.) with external devices such as other vehicles, base stations (e.g. base stations, road side units, etc.), and servers. The control unit 120 may control elements of the vehicle or autonomous vehicle 100 to perform various operations. The control unit 120 may include an Electronic Control Unit (ECU). The driving unit 140a can drive the vehicle or autonomous vehicle 100 on the ground. The driving unit 140a may include an engine, motor, power train, wheels, brakes, steering device, etc. The power supply unit 140b supplies power to the vehicle or autonomous vehicle 100 and may include a wired/wireless charging circuit, a battery, etc. The sensor unit 140c can obtain vehicle status, surrounding environment information, user information, etc. The sensor unit 140c includes an inertial measurement unit (IMU) sensor, a collision sensor, a wheel sensor, a speed sensor, an inclination sensor, a weight sensor, a heading sensor, a position module, and a vehicle forward sensor. /May include a reverse sensor, battery sensor, fuel sensor, tire sensor, steering sensor, temperature sensor, humidity sensor, ultrasonic sensor, illuminance sensor, pedal position sensor, etc. The autonomous driving unit 140d includes technology for maintaining the driving lane, technology for automatically adjusting speed such as adaptive cruise control, technology for automatically driving along a set route, and technology for automatically setting and driving when a destination is set. Technology, etc. can be implemented.

For example, the communication unit 110 may receive map data, traffic information data, etc. from an external server. The autonomous driving unit 140d can create an autonomous driving route and driving plan based on the acquired data. The control unit 120 may control the driving unit 140a so that the vehicle or autonomous vehicle 100 moves along the autonomous driving path according to the driving plan (e.g., speed/direction control). During autonomous driving, the communication unit 110 may acquire the latest traffic information data from an external server irregularly/periodically and obtain surrounding traffic information data from surrounding vehicles. Additionally, during autonomous driving, the sensor unit 140c can obtain vehicle status and surrounding environment information. The autonomous driving unit 140d may update the autonomous driving route and driving plan based on newly acquired data/information. The communication unit 110 may transmit information about vehicle location, autonomous driving route, driving plan, etc. to an external server. An external server can predict traffic information data in advance using AI technology, etc., based on information collected from vehicles or self-driving vehicles, and provide the predicted traffic information data to the vehicles or self-driving vehicles.

47 illustrates a vehicle applied to various embodiments of the present disclosure. Vehicles can also be implemented as transportation, trains, airplanes, ships, etc.

Referring to FIG. 47, the vehicle 100 may include a communication unit 110, a control unit 120, a memory unit 130, an input/output unit 140a, and a position measurement unit 140b. Here, blocks 110 to 130/140a to 140b correspond to blocks 110 to 130/140 in FIG. 44, respectively.

The communication unit 110 may transmit and receive signals (eg, data, control signals, etc.) with other vehicles or external devices such as a base station. The control unit 120 can control components of the vehicle 100 to perform various operations. The memory unit 130 may store data/parameters/programs/codes/commands that support various functions of the vehicle 100. The input/output unit 140a may output an AR/VR object based on information in the memory unit 130. The input/output unit 140a may include a HUD. The location measuring unit 140b may obtain location information of the vehicle 100. The location information may include absolute location information of the vehicle 100, location information within the driving line, acceleration information, location information with surrounding vehicles, etc. The location measuring unit 140b may include GPS and various sensors.

For example, the communication unit 110 of the vehicle 100 may receive map information, traffic information, etc. from an external server and store them in the memory unit 130. The location measurement unit 140b may acquire vehicle location information through GPS and various sensors and store it in the memory unit 130. The control unit 120 creates a virtual object based on map information, traffic information, and vehicle location information, and the input/output unit 140a can display the generated virtual object on the window of the vehicle (1410, 1420). Additionally, the control unit 120 may determine whether the vehicle 100 is operating normally within the driving line based on vehicle location information. If the vehicle 100 deviates from the driving line abnormally, the control unit 120 may display a warning on the window of the vehicle through the input/output unit 140a. Additionally, the control unit 120 may broadcast a warning message regarding driving abnormalities to surrounding vehicles through the communication unit 110. Depending on the situation, the control unit 120 may transmit location information of the vehicle and information about driving/vehicle abnormalities to the relevant organizations through the communication unit 110.

