[go: up one dir, main page]

WO2023279366A1 - Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage - Google Patents

Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage Download PDF

Info

Publication number
WO2023279366A1
WO2023279366A1 PCT/CN2021/105462 CN2021105462W WO2023279366A1 WO 2023279366 A1 WO2023279366 A1 WO 2023279366A1 CN 2021105462 W CN2021105462 W CN 2021105462W WO 2023279366 A1 WO2023279366 A1 WO 2023279366A1
Authority
WO
WIPO (PCT)
Prior art keywords
noise reduction
reference signal
measurement value
data set
signal measurement
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2021/105462
Other languages
English (en)
Chinese (zh)
Inventor
刘文东
田文强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Oppo Mobile Telecommunications Corp Ltd
Original Assignee
Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Oppo Mobile Telecommunications Corp Ltd filed Critical Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority to PCT/CN2021/105462 priority Critical patent/WO2023279366A1/fr
Priority to CN202180095342.4A priority patent/CN116941185A/zh
Publication of WO2023279366A1 publication Critical patent/WO2023279366A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
    • H04B1/06—Receivers
    • H04B1/10—Means associated with receiver for limiting or suppressing noise or interference
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence

Definitions

  • the present application relates to the communication field, and in particular to a transfer learning-based noise reduction method, terminal equipment, network equipment, and storage media.
  • noise reduction networks based on artificial intelligence (AI) are trained and deployed in specific signal-to-noise ratio scenarios, and their generalization performance is not good, so they are subject to many limitations in practical applications.
  • AI artificial intelligence
  • the embodiment of the present application provides a noise reduction method based on migration learning, a terminal device, a network device, and a storage medium, which are used to propose a noise reduction model based on migration training, so that the noise reduction model in downlink or uplink transmission can be adapted to correspond to The variable reference signal measurement value in the link environment achieves good noise reduction effect.
  • the first aspect of the embodiments of the present application provides a method for noise reduction based on transfer learning, which may include: a terminal device acquires a noise reduction model based on transfer learning; and the terminal device performs noise reduction processing according to the noise reduction model.
  • the second aspect of the embodiments of the present application provides a noise reduction method based on transfer learning, which may include: a network device acquires a current reference signal measurement value; the network device obtains a current reference signal measurement value and a preset reference signal measurement value A set of intervals is used to acquire a noise reduction model corresponding to the measured value of the current reference signal, or a target data set, and the noise reduction model or the target data set is used for noise reduction processing.
  • the third aspect of the embodiments of the present invention provides a terminal device with a noise reduction model based on migration training, so that the noise reduction