WO2024093997A1 - Method and apparatus for determining model applicability, and communication device - Google Patents

Abstract

The present application belongs to the technical field of communications. Disclosed are a method and apparatus for determining model applicability, and a communication device. The method for determining model applicability in the embodiments of the present application comprises: a communication device determining target feature information corresponding to target channel information; and according to the target feature information, determining that a target AI model is applicable or is not applicable.

Description

Method, device and communication equipment for determining model applicability

cross reference

This application claims priority to a Chinese patent application filed with the China Patent Office on November 4, 2022, with application number 202211379119.5 and invention name “Method, device and communication equipment for determining model applicability”. The entire contents of the application are incorporated by reference into this application.

Technical Field

The present application belongs to the field of communication technology, and specifically relates to a method, device and communication equipment for determining the applicability of a model.

Background technique

Artificial Intelligence (AI) has been widely used in various fields. For example, integrating AI into wireless communication networks can significantly improve technical indicators such as throughput, latency, and user capacity of wireless communication networks, and is an important research and development direction for future wireless communication networks.

Regarding the application of AI in wireless communication systems, the relevant technologies generally use communication equipment (such as terminals or network-side equipment, etc.) to constantly observe the output of the AI model and/or the system performance of the communication system to determine whether the AI model is applicable in the current communication environment (or wireless communication system). However, the aforementioned AI model applicability determination scheme will cause the AI model to continue to work for a long time in an inappropriate communication environment, affecting the performance of the communication system.

Summary of the invention

The embodiments of the present application provide a method, apparatus, and communication device for determining the applicability of a model, which can avoid the problem that the AI model continues to work for a long time in an inappropriate communication environment and ensure the performance of the communication system.

In a first aspect, a method for determining model applicability is provided, including: a communication device determines target feature information corresponding to target channel information; and determines whether the target AI model is applicable or not based on the target feature information.

In a second aspect, a device for determining the applicability of a model is provided, including: a determination module for determining target feature information corresponding to target channel information; and determining whether the target AI model is applicable or not based on the target feature information.

According to a third aspect, a communication device is provided, comprising a processor and a memory, wherein the memory stores a program or instruction that can be executed on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the first aspect are implemented.

In a fourth aspect, a communication device is provided, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps of the method described in the first aspect.

In a fifth aspect, a communication system is provided, comprising: at least one communication device, wherein the communication device can be used to perform Perform the steps of the method described in the first aspect.

In a sixth aspect, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.

In a seventh aspect, a chip is provided, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps of the method described in the first aspect.

In an eighth aspect, a computer program product/program product is provided, wherein the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the steps of the method described in the first aspect.

In an embodiment of the present application, the communication device determines the target feature information corresponding to the target channel information, and then determines whether the target AI model is applicable based on the target feature information. This can improve the efficiency of determining the applicability of the target AI model, avoid the problem in the related technology that the communication device needs to constantly observe the output of the AI model and/or the system performance of the communication system to determine the applicability of the AI model, and cause the AI model to need to continue to work for a long time in an inappropriate communication environment, thereby effectively ensuring the performance of the communication system.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the structure of a wireless communication system provided by an exemplary embodiment of the present application.

FIG. 2 is a flowchart of a method for determining model applicability provided by an exemplary embodiment of the present application.

FIG. 3 is a second flowchart of a method for determining model applicability provided by an exemplary embodiment of the present application.

FIG. 4 is a third flowchart of a method for determining model applicability provided by an exemplary embodiment of the present application.

FIG5 is a schematic diagram of the structure of an apparatus for determining model applicability provided by an exemplary embodiment of the present application.

FIG. 6 is a schematic diagram of the structure of a communication device provided by an exemplary embodiment of the present application.

FIG. 7 is a schematic diagram of the structure of a terminal provided by an exemplary embodiment of the present application.

FIG8 is a schematic diagram of the structure of a network side device provided by an exemplary embodiment of the present application.

Detailed ways

The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of this application.

The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first" and "second" are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and/or" in the specification and claims represents at least one of the connected objects, and the character "/" generally represents that the objects associated with each other are in an "or" relationship.

It is worth noting that the technology described in the embodiments of the present application is not limited to the Long Term Evolution (LTE). The present invention relates to an LTE/LTE evolution (LTE-Advanced, LTE-A) system, and can also be used for other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA) and other systems. The terms "system" and "network" in the embodiments of the present application are often used interchangeably, and the described techniques can be used for the systems and radio technologies mentioned above as well as for other systems and radio technologies. The following description describes a New Radio (NR) system for example purposes, and NR terminology is used in most of the following descriptions, but these techniques can also be applied to applications other than NR system applications, such as the ^6th Generation (6G) communication system.

FIG1 shows a block diagram of a wireless communication system applicable to an embodiment of the present application. The wireless communication system includes a terminal 11 and a network side device 12. The terminal 11 may be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer) or a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a handheld computer, a netbook, an ultra-mobile personal computer (ultra-mobile personal computer, UMPC), a mobile Internet device (Mobile Internet Device, MID), an augmented reality (augmented reality, AR)/virtual reality (virtual reality, VR) device The terminal side devices 12 include: smart devices, robots, wearable devices (Wearable Device), vehicle-mounted equipment (VUE), pedestrian terminals (PUE), smart homes (home appliances with wireless communication functions, such as refrigerators, televisions, washing machines or furniture, etc.), game consoles, personal computers (personal computers, PCs), ATMs or self-service machines, etc., and wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. It should be noted that the specific type of the terminal 11 is not limited in the embodiments of the present application. The network side device 12 may include an access network device or a core network device, wherein the access network device 12 may also be called a wireless access network device, a wireless access network (Radio Access Network, RAN), a wireless access network function or a wireless access network unit. The access network device 12 may include a base station, a wireless local area network (WLAN) access point or a wireless fidelity (WiFi) node, etc. The base station may be referred to as a node B, an evolved node B (eNB), an access point, a base transceiver station (BTS), a radio base station, a radio transceiver, a basic service set (BSS), an extended service set (ESS), a home node B, a home evolved node B, a transmitting and receiving point (TRP) or some other appropriate term in the field. As long as the same technical effect is achieved, the base station is not limited to a specific technical vocabulary. It should be noted that in the embodiment of the present application, only the base station in the NR system is used as an example for introduction, and the specific type of the base station is not limited. In conjunction with the accompanying drawings, the technical solution provided in the embodiment of the present application is described in detail through some embodiments and their application scenarios.

As shown in FIG2 , a flow chart of a method 200 for determining model applicability provided by an exemplary embodiment of the present application is provided. The method 200 may be, but is not limited to, executed by a communication device (such as a terminal or a network-side device), and may be specifically executed by hardware and/or software installed in the communication device. In this embodiment, the method 200 may at least include the following steps.

S210, the communication device determines target characteristic information corresponding to the target channel information.

Wherein, when the communication device determines the target characteristic information corresponding to the target channel information, it can be implemented by information statistics and the like, which is not limited here.

The target channel information can be collected by the communication device according to the protocol agreement or network side configuration, or it can be collected according to the channel information applicable to the target AI model, and there is no limitation here.

In this embodiment, the target characteristic information may be different according to different target channel information. As an implementation manner, the target characteristic information corresponding to the target channel information may include at least one of the following (11)-(19).

(11) Spatial beam information.

Optionally, in this embodiment, the spatial beam information may include at least one of the correlation between the index distribution vector of each beam and the first distribution vector, the first quantity, and the second quantity. The index of the beam may be the energy or power of the beam; the correlation is an indicator that measures the degree of association or distance between two vectors, for example, the correlation may be cosine similarity or the square of cosine similarity, etc., which are not listed here one by one.