Figure 48 illustrates an XR device applied to various embodiments of the present disclosure. XR devices can be implemented as HMDs, HUDs (Head-Up Displays) installed in vehicles, televisions, smartphones, computers, wearable devices, home appliances, digital signage, vehicles, robots, etc.

Referring to FIG. 48, the XR device 100a may include a communication unit 110, a control unit 120, a memory unit 130, an input/output unit 140a, a sensor unit 140b, and a power supply unit 140c. . Here, blocks 110 to 130/140a to 140c correspond to blocks 110 to 130/140 in FIG. 44, respectively.

The communication unit 110 may transmit and receive signals (eg, media data, control signals, etc.) with external devices such as other wireless devices, mobile devices, or media servers. Media data may include video, images, sound, etc. The control unit 120 may perform various operations by controlling the components of the XR device 100a. For example, the control unit 120 may be configured to control and/or perform procedures such as video/image acquisition, (video/image) encoding, and metadata generation and processing. The memory unit 130 may store data/parameters/programs/codes/commands necessary for driving the XR device 100a/creating an XR object. The input/output unit 140a may obtain control information, data, etc. from the outside and output the generated XR object. The input/output unit 140a may include a camera, microphone, user input unit, display unit, speaker, and/or haptic module. The sensor unit 140b can obtain XR device status, surrounding environment information, user information, etc. The sensor unit 140b may include a proximity sensor, an illumination sensor, an acceleration sensor, a magnetic sensor, a gyro sensor, an inertial sensor, an RGB sensor, an IR sensor, a fingerprint recognition sensor, an ultrasonic sensor, an optical sensor, a microphone, and/or a radar. there is. The power supply unit 140c supplies power to the XR device 100a and may include a wired/wireless charging circuit, a battery, etc.

As an example, the memory unit 130 of the XR device 100a may include information (eg, data, etc.) necessary for creating an XR object (eg, AR/VR/MR object). The input/output unit 140a can obtain a command to operate the XR device 100a from the user, and the control unit 120 can drive the XR device 100a according to the user's driving command. For example, when a user tries to watch a movie, news, etc. through the XR device 100a, the control unit 120 sends content request information to another device (e.g., mobile device 100b) or It can be transmitted to a media server. The communication unit 130 may download/stream content such as movies and news from another device (eg, mobile device 100b) or a media server to the memory unit 130. The control unit 120 controls and/or performs procedures such as video/image acquisition, (video/image) encoding, and metadata creation/processing for the content, and acquires it through the input/output unit 140a/sensor unit 140b. XR objects can be created/output based on information about surrounding space or real objects.

Additionally, the XR device 100a is wirelessly connected to the mobile device 100b through the communication unit 110, and the operation of the XR device 100a can be controlled by the mobile device 100b. For example, the mobile device 100b may operate as a controller for the XR device 100a. To this end, the XR device 100a may obtain 3D location information of the mobile device 100b and then generate and output an XR object corresponding to the mobile device 100b.

Figure 49 illustrates a robot applied to various embodiments of the present disclosure. Robots can be classified into industrial, medical, household, military, etc. depending on the purpose or field of use.

Referring to FIG. 49, the robot 100 may include a communication unit 110, a control unit 120, a memory unit 130, an input/output unit 140a, a sensor unit 140b, and a driver 140c. Here, blocks 110 to 130/140a to 140c correspond to blocks 110 to 130/140 in FIG. 44, respectively.

The communication unit 110 may transmit and receive signals (e.g., driving information, control signals, etc.) with external devices such as other wireless devices, other robots, or control servers. The control unit 120 can control the components of the robot 100 to perform various operations. The memory unit 130 may store data/parameters/programs/codes/commands that support various functions of the robot 100. The input/output unit 140a may obtain information from the outside of the robot 100 and output the information to the outside of the robot 100. The input/output unit 140a may include a camera, microphone, user input unit, display unit, speaker, and/or haptic module. The sensor unit 140b can obtain internal information of the robot 100, surrounding environment information, user information, etc. The sensor unit 140b may include a proximity sensor, an illumination sensor, an acceleration sensor, a magnetic sensor, a gyro sensor, an inertial sensor, an IR sensor, a fingerprint recognition sensor, an ultrasonic sensor, an optical sensor, a microphone, a radar, etc. The driving unit 140c can perform various physical operations such as moving robot joints. Additionally, the driving unit 140c can cause the robot 100 to run on the ground or fly in the air. The driving unit 140c may include an actuator, motor, wheel, brake, propeller, etc.

Figure 50 illustrates an AI device applied to various embodiments of the present disclosure.

AI devices are fixed or mobile devices such as TVs, projectors, smartphones, PCs, laptops, digital broadcasting terminals, tablet PCs, wearable devices, set-top boxes (STBs), radios, washing machines, refrigerators, digital signage, robots, vehicles, etc. It can be implemented with available devices, etc.

Referring to FIG. 50, the AI device 100 includes a communication unit 110, a control unit 120, a memory unit 130, an input/output unit (140a/140b), a learning processor unit 140c, and a sensor unit 140d. may include. Blocks 110 to 130/140a to 140d correspond to blocks 110 to 130/140 in FIG. 44, respectively.

The communication unit 110 uses wired and wireless communication technology to communicate with external devices such as other AI devices (e.g., Figure W1, 100x, 200, 400) or the AI server 200, and wired and wireless signals (e.g., sensor information, user input, learning models, control signals, etc.) can be transmitted and received. To this end, the communication unit 110 may transmit information in the memory unit 130 to an external device or transmit a signal received from an external device to the memory unit 130.

The control unit 120 may determine at least one executable operation of the AI device 100 based on information determined or generated using a data analysis algorithm or a machine learning algorithm. And, the control unit 120 can control the components of the AI device 100 to perform the determined operation. For example, the control unit 120 may request, search, receive, or utilize data from the learning processor unit 140c or the memory unit 130, and may select at least one executable operation that is predicted or is determined to be desirable. Components of the AI device 100 can be controlled to execute operations. In addition, the control unit 120 collects history information including the operation content of the AI device 100 or the user's feedback on the operation, and stores it in the memory unit 130 or the learning processor unit 140c, or the AI server ( It can be transmitted to an external device such as Figure W1, 400). The collected historical information can be used to update the learning model.

The memory unit 130 can store data supporting various functions of the AI device 100. For example, the memory unit 130 may store data obtained from the input unit 140a, data obtained from the communication unit 110, output data from the learning processor unit 140c, and data obtained from the sensing unit 140. Additionally, the memory unit 130 may store control information and/or software codes necessary for operation/execution of the control unit 120.

The input unit 140a can obtain various types of data from outside the AI device 100. For example, the input unit 120 may acquire training data for model training and input data to which the learning model will be applied. The input unit 140a may include a camera, microphone, and/or a user input unit. The output unit 140b may generate output related to vision, hearing, or tactile sensation. The output unit 140b may include a display unit, a speaker, and/or a haptic module. The sensing unit 140 may obtain at least one of internal information of the AI device 100, surrounding environment information of the AI device 100, and user information using various sensors. The sensing unit 140 may include a proximity sensor, an illumination sensor, an acceleration sensor, a magnetic sensor, a gyro sensor, an inertial sensor, an RGB sensor, an IR sensor, a fingerprint recognition sensor, an ultrasonic sensor, an optical sensor, a microphone, and/or a radar. there is.

The learning processor unit 140c can train a model composed of an artificial neural network using training data. The learning processor unit 140c may perform AI processing together with the learning processor unit of the AI server (FIG. W1, 400). The learning processor unit 140c may process information received from an external device through the communication unit 110 and/or information stored in the memory unit 130. Additionally, the output value of the learning processor unit 140c may be transmitted to an external device through the communication unit 110 and/or stored in the memory unit 130.

Claims set forth in various embodiments of the present disclosure may be combined in various ways. For example, technical features of method claims of various embodiments of the present disclosure may be combined and implemented as a device, and technical features of device claims of various embodiments of the present disclosure may be combined and implemented as a method. In addition, the technical features of the method claims and the technical features of the device claims of various embodiments of the present disclosure may be combined to implement a device, and the technical features of the method claims and the device claims of various embodiments of the present disclosure may be combined. It can be implemented in this way.