Based on this, the first number can be the number of beams corresponding to the multiple beams when the ratio of the sum of the indicators of the multiple beams to the total indicators of the beams reaches or exceeds the first threshold. For example, assuming that there are 10 beams in total, such as beam 1, beam 2, beam 3, beam 4, beam 5, beam 6, beam 7, beam 8, beam 9, and beam 10, and the first threshold is X1, then, if the sum of the indicators (such as power or energy) of 5 beams (such as beam 1, beam 4, beam 5, beam 8, and beam 9) among the 10 beams accounts for a ratio of the sum of the indicators of the 10 beams that reaches or exceeds the first threshold X1, then the first number is 5. Optionally, the X1 can be implemented by protocol default, network configuration, or terminal reporting, and is not limited here. Optionally, before calculating the sum of the indicators of the multiple beams, all beams can be sorted according to their indicators, for example, from large to small, or from small to large.

The second number is the number of single beams corresponding to the ratio of the index of a single beam to the total index of the beams when it reaches or exceeds the second threshold. For example, assuming there are 10 beams in total, such as beam 1, beam 2, beam 3, beam 4, beam 5, beam 6, beam 7, beam 8, beam 9, beam 10, and the second threshold is X2, then, if the ratio of the index of beam 3, beam 7, and beam 10 among the 10 beams to the sum of the indexes of the 10 beams reaches or exceeds the second threshold X2, then the second number is 3, i.e., beam 3, beam 7, and beam 10. Optionally, X2 can be implemented by protocol default, network configuration, or terminal reporting, which is not limited here.

The first distribution vector is a beam index distribution vector adapted to the target AI model. Of course, the index distribution vectors of the aforementioned beams can be understood as follows: Assuming that the index of the beam is energy or power, and there are N0 beams in total, then these N0 beams can be recorded as vectors in the form of [energy or power of the first beam, energy or power of the second beam, ..., energy of the N0th beam].

In addition, in one implementation, the index distribution vector of each beam can be obtained by shifting or cyclically shifting each beam according to the index size of each beam. For example, each time statistics are taken, the beam with the highest power is fixed to the N1th beam, and then all beam energy graphs are cyclically shifted. For example, at this time, the highest power beam is the N2th beam. If N2>N1, the beam energy graph is shifted to the left by (N2-N1) beams according to the beam ID, that is, beam N3 at the time of statistics is actually the energy of beam (N3+N2-N1); if N2<N1, the beam energy graph is shifted to the right by (N1-N2) beams according to the beam ID, that is, beam N3 at the time of statistics is actually the energy of beam (N3+N2-N1); if N2＝N1, the beam energy graph No need to move.

Alternatively, the index distribution vectors of the beams may also be obtained by arranging the index sizes of the beams in a descending order or a descending order.

(12) Channel Impulse Response (CIR).

In this embodiment, the CIR includes at least one of the correlation between the indicator distribution vector of each path and the second distribution vector, the third quantity, the fourth quantity, the first path position, the first path indicator, the main path position, and the main path indicator, and the second distribution vector is a path distribution vector adapted to the target AI model. Among them, the correlation can be an indicator to measure the degree of association or distance between two vectors. For example, the correlation can be cosine similarity or the square of cosine similarity, etc., which are not listed here one by one. It should be noted that the path indicators mentioned in this embodiment may include at least one of energy, power, reference signal received power (Reference Signal Received Power, RSRP), and reference signal time difference (Reference Signal Time Difference, RSTD).

Among them, the third number is the number of paths corresponding to the multiple paths when the ratio of the sum of the indicators of the multiple paths to the total indicators of the paths reaches or exceeds the third threshold. For example, assuming that there are 10 paths in total, such as path 1, path 2, path 3, path 4, path 5, path 6, path 7, path 8, path 9, and path 10, and the third threshold is X3, then, if the sum of the indicators of 5 paths (such as path 1, path 4, path 5, path 8, and path 9) among the 10 paths accounts for a ratio of the sum of the indicators of the 10 paths that reaches or exceeds the third threshold X3, then the third number is 5. The X3 can be implemented by protocol default, network configuration, or terminal reporting, and is not limited here. Optionally, before calculating the sum of the indicators of the multiple paths, all paths need to be sorted according to their indicators, for example, from large to small, or from small to large.

The fourth number is the number of paths corresponding to a single path when the ratio of the index of a single path to the total index of the paths reaches or exceeds the fourth threshold. For example, assuming there are 10 paths in total, such as path 1, path 2, path 3, path 4, path 5, path 6, path 7, path 8, path 9, and path 10, and the fourth threshold is X4, then if the ratio of the index of path 3, path 7, and path 10 among the 10 paths to the sum of the indexes of the 10 paths reaches or exceeds the fourth threshold X4, then the fourth number is 3, i.e., path 3, path 7, and path 10. The X4 can be implemented by protocol default, network configuration, or terminal reporting, and is not limited here.

Among them, the energy or power distribution vector of each path can be understood as: assuming that there are N0 paths in total, and the index of the path is energy or power, then it can be recorded in vector form as [energy or power of the first path, energy or power of the second path, ..., energy of the N0th path].

In addition, in one implementation, the index distribution vector of each path is obtained by shifting or cyclically shifting each path according to the index size of each path. For example, each time statistics are taken, the path with the highest power is fixed as the N1th path, and then all path energy graphs are cyclically shifted. For example, at this time, the highest power path is the N2th path. If N2>N1, the path energy graph is shifted to the left by (N2-N1) paths according to the path ID, that is, the path N3 during statistics is actually the energy of the path (N3+N2-N1); if N2<N1, the path energy graph is shifted to the right by (N1-N2) paths according to the path identifier (ID), that is, the path N3 during statistics is actually the energy of the path (N3+N2-N1); if N2=N1, the path energy graph does not need to be moved.

Alternatively, the index distribution vectors of the paths may be obtained by arranging the index sizes of the paths in a descending order or in a descending order.

(13) Power Delay Profile (PDP) information.

In this embodiment, similar to the aforementioned CIR, the PDP information includes at least one of the correlation between the indicator distribution vector of each path and the second distribution vector, the third quantity, the fourth quantity, the first path position, the first path indicator, the main path position, and the main path indicator; wherein the third quantity is the number of paths corresponding to the multiple paths when the ratio of the sum of the indicators of multiple paths to the total path indicator reaches or exceeds the third threshold, the fourth quantity is the number of paths corresponding to the single path when the ratio of the indicator of a single path to the total path indicator reaches or exceeds the fourth threshold, the second distribution vector is the path distribution vector adapted to the target AI model, and the path indicator includes at least one of energy, power, RSRP, and RSTD. It can be understood that the PDP information can refer to the relevant description in the aforementioned CIR, which will not be repeated here.

(14) Delay spread information.

(15) Doppler information.

(16) Time of Arrival (TOA) information.

(17) Line of Sight (LOS) information.

(18) Non-Line-of-Sight (NLOS) information.

(19) Rank-related information: The rank-related information can be understood as the distribution of the feature vectors of each data stream or the gap between them, so as to characterize the concentration of indicators such as energy of the data stream.

It should be noted that, in a possible implementation, the target channel information and/or target feature information mentioned above may be related to the scope of application of the target AI model in S220. For example, for the target feature information, assuming that the target AI model is applicable to the processing of PDP and TOA, then the target feature information may be PDP and TOA, and this embodiment does not limit this.

In this embodiment, the rank-related information may include at least one of the correlation between the index distribution vector of each data stream and the third distribution vector, the fifth quantity, and the sixth quantity. The third distribution vector is a data stream distribution vector adapted to the target AI model, and the index of the data stream includes at least one of energy, power, eigenvalue, and singular value.