Claims (20)

In a method of operating a terminal (user equipment, UE) in a wireless communication system, Receiving one or more synchronization signals from a base station (BS); Receiving system information from the base station; Receiving a reference signal from the base station; Generating channel state information (CSI) based on the reference signal; generating a plurality of signals for the CSI based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel; Including transmitting the plurality of signals to the base station, method. According to claim 1, The CSI can be restored based on a first signal among the plurality of signals, The CSI can be restored based on the first signal and one or more other signals among the plurality of signals, Restoration of the CSI is not possible with only the one or more signals other than the first signal, method. According to clause 2, When based on the first signal and the other one or more signals, the restoration performance of the CSI is higher than when based on the first signal alone, method. According to claim 1, Each of the first signal and the other one or more signals correspond to each of the plurality of encoder components, method. According to clause 2, When restoration of the CSI is performed based on only the first signal, and when restoration of the CSI is performed based on the first signal and one or more other signals, the same decoder NN model is used, method. According to clause 5, When restoration of the CSI is performed based on the first signal and the one or more other signals, The first signal and the other one or more signals are combined into one signal through a weighted sum and then input to the decoder NN model, method. According to clause 6, The decoder NN model adaptively performs the weighted sum for the first signal and the other one or more signals according to a CSI feedback rate that varies based on at least one of a channel state and an uplink resource state. composed, method. According to claim 1, Each of the plurality of encoder components includes one or more convolutional layers, The encoder NN model consists of a parameter set for convolutional layers included in the plurality of encoder components, method. In a method of operating a base station (BS) in a wireless communication system, Transmitting one or more synchronization signals to a user equipment (UE); Transmitting system information to the terminal; Transmitting a reference signal to the terminal; Comprising receiving a plurality of signals from the terminal, The plurality of signals are related to channel state information (CSI) based on the reference signal, The plurality of signals are generated based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel, method. According to clause 9, The CSI can be restored based on a first signal among the plurality of signals, The CSI can be restored based on the first signal and one or more other signals among the plurality of signals, Restoration of the CSI is not possible with only the one or more signals other than the first signal, method. According to claim 10, When based on the first signal and the other one or more signals, the restoration performance of the CSI is higher than when based on the first signal alone, method. According to clause 9, Each of the first signal and the other one or more signals correspond to each of the plurality of encoder components, method. According to claim 10, When restoration of the CSI is performed based on only the first signal, and when restoration of the CSI is performed based on the first signal and one or more other signals, the same decoder NN model is used, method. According to claim 13, When restoration of the CSI is performed based on the first signal and the one or more other signals, The first signal and the other one or more signals are combined into one signal through a weighted sum and then input to the decoder NN model, method. According to claim 14, The decoder NN model adaptively performs the weighted sum for the first signal and the other one or more signals according to a CSI feedback rate that varies based on at least one of a channel state and an uplink resource state. composed, method. According to clause 9, Each of the plurality of encoder components includes one or more convolutional layers, The encoder NN model consists of a parameter set for convolutional layers included in the plurality of encoder components, method. In a terminal (user equipment, UE) in a wireless communication system, transceiver; and Contains at least one processor, The at least one processor, Receive one or more synchronization signals from a base station (BS), Receive system information from the base station, Receive a reference signal from the base station, Generate channel state information (CSI) based on the reference signal, Generating a plurality of signals for the CSI based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel, configured to transmit the plurality of signals to the base station, Terminal. In a base station (BS) in a wireless communication system, transceiver; and Contains at least one processor, The at least one processor, Transmit one or more synchronization signals to a user equipment (UE), Transmit system information to the terminal, Transmit a reference signal to the terminal, Configured to receive a plurality of signals from the terminal, The plurality of signals are related to channel state information (CSI) based on the reference signal, The plurality of signals are generated based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel, Base station. In one or more non-transitory computer-readable media storing one or more instructions, the one or more instructions perform operations based on execution by one or more processors; The above operations are: Receiving one or more synchronization signals from a base station (BS); Receiving system information from the base station; Receiving a reference signal from the base station; Generating channel state information (CSI) based on the reference signal; generating a plurality of signals for the CSI based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel; Including transmitting the plurality of signals to the base station, Computer-readable media. In one or more non-transitory computer-readable media storing one or more instructions, the one or more instructions perform operations based on execution by one or more processors; The above operations are: Transmitting one or more synchronization signals to a user equipment (UE); Transmitting system information to the terminal; Transmitting a reference signal to the terminal; Comprising receiving a plurality of signals from the terminal, The plurality of signals are related to channel state information (CSI) based on the reference signal, The plurality of signals are generated based on an encoder neural network (NN) model including a plurality of encoder components arranged in parallel, Computer-readable media.

Images