The fifth quantity is the number of data streams corresponding to the multiple data streams when the ratio of the sum of the indicators of the multiple data streams to the total indicators of the data streams reaches or exceeds the fifth threshold. For example, taking the indicator of the data stream as total energy as an example, assuming that there are 5 data streams in total, such as data stream 1, data stream 2, data stream 3, data stream 4, and data stream 5, and the fifth threshold is X5, then, if the sum of the total energy of 2 data streams (such as data stream 1 and data stream 4) among the 5 data streams accounts for the sum of the total energy of the 5 data streams. The ratio reaches or exceeds the fifth threshold X5, then the third quantity is 2. The X5 can be implemented by protocol default, network configuration or terminal reporting. Optionally, before calculating the sum of the indicators of the multiple data streams, all data streams can be sorted according to their indicators, or all data streams need to be sorted according to their data stream identifiers, such as sorting from large to small, or from small to large.

The sixth quantity is the number of data streams corresponding to a single data stream when the ratio of the index of a single data stream to the total index of the data stream reaches or exceeds the sixth threshold. For example, taking the index of the data stream as total energy, assuming that there are 5 data streams in total, such as data stream 1, data stream 2, data stream 3, data stream 4, and data stream 5, and the sixth threshold is X6, then, if the ratio of the energy of data stream 3 and data stream 4 in the 5 data streams to the sum of the indexes of the 5 data streams reaches or exceeds the sixth threshold X6, then the fourth quantity is 3, that is, data stream 3 and data stream 4. The X6 can be The protocol default, network configuration or terminal reporting implementation is not restricted here.

It can be understood that the aforementioned data stream can also be understood as a data block or layer (Layer), and this embodiment does not limit this.

S220, determining whether the target AI model is applicable or not based on the target feature information.

Among them, the target AI model mentioned in the context of this application has multiple implementation methods, such as the target AI model can be a neural network, a decision tree, a support vector machine, a Bayesian classifier, etc.

In addition, compared with the related art that requires constant observation of the output of the AI model and/or the performance of the communication system implemented based on the AI model to determine whether the AI model is suitable for the current communication environment, this embodiment directly determines whether the target AI model is suitable based on the target feature information corresponding to the target channel information. This can more efficiently determine whether the AI model is suitable, and avoids the problem that the AI model needs to work for a long time in an inapplicable environment, thereby effectively ensuring the performance of the communication system.

It should be noted that, as a possible implementation method, the applicable scope of the target AI model is determined by the feature information corresponding to the training data (i.e., the data used for training the target AI model). When using the training data to train the target AI model, the corresponding feature information of the training data is determined as the applicable scope of the target AI model. For example, the target AI model is used for positioning, the target feature information is the LOS information mean, and the corresponding LOS information mean of the training data is Y1, then the applicable scope of the target AI model is near the LOS information mean Y1.

Of course, the training data used for training the target AI model is determined according to the purpose of the target AI model. For example, assuming that the target AI model is used for signal processing, then the training data is related to signal processing. For another example, assuming that the target AI model is used for channel prediction, then the training data is related to channel prediction, etc., and there is no limitation here.

Based on this, in one implementation, the purpose of the target AI model in this embodiment may include at least one of the following.

Signal processing. The signal processing includes signal detection, signal filtering, signal equalization, etc. Of course, the signal can be a demodulation reference signal (Demodulation Reference Signal, DMRS), a sounding reference signal (Sounding Reference Signal, SRS), a synchronization signal (Synchronization Signal Block, SSB), a phase reference signal (Tracking Reference Signal, TRS), a phase tracking reference signal (Phase-tracking reference signal, PTRS), a channel state information reference signal (Channel State Information Reference Signal, CSI-RS), etc.

Signal demodulation. The signal demodulation may be the demodulation of signals such as the physical downlink control channel (PDCCH), the physical downlink shared channel (PDSCH), the physical uplink control channel (PUCCH), the physical uplink shared channel (PUSCH), the physical random access channel (PRACH), and the physical broadcast channel (PBCH).

Signal transmission and reception: The signal transmission and reception may be transmission and reception of PDCCH, PDSCH, PUCCH, PUSCH, PRACH, PBCH and other signals.

Channel state information acquisition. The channel state information acquisition includes signal state information feedback and frequency division multiplexing (FDD) uplink and downlink partial reciprocity acquisition, etc. Among them, the signal state information feedback may include channel related information, channel matrix related information, channel characteristic information, channel matrix characteristic information, precoding matrix indicator (Precoding matrix indicator, PMI), rank indicator (Rank indicator, RI), CSI-RS resource indicator (CSI-RS Resource Indicator, CRI), channel quality indicator (CQI), layer indicator (Layer Indicator, LI) and other information feedback. The FDD uplink and downlink partial reciprocity can be understood as: for the FDD system, according to the partial reciprocity, the base station and other network side devices obtain the angle and delay information according to the uplink channel, and can notify the terminal of the angle information and delay information through CSI-RS precoding or direct indication. The terminal reports according to the indication of the base station or selects and reports within the indication range of the base station, thereby reducing the calculation amount of the terminal and the overhead of CSI reporting.

Beam management: The beam management may include beam measurement, beam reporting, beam prediction, beam failure detection, beam failure recovery, new beam indication in beam failure recovery, etc.

Channel prediction: The channel prediction may include prediction of channel state information, beam prediction, etc.

Interference suppression: The interference suppression may include suppression of intra-cell interference, inter-cell interference, out-of-band interference, intermodulation interference, etc.

Terminal positioning: The terminal positioning includes estimating the specific position (including horizontal position and/or vertical position) or possible future trajectory of the terminal through a reference signal (such as SRS), or information to assist position estimation or trajectory estimation.

Prediction and management of high-level services and parameters: The prediction and management of high-level services and parameters may include throughput, required data packet size, service requirements, mobile speed, noise information, etc.

Analysis of control signaling: The analysis of control signaling may include analysis of power control related signaling, beam management related signaling, etc.

In this embodiment, by determining the target feature information corresponding to the target channel information, and then determining whether the target AI model is applicable based on the target feature information, the efficiency of determining the applicability of the target AI model can be improved, thereby avoiding the problem in related technologies that the communication equipment needs to constantly observe the output of the AI model and/or the system performance of the communication system to determine the applicability of the AI model, resulting in the AI model needing to continue to work for a long time in an inappropriate communication environment, thereby effectively ensuring the performance of the communication system.

In addition, if multiple AI models are stored in the communication device, the target feature information corresponding to the target channel information is periodically determined through the present application, and the applicability of the target AI model is determined based on the target feature information to flexibly select the most matching AI model, thereby greatly improving the generalization performance of the AI model in different communication environments and ensuring the flexibility and stability of the communication system.

As shown in FIG3 , a flow chart of a method 300 for determining model applicability provided by an exemplary embodiment of the present application is provided. The method 300 may be, but is not limited to, executed by a communication device (such as a terminal or a network-side device), and may be specifically executed by hardware and/or software installed in the communication device. In this embodiment, the method 300 may at least include the following steps.

S310, the communication device determines target characteristic information corresponding to the target channel information.

It can be understood that the implementation process of S310 can refer to the relevant description in the method embodiment 200. In the implementation method, when determining the target characteristic information corresponding to the target channel information, the communication device can determine the target characteristic information of the target channel information in the target domain; wherein the target domain may include but is not limited to at least one of the delay domain, beam domain, Doppler domain, and space domain.

Of course, if the communication device determines the target characteristic information of the target channel information, and the determined target domain is different from the domain corresponding to the acquired target channel information, conversion between different domains can be performed. For example, when the target domain includes the delay domain, the target channel information in the frequency domain is converted to the delay domain; and/or, when the target domain includes the beam domain, the target channel information in the antenna domain is converted to the beam domain; and/or, when the target domain includes the Doppler domain, the target channel information in the time domain is converted to the Doppler domain.

Based on this, in one implementation, the target domain determined by the communication device may be one or more. Then, when there are multiple target domains, the communication device may combine the characteristic information of multiple target domains for comparison to determine whether the target AI model is applicable (or whether the target AI model is invalid). For example, when the target domain includes the delay domain, the Doppler domain, and the beam domain, it is possible to simultaneously determine whether the target AI model is applicable based on the characteristic information of the delay domain, the characteristic information of the Doppler domain, and the characteristic information of the beam domain. For example, when the characteristic information of the delay domain, the characteristic information of the Doppler domain, and the characteristic information of the beam domain are all within the applicable scope of the target AI model, it is determined that the target AI model is applicable. Otherwise, it is determined that the target AI model is not applicable.

Furthermore, when determining whether the target AI model is applicable, in addition to determining the target feature information on the target domain, in another implementation method, the communication device may determine the target feature information corresponding to the target channel information in any one of the following methods 1-3.

Mode 1: determining the statistical value of the target feature information according to the statistical value of the previous target feature information and the currently collected target feature information.

For example, assuming that the statistical value of the last target feature information is X, and the currently collected target feature information is Y, then the statistical value of the target feature information is alpha*X+beta*Y, where alpha and beta are weights. In this embodiment, alpha and beta can be implemented by protocol agreement, high-level configuration, etc. Optionally, the values of alpha and beta can both be 1.

Mode 2: determining the statistical value of the target feature information according to the average value of all target feature information collected within the first time.

The value of the first time may be implemented by protocol default or network configuration or terminal reporting. The average value may be a geometric average, an arithmetic average, a weighted average, etc. The weighted value of the weighted average may be implemented by protocol default or network configuration or terminal reporting, which is not limited here.

Mode 3, determining the statistical value of the target feature information based on a Gaussian mixture model (GMM). The Gaussian mixture model is to accurately quantify things using a Gaussian probability density function (normal distribution curve). It is a model that decomposes things into several Gaussian probability density functions. The GMM is a method for obtaining statistical information. Optionally, in this embodiment, several model parameters formed based on Gaussian probability density functions are used as the statistical value of the target feature information, such as the mean, variance and ratio/probability/contribution of each decomposed Gaussian probability density function to the total model as the statistical value of the target feature information.

For example, assuming that the first path delay information of N terminals in the area with an ID of S is obtained, the mean and variance of T Gaussian distributions can be obtained through the Gaussian mixture model to serve as the statistical value of the target feature information.

It should be noted that when the terminal determines the target feature information, which of the aforementioned methods 1-3 is adopted can be determined by protocol agreement, high-level configuration or terminal autonomy, etc., and is not restricted here.

In addition, the type of the target domain and/or the target feature information mentioned in this embodiment can be determined by at least one of the following (21)-(24).

(21) Network side instructions.

(22) Determined according to the configuration information of the target AI model, such as the target domain or the target feature information can be configured or included in the configuration information of the target AI model.

(23) Determined according to the description information of the target AI model, such as the target domain or the target feature information can be configured or included in the description information of the target AI model.

(24) obtained interactively during the training process of the target AI model.

S320: Determine whether the target AI model is applicable or not based on the target feature information.

It can be understood that in addition to referring to the relevant description in method embodiment 200, the implementation process of S320, as shown in Figure 3, in one implementation, determining whether the target AI model is applicable or not based on the target feature information may include the following S321 and/or S322.

S321: When the target feature information is within the applicable scope of the target AI model, determine that the target AI model is applicable.

Among them, the target feature information described in S321-S322 can be determined based on the target domain, and/or, the target feature information can be determined based on any of the aforementioned methods 1-3, which is not limited here.

In addition, in one implementation, when the target AI model is applicable, the communication device reports (or triggers) first information, or does not report any information, and the first information is used to indicate that the target AI model is available or can work normally. Of course, whether the communication device reports the first information can be implemented by protocol agreement or network side configuration, and is not limited here.

S322: When the target feature information is not within the applicable scope of the target AI model, determine that the target AI model is not applicable.

Optionally, when the target AI model is not applicable, second information is reported (or triggered), and the second information is used to indicate or request at least one of model switching, model deactivation, and enabling a non-AI algorithm. Of course, when the communication device indicates or requests model switching or enabling a non-AI algorithm, the target feature information is within the applicable scope of the AI model or algorithm to be switched or enabled.

In this embodiment, by adopting different methods of determining target feature information, it is possible to determine with low overhead whether the target AI model is applicable to the current environment, which can further improve the applicability and flexibility of the target AI model and ensure the performance of the communication system.

As shown in FIG. 4 , a flow chart of a method 400 for determining model applicability provided by an exemplary embodiment of the present application is shown. The method 400 may be, but is not limited to, executed by a communication device (such as a terminal or a network side device), and may be specifically executed by a communication device installed in The hardware and/or software in the communication device executes. In this embodiment, the method 400 may at least include the following steps.

S410, collecting or counting the target channel information or target feature information in a predetermined manner.

The predetermined method includes at least one of the following methods 1 to 6.

Mode 1: collecting or counting the target channel information or target feature information in real time.

Mode 2, based on the observation period, collects or counts the target channel information or target characteristic information at a second time interval; wherein the second time can be understood as the observation period. It can be understood that the observation period and the value of the second time can be implemented by protocol agreement, high-level configuration or network-side configuration, and is not limited here.

Mode 3, collect or count the target channel information or target feature information within the observation window. For example, within 200ms, only 1ms-10ms is within the observation window, then the communication device can only collect or count the target channel information or target feature information within 1ms-10ms to determine the applicability of the target AI model. It can be understood that the size of the observation window can be achieved by protocol agreement, high-level configuration or network-side configuration, and is not limited here.

Mode 4, based on the first observation position, the target channel information or target feature information is collected or counted when the communication device moves beyond a predetermined distance. For example, assuming that the first observation position is point A, then when the distance of the communication device from point A exceeds a predetermined distance, the communication device collects or counts the target channel information or target feature information. It can be understood that the values of the first observation position and the predetermined distance can be implemented by protocol predetermination, network side configuration or high-level configuration, and are not limited here.

Mode 5: Based on the second observation position, the target channel information or target feature information is collected or counted when the communication device leaves the designated area, and the designated area is the area where the target channel information or target feature information was collected last time. The second observation position can be implemented by protocol agreement, high-level configuration, network side configuration, etc., which is not limited here.

Mode 6: Based on the third observation position, when the change in the physical position of the communication device exceeds a predetermined value, the target channel information or target feature information is collected or counted. The third observation position and the predetermined value can be implemented by protocol agreement, high-level configuration, network side configuration, etc., and are not limited here.

It can be understood that for the aforementioned methods 4 to 6, if the target AI model is used for purposes such as terminal positioning, then when the communication device moves a predetermined distance or leaves a designated area or changes in physical position, it may cause the spatial statistical information to change, that is, the communication environment has changed. Therefore, the communication device verifies the applicability of the target AI model by collecting or counting target channel information or target feature information, thereby ensuring the performance of the communication system.

In addition, in the target channel information or target feature information statistics or collection methods given in the aforementioned methods 1-6, if the communication device directly collects or counts the target feature information based on methods 1-6, then the communication device can execute S430 based on the collected or counted target feature information to determine whether the target AI model is applicable. The determination process can refer to the relevant description in the aforementioned method embodiments 200-300, which will not be repeated here.

Of course, if the communication device directly collects or counts the target channel information based on method 1-method 6, then the communication device needs to execute S420 based on the collected or counted target channel information to obtain the target feature information corresponding to the target channel information, and then determine whether the target AI model is applicable based on the target feature information.

S420, the communication device determines target characteristic information corresponding to the target channel information.

It can be understood that the implementation process of S420 can refer to the relevant description in the aforementioned method embodiments 200 and/or 300. To avoid repetition, this embodiment will not be repeated here.

S430: Determine whether the target AI model is applicable or not based on the target feature information.

It can be understood that in addition to referring to the relevant descriptions in the aforementioned method embodiments 200 and/or 300, the implementation process of S430 can, as a possible implementation method, when the communication device determines whether the target AI model is applicable or not based on the target feature information, it can also calculate the target statistic corresponding to the target feature information corresponding to the target channel information, and determine whether the target AI model is applicable or not based on the target statistic (for example, whether the distance between the target statistic and a preset statistic is less than a threshold, and the preset statistic is determined based on the applicable scope of the AI model); wherein the target statistic includes at least one of the following (31)-(33).

(31)Mean.

(32) Variance.

Among them, the mean and variance in the above (31)-(32) can be first-order statistics, second-order statistics or higher-order statistics, and are not limited here.

For example, assuming that the target statistic is a first-order statistic, then after obtaining the target feature information, the communication device can calculate the mean and/or variance of the target feature information, and determine whether the target AI model is applicable or not based on the mean and/or variance.

(33) A statistic determined based on at least one of the following: Cumulative Distribution Function (CDF), Probability Density Function (PDF), or Probability Mass Function (PMF).

For example, after obtaining the target feature information, the communication device can calculate the target statistics corresponding to the target feature information based on CDF, PDF or PMF, and determine whether the target AI model is applicable or not based on the target statistics.

In this embodiment, by further providing different statistical or collection methods to obtain target channel information or target feature information, the generalization performance of the AI model in complex environments can be effectively improved, and the flexibility and stability of the communication system can be enhanced.

The method 200-400 for determining model applicability provided in the embodiments of the present application may be performed by a device for determining model applicability. In the embodiments of the present application, the device for determining model applicability is described by taking the method for determining model applicability performed by the device for determining model applicability as an example.

As shown in Figure 5, it is a structural diagram of an apparatus 500 for determining model applicability provided in an exemplary embodiment of the present application. The apparatus 500 includes a first determination module 510 for determining target feature information corresponding to target channel information; and a second determination module 520 for determining whether the target AI model is applicable or not based on the target feature information.

Optionally, the first determination module 510 determines the target characteristic information corresponding to the target channel information, including: determining the target characteristic information of the target channel information in the target domain; wherein the target domain includes at least one of a delay domain, a beam domain, and a Doppler domain.

Optionally, the first determination module 510 determines the target characteristic information of the target channel information, and further includes at least one of the following: when the target domain includes the delay domain, converting the target channel information in the frequency domain to the delay domain; when the target domain includes the beam domain, converting the target channel information in the antenna domain to the beam domain; When the target domain includes the Doppler domain, the target channel information in the time domain is converted into the Doppler domain.

Optionally, the second determination module 510 determines the target feature information corresponding to the target channel information, including any one of the following: determining the statistical value of the target feature information based on the statistical value of the previous target feature information and the currently collected target feature information; determining the statistical value of the target feature information based on the average value of all target feature information collected within the first time; determining the statistical value of the target feature information based on a Gaussian mixture model GMM.

Optionally, the second determination module 520 determines whether the target AI model is applicable or not based on the target feature information, including any one of the following: when the target feature information on the target domain is within the scope of application of the target AI model, determining that the target AI model is applicable; when the target feature information on the target domain is not within the scope of application of the target AI model, determining that the target AI model is not applicable.

Optionally, the second determination module 520 determines whether the target AI model is applicable or not based on the target feature information, including: calculating the target statistic corresponding to the target feature information; determining whether the target AI model is applicable or not based on the target statistic; wherein the target statistic includes at least one of the following: mean; variance; a statistic determined based on at least one of the cumulative distribution function CDF, the probability density function PDF, and the probability mass function PMF.

Optionally, the scope of application of the target AI model is determined by feature information corresponding to the training data.

Optionally, the device also includes a reporting module, used for any of the following: when the target AI model is applicable, reporting first information, or not reporting any information, the first information being used to indicate that the target AI model is available or can work normally; when the target AI model is not applicable, reporting second information, the second information being used to indicate or request at least one of model switching, model deactivation, and enabling of a non-AI algorithm.

Optionally, the first determination module 510 is also used to: collect or count the target channel information or target feature information in a predetermined manner; wherein the predetermined manner includes at least one of the following: real-time collection or counting of the target channel information or target feature information; based on an observation period, collecting or counting the target channel information or target feature information at second intervals; collecting or counting the target channel information or target feature information located in an observation window; based on a first observation position, collecting or counting the target channel information or target feature information when the communication device moves more than a predetermined distance; based on a second observation position, collecting or counting the target channel information or target feature information when the communication device leaves a designated area, the designated area being the area where the target channel information or target feature information was last collected; based on a third observation position, collecting or counting the target channel information or target feature information when the change in the physical position of the communication device exceeds a predetermined value.

Optionally, the target characteristic information corresponding to the target channel information includes at least one of the following: spatial beam information; channel impulse response CIR; power delay spectrum PDP information; delay spread Delay spread information; Doppler information; arrival time TOA information; line-of-sight transmission LOS information; non-line-of-sight transmission NLOS information; and rank-related information.

Optionally, the spatial beam information includes at least one of a correlation between an index distribution vector of each beam and a first distribution vector, a first quantity, and a second quantity; wherein the first quantity is the number of beams corresponding to the multiple beams when the ratio of the sum of the indexes of the multiple beams to the total index of the beams reaches or exceeds a first threshold, and the second quantity is the number of beams corresponding to the single beam when the ratio of the index of the single beam to the total index of the beam reaches or exceeds a second threshold, and the The first distribution vector is a beam index distribution vector adapted to the target AI model, and the index of the beam includes the energy or power of the beam.

Optionally, the index distribution vector of each beam is obtained by shifting or cyclically shifting each beam according to the index size of each beam.

Optionally, the CIR or PDP includes at least one of the correlation between the indicator distribution vector of each path and the second distribution vector, a third quantity, a fourth quantity, the first path position, the first path indicator, the main path position, and the main path indicator; wherein, the third quantity is the number of path corresponding to the multiple path when the ratio of the sum of the indicators of multiple path to the total path indicator reaches or exceeds the third threshold, the fourth quantity is the number of path corresponding to the single path when the ratio of the indicator of a single path to the total path indicator reaches or exceeds the fourth threshold, the second distribution vector is a path distribution vector adapted to the target AI model, and the path indicators include at least one of energy, power, reference signal received power RSRP, and reference signal time difference RSTD.

Optionally, the rank-related information includes at least one of the correlation between the indicator distribution vector of each data stream and the third distribution vector, a fifth quantity, and a sixth quantity; wherein, the fifth quantity is the number of data streams corresponding to the multiple data streams when the proportion of the sum of the indicators of multiple data streams to the total indicators of the data streams reaches or exceeds the fifth threshold, and the sixth quantity is the number of data streams corresponding to the single data stream when the proportion of the indicators of a single data stream to the total indicators of the data stream reaches or exceeds the sixth threshold. The third distribution vector is a data stream distribution vector adapted to the target AI model, and the indicators of the data stream include at least one of energy, power, eigenvalues, and singular values.

Optionally, the type of the target domain and the target feature information is determined by at least one of the following: network side indication; determined according to configuration information of the target AI model; determined according to description information of the target AI model; or obtained interactively during the training process of the target AI model.

Optionally, the uses of the target AI model include at least one of the following: signal processing; signal demodulation; signal reception and transmission; channel state information acquisition; beam management; channel prediction; interference suppression; terminal positioning; prediction and management of high-level services and parameters; and analysis of control signaling.

The device 500 for determining the applicability of the model in the embodiment of the present application may be a terminal or a network-side device. By way of example, the terminal may include but is not limited to the types of the terminal 11 listed above, and the network-side device may include but is not limited to the types of the network-side device 12 listed above, etc., which are not specifically limited in the embodiment of the present application.

The device 500 for determining model applicability provided in the embodiment of the present application can implement the various processes implemented in the method embodiments of Figures 2 to 4 and achieve the same technical effects. To avoid repetition, they will not be described here.

Optionally, as shown in FIG6 , the embodiment of the present application further provides a communication device 600, including a processor 601 and a memory 602, wherein the memory 602 stores a program or instruction that can be run on the processor 601. For example, when the communication device 600 is a terminal, the program or instruction is executed by the processor 601 to implement the various steps of the above method embodiments 200-400, and can achieve the same technical effect. When the communication device 600 is a network side device, the program or instruction is executed by the processor 601 to implement the various steps of the above method embodiments 200-400, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

In one implementation, the communication device may be a terminal, which may include a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps of the method described in method embodiments 200-400. This terminal embodiment corresponds to the above-mentioned communication device side method embodiment, and each implementation process and implementation method of the above-mentioned method embodiment can be applied to the terminal embodiment and can achieve the same technical effect. Specifically, Figure 7 is a schematic diagram of the hardware structure of a terminal implementing an embodiment of the present application.

The terminal 700 includes but is not limited to: a radio frequency unit 701, a network module 702, an audio output unit 703, an input unit 704, a sensor 705, a display unit 706, a user input unit 707, an interface unit 708, a memory 709, and at least some of the components of a processor 710.

Those skilled in the art will appreciate that the terminal 700 may also include a power source (such as a battery) for supplying power to each component, and the power source may be logically connected to the processor 710 through a power management system, so as to implement functions such as managing charging, discharging, and power consumption management through the power management system. The terminal structure shown in FIG7 does not constitute a limitation on the terminal, and the terminal may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently, which will not be described in detail here.

It should be understood that in the embodiment of the present application, the input unit 704 may include a graphics processing unit (GPU) 1041 and a microphone 7042, and the graphics processor 7041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The display unit 706 may include a display panel 7061, and the display panel 7061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 707 includes a touch panel 7071 and at least one of other input devices 7072. The touch panel 7071 is also called a touch screen. The touch panel 7071 may include two parts: a touch detection device and a touch controller. Other input devices 7072 may include, but are not limited to, a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.

In the embodiment of the present application, after receiving downlink data from the network side device, the RF unit 701 can transmit the data to the processor 710 for processing; in addition, the RF unit 701 can send uplink data to the network side device. Generally, the RF unit 701 includes but is not limited to an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.

The memory 709 can be used to store software programs or instructions and various data. The memory 709 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instruction required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 709 may include a volatile memory or a non-volatile memory, or the memory 709 may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct RAM bus random access memory (DRRAM). The memory in the embodiments of the present application 709 includes, but is not limited to, these and any other suitable types of memory.

The processor 710 may include one or more processing units; optionally, the processor 710 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 710.

Among them, the processor 710 is used to determine the target feature information corresponding to the target channel information, and determine whether the target AI model is applicable or not based on the target feature information.

Optionally, the processor 710 determines the target characteristic information corresponding to the target channel information, including: determining the target characteristic information of the target channel information in a target domain; wherein the target domain includes at least one of a delay domain, a beam domain, and a Doppler domain.

Optionally, the processor 710 determines the target characteristic information of the target channel information, and also includes at least one of the following: when the target domain includes the delay domain, converting the target channel information in the frequency domain to the delay domain; when the target domain includes the beam domain, converting the target channel information in the antenna domain to the beam domain; when the target domain includes the Doppler domain, converting the target channel information in the time domain to the Doppler domain.

Optionally, the processor 710 determines the target feature information corresponding to the target channel information, including any one of the following: determining the statistical value of the target feature information based on the last calculated value of the target feature information and the currently collected target feature information; determining the statistical value of the target feature information based on the average value of all target feature information collected within the first time; determining the statistical value of the target feature information based on a Gaussian mixture model GMM.

Optionally, the processor 710 determines whether the target AI model is applicable or not based on the target feature information, including any one of the following: when the target feature information on the target domain is within the applicable scope of the target AI model, determining that the target AI model is applicable; when the target feature information on the target domain is not within the applicable scope of the target AI model, determining that the target AI model is not applicable.

Optionally, the processor 710 determines whether the target AI model is applicable or not based on the target feature information, including: calculating a target statistic corresponding to the target feature information; determining whether the target AI model is applicable or not based on the target statistic; wherein the target statistic includes at least one of the following: mean; variance; a statistic determined based on at least one of the cumulative distribution function CDF, the probability density function PDF, and the probability mass function PMF.

Optionally, the scope of application of the target AI model is determined by feature information corresponding to the training data.

Optionally, the radio frequency unit 701 is used for any of the following: when the target AI model is applicable, reporting first information, or not reporting any information, the first information being used to indicate that the target AI model is available or can work normally; when the target AI model is not applicable, reporting second information, the second information being used to indicate or request at least one of model switching, model deactivation, and enabling of a non-AI algorithm.

Optionally, the processor 710 is further configured to: collect or count the target channel information or target feature information in a predetermined manner; wherein the predetermined manner includes at least one of the following: real-time collection or counting of the target channel information or target feature information; based on an observation period, collecting or counting the target channel information or target feature information at second intervals; collecting or counting the target channel information or target feature information located within an observation window; based on a first observation position, The target channel information or target characteristic information is collected or counted when the communication device moves more than a predetermined distance; based on a second observation position, the target channel information or target characteristic information is collected or counted when the communication device leaves a designated area, and the designated area is the area where the target channel information or target characteristic information was collected last time; based on a third observation position, the target channel information or target characteristic information is collected or counted when the change in the physical position of the communication device exceeds a predetermined value.

Optionally, the target characteristic information corresponding to the target channel information includes at least one of the following: spatial beam information; channel impulse response CIR; power delay spectrum PDP information; delay spread Delay spread information; Doppler information; arrival time TOA information; line-of-sight transmission LOS information; non-line-of-sight transmission NLOS information; and rank-related information.

Optionally, the spatial beam information includes at least one of the correlation between the index distribution vector of each beam and the first distribution vector, a first quantity, and a second quantity; wherein, the first quantity is the number of beams corresponding to the multiple beams when the ratio of the sum of the indicators of multiple beams to the total beam indicators reaches or exceeds the first threshold, and the second quantity is the number corresponding to the single beam when the ratio of the indicators of a single beam to the total beam indicators reaches or exceeds the second threshold, and the first distribution vector is a beam indicator distribution vector adapted to the target AI model, and the indicators of the beam include the energy or power of the beam.

Optionally, the index distribution vector of each beam is obtained by shifting or cyclically shifting each beam according to the index size of each beam.

Optionally, the CIR or PDP includes at least one of the correlation between the indicator distribution vector of each path and the second distribution vector, a third quantity, a fourth quantity, the first path position, the first path indicator, the main path position, and the main path indicator; wherein, the third quantity is the number of path corresponding to the multiple path when the ratio of the sum of the indicators of multiple path to the total path indicator reaches or exceeds the third threshold, the fourth quantity is the number of path corresponding to the single path when the ratio of the indicator of a single path to the total path indicator reaches or exceeds the fourth threshold, the second distribution vector is a path distribution vector adapted to the target AI model, and the path indicators include at least one of energy, power, reference signal received power RSRP, and reference signal time difference RSTD.

Optionally, the rank-related information includes at least one of the correlation between the indicator distribution vector of each data stream and the third distribution vector, a fifth quantity, and a sixth quantity; wherein, the fifth quantity is the number of data streams corresponding to the multiple data streams when the proportion of the sum of the indicators of multiple data streams to the total indicators of the data streams reaches or exceeds the fifth threshold, and the sixth quantity is the number of data streams corresponding to the single data stream when the proportion of the indicators of a single data stream to the total indicators of the data stream reaches or exceeds the sixth threshold. The third distribution vector is a data stream distribution vector adapted to the target AI model, and the indicators of the data stream include at least one of energy, power, eigenvalues, and singular values.

Optionally, the type of the target domain and the target feature information is determined by at least one of the following: network side indication; determined according to configuration information of the target AI model; determined according to description information of the target AI model; or obtained interactively during the training process of the target AI model.

Optionally, the uses of the target AI model include at least one of the following: signal processing; signal demodulation; signal reception and transmission; channel state information acquisition; beam management; channel prediction; interference suppression; terminal positioning; prediction and management of high-level services and parameters; and analysis of control signaling.

In another implementation, the communication device 600 may also be a network side device, which may include a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps of the method described in embodiments 200-400. The network side device embodiment corresponds to the above communication device side method embodiment, and each implementation process and implementation method of the above method embodiment can be applied to the network side device embodiment, and can achieve the same technical effect.

Specifically, the embodiment of the present application also provides a network side device. As shown in Figure 8, the network side device 800 includes: an antenna 801, a radio frequency device 802, a baseband device 803, a processor 804 and a memory 805. The antenna 801 is connected to the radio frequency device 802. In the uplink direction, the radio frequency device 802 receives information through the antenna 801 and sends the received information to the baseband device 803 for processing. In the downlink direction, the baseband device 803 processes the information to be sent and sends it to the radio frequency device 802. The radio frequency device 802 processes the received information and sends it out through the antenna 801.

The method executed by the network-side device in the above embodiment may be implemented in the baseband device 803, which includes a baseband processor.

The baseband device 803 may include, for example, at least one baseband board, on which multiple chips are arranged, as shown in Figure 8, one of which is, for example, a baseband processor, which is connected to the memory 805 through a bus interface to call the program in the memory 805 and execute the network device operations shown in the above method embodiment.

The network side device may also include a network interface 806, which is, for example, a common public radio interface (CPRI).

Specifically, the network side device 800 of the embodiment of the present disclosure also includes: instructions or programs stored in the memory 805 and executable on the processor 804. The processor 804 calls the instructions or programs in the memory 805 to execute the methods executed by the modules shown in Figure 5 and achieve the same technical effect. To avoid repetition, it will not be repeated here.

An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned method embodiments 200-400 are implemented and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

The processor is the processor in the terminal described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run network-side device programs or instructions to implement the various processes of the above-mentioned method embodiments 200-400, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

An embodiment of the present application also provides a computer program product, which includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor. When the program or instruction is executed by the processor, each process of the above-mentioned method embodiments 200-400 is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

An embodiment of the present application also provides a communication system, including: at least one communication device, which can be used to execute the various processes of the method embodiments 200-400 as described above, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises one..." does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be noted that the scope of the method and device in the embodiment of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM/RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims (34)

A method for determining the suitability of a model, comprising: The communication device determines the target characteristic information corresponding to the target channel information; Determine whether the target AI model is applicable or not based on the target feature information. The method of claim 1, wherein the communication device determines the target characteristic information corresponding to the target channel information, comprising: The communication device determines target characteristic information of the target channel information in the target domain; The target domain includes at least one of a delay domain, a beam domain, and a Doppler domain. The method of claim 2, wherein the communication device determines the target characteristic information of the target channel information, further comprising at least one of the following: In a case where the target domain includes the delay domain, converting the target channel information in the frequency domain to the delay domain; In a case where the target domain includes the beam domain, converting target channel information of the antenna domain to the beam domain; In a case where the target domain includes the Doppler domain, target channel information in the time domain is converted into the Doppler domain. The method according to any one of claims 1 to 3, wherein the communication device determines the target characteristic information corresponding to the target channel information, including any one of the following: Determine the statistical value of the target feature information according to the statistical value of the previous target feature information and the currently collected target feature information; Determine the statistical value of the target feature information according to the average value of all target feature information collected within the first time; The statistical value of the target feature information is determined based on a Gaussian mixture model GMM. The method according to any one of claims 1 to 4, wherein determining whether the target AI model is applicable or not based on the target feature information comprises: Calculating target statistics corresponding to the target feature information; Determining whether the target AI model is applicable or not based on the target statistic; The target statistic includes at least one of the following: Mean; variance; A statistic determined based on at least one of a cumulative distribution function CDF, a probability density function PDF, and a probability mass function PMF. The method according to any one of claims 2 to 5, wherein determining whether the target AI model is applicable or not based on the target feature information comprises any one of the following: When the target feature information is within the applicable scope of the target AI model, determining that the target AI model is applicable; When the target feature information is not within the applicable scope of the target AI model, determine the target AI Model not applicable. The method according to any one of claims 1 to 6, wherein the scope of application of the target AI model is determined by feature information corresponding to the training data. The method according to any one of claims 1 to 6, wherein the method further comprises any one of the following: If the target AI model is applicable, reporting first information, or not reporting any information, wherein the first information is used to indicate that the target AI model is available or can work normally; When the target AI model is not applicable, second information is reported, where the second information is used to indicate or request at least one of model switching, model deactivation, and enabling of a non-AI algorithm. The method according to any one of claims 1 to 8, wherein the method further comprises: Collecting or counting the target channel information or target feature information in a predetermined manner; The predetermined method includes at least one of the following: Collecting or counting the target channel information or target feature information in real time; Based on the observation period, collecting or counting the target channel information or target feature information at second intervals; Collecting or counting the target channel information or target feature information located in the observation window; Based on the first observation position, collecting or counting the target channel information or target feature information when the communication device moves beyond a predetermined distance; Based on the second observation position, collecting or counting the target channel information or target feature information when the communication device leaves a designated area, wherein the designated area is an area where the target channel information or target feature information was collected last time; Based on the third observation position, the target channel information or target feature information is collected or counted when the change in the physical position of the communication device exceeds a predetermined value. The method according to any one of claims 1 to 9, wherein the target characteristic information corresponding to the target channel information includes at least one of the following: Spatial beam information; Channel impulse response CIR; Power delay profile PDP information; Delay spread Delay spread information; Doppler information; Time of arrival TOA information; Line of sight transmits LOS information; Non-line-of-sight transmission of NLOS information; Rank related information. The method of claim 10, wherein the spatial beam information comprises at least one of a correlation between the index distribution vector of each beam and the first distribution vector, a first quantity, and a second quantity; The first number is the number of beams corresponding to the multiple beams when the ratio of the sum of the indicators of the multiple beams to the total indicators of the beams reaches or exceeds the first threshold, and the second number is the ratio of the indicator of a single beam to the total indicators of the beams. When the second threshold is reached or exceeded, the number corresponding to the single beam, the first distribution vector is a beam index distribution vector adapted to the target AI model, and the index of the beam includes the energy or power of the beam. The method as claimed in claim 11, wherein the indicator distribution vector of each beam is obtained by shifting or cyclically shifting each beam according to the indicator size of each beam. The method of claim 10, wherein the CIR or PDP includes at least one of a correlation between an index distribution vector of each path and a second distribution vector, a third quantity, a fourth quantity, a first path position, a first path index, a main path position, and a main path index; Among them, the third quantity is the number of path numbers corresponding to the multiple path numbers when the ratio of the sum of the indicators of multiple path numbers to the total path indicators reaches or exceeds the third threshold; the fourth quantity is the number of path numbers corresponding to the single path when the ratio of the indicators of a single path to the total path indicators reaches or exceeds the fourth threshold; the second distribution vector is a path distribution vector adapted to the target AI model, and the path indicators include at least one of energy, power, reference signal received power RSRP, and reference signal time difference RSTD. The method of claim 10, wherein the rank-related information comprises at least one of a correlation between the indicator distribution vector of each data stream and the third distribution vector, a fifth quantity, and a sixth quantity; Among them, the fifth number is the number of data streams corresponding to the multiple data streams when the ratio of the sum of the indicators of multiple data streams to the total indicators of the data streams reaches or exceeds the fifth threshold; the sixth number is the number of data streams corresponding to the single data stream when the ratio of the indicators of a single data stream to the total indicators of the data streams reaches or exceeds the sixth threshold; the third distribution vector is a data stream distribution vector adapted to the target AI model, and the indicators of the data stream include at least one of energy, power, eigenvalues, and singular values. The method according to any one of claims 1 to 14, wherein the type of the target domain and the target feature information is determined by at least one of the following: Network side indication; Determined according to the configuration information of the target AI model; Determined according to the description information of the target AI model; It is obtained interactively during the training process of the target AI model. The method according to any one of claims 1 to 15, wherein the purpose of the target AI model includes at least one of the following: Signal processing; Signal demodulation; Signal transmission and reception; Channel state information acquisition; Beam management; Channel prediction; Interference suppression; Terminal positioning; Prediction and management of high-level business,parameters; Parsing of control signaling. An apparatus for determining the suitability of a model, comprising: A determination module is used to determine target feature information corresponding to the target channel information; and determine whether the target AI model is applicable or not based on the target feature information. The apparatus according to claim 17, wherein the determining module determines the target characteristic information corresponding to the target channel information, comprising: Determining target feature information of target channel information in a target domain; The target domain includes at least one of a delay domain, a beam domain, and a Doppler domain. The apparatus of claim 18, wherein the determining module determines the target characteristic information of the target channel information, further comprising at least one of the following: In a case where the target domain includes the delay domain, converting the target channel information in the frequency domain to the delay domain; In a case where the target domain includes the beam domain, converting target channel information of the antenna domain to the beam domain; In a case where the target domain includes the Doppler domain, target channel information in the time domain is converted into the Doppler domain. The apparatus according to any one of claims 17 to 19, wherein the determination module determines the target characteristic information corresponding to the target channel information, including any one of the following: Determine the statistical value of the target feature information according to the statistical value of the previous target feature information and the currently collected target feature information; Determine the statistical value of the target feature information according to the average value of all target feature information collected within the first time; The statistical value of the target feature information is determined based on a Gaussian mixture model GMM. The device according to any one of claims 17 to 20, wherein the determination module determines whether the target AI model is applicable or not based on the target feature information, comprising: Calculating target statistics corresponding to the target feature information; Determining whether the target AI model is applicable or not based on the target statistic; The target statistic includes at least one of the following: Mean; variance; A statistic determined based on at least one of a cumulative distribution function CDF, a probability density function PDF, and a probability mass function PMF. The device according to any one of claims 17 to 21, wherein the determination module determines whether the target AI model is applicable or not applicable according to the target feature information, including any one of the following: When the target feature information is within the applicable scope of the target AI model, determining that the target AI model is applicable; When the target feature information is not within the applicable scope of the target AI model, it is determined that the target AI model is not applicable. The device as described in any one of claims 17-22, wherein the scope of application of the target AI model is determined by feature information corresponding to the training data. The apparatus according to any one of claims 17 to 22, wherein the apparatus further comprises a reporting module configured to: If the target AI model is applicable, reporting first information, or not reporting any information, wherein the first information is used to indicate that the target AI model is available or can work normally; When the target AI model is not applicable, second information is reported, where the second information is used to indicate or request at least one of model switching, model deactivation, and enabling of a non-AI algorithm. The device according to any one of claims 17 to 24, wherein the determination module is further used to: collect or count the target channel information or target feature information in a predetermined manner; The predetermined method includes at least one of the following: Collecting or counting the target channel information or target feature information in real time; Based on the observation period, collecting or counting the target channel information or target feature information at second intervals; Collecting or counting the target channel information or target feature information located in the observation window; Based on the first observation position, collecting or counting the target channel information or target feature information when the communication device moves beyond a predetermined distance; Based on the second observation position, collecting or counting the target channel information or target feature information when the communication device leaves a designated area, wherein the designated area is an area where the target channel information or target feature information was collected last time; Based on the third observation position, the target channel information or target feature information is collected or counted when the change in the physical position of the communication device exceeds a predetermined value. The apparatus according to any one of claims 17 to 25, wherein the target characteristic information corresponding to the target channel information includes at least one of the following: Spatial beam information; Channel impulse response CIR; Power delay profile PDP information; Delay spread Delay spread information; Doppler information; Time of arrival TOA information; Line of sight transmits LOS information; Non-line-of-sight transmission of NLOS information; Rank related information. The apparatus of claim 26, wherein the spatial beam information comprises at least one of a correlation between the indicator distribution vector of each beam and the first distribution vector, a first quantity, and a second quantity; Among them, the first number is the number of beams corresponding to the multiple beams when the ratio of the sum of the indicators of multiple beams to the total indicators of the beams reaches or exceeds the first threshold; the second number is the number corresponding to the single beam when the ratio of the indicator of a single beam to the total indicators of the beams reaches or exceeds the second threshold; the first distribution vector is a beam indicator distribution vector adapted to the target AI model, and the indicators of the beam include the energy or power of the beam. The device as claimed in claim 27, wherein the indicator distribution vector of each beam is obtained by shifting or cyclically shifting each beam according to the indicator size of each beam. The apparatus of claim 26, wherein the CIR or PDP comprises at least one of a correlation between an index distribution vector of each path and a second distribution vector, a third quantity, a fourth quantity, a first path position, a first path index, a main path position, and a main path index; Among them, the third quantity is the number of path numbers corresponding to the multiple path numbers when the ratio of the sum of the indicators of multiple path numbers to the total path indicators reaches or exceeds the third threshold; the fourth quantity is the number of path numbers corresponding to the single path when the ratio of the indicators of a single path to the total path indicators reaches or exceeds the fourth threshold; the second distribution vector is a path distribution vector adapted to the target AI model, and the path indicators include at least one of energy, power, reference signal received power RSRP, and reference signal time difference RSTD. The apparatus of claim 26, wherein the rank-related information comprises at least one of a correlation between the indicator distribution vector of each data stream and the third distribution vector, a fifth quantity, and a sixth quantity; Among them, the fifth number is the number of data streams corresponding to the multiple data streams when the ratio of the sum of the indicators of multiple data streams to the total indicators of the data streams reaches or exceeds the fifth threshold; the sixth number is the number of data streams corresponding to the single data stream when the ratio of the indicators of a single data stream to the total indicators of the data streams reaches or exceeds the sixth threshold; the third distribution vector is a data stream distribution vector adapted to the target AI model, and the indicators of the data stream include at least one of energy, power, eigenvalues, and singular values. The apparatus according to any one of claims 17 to 30, wherein the target domain and the type of the target feature information are determined by at least one of the following: Network side indication; Determined according to the configuration information of the target AI model; Determined according to the description information of the target AI model; It is obtained interactively during the training process of the target AI model. The device according to any one of claims 17 to 31, wherein the purpose of the target AI model includes at least one of the following: Signal processing; Signal demodulation; Signal transmission and reception; Channel state information acquisition; Beam management; Channel prediction; Interference suppression; Terminal positioning; Prediction and management of high-level business,parameters; Parsing of control signaling. A communication device comprises a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of the method according to any one of claims 1 to 16 are implemented. A readable storage medium stores a program or instruction, and when the program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 16 are implemented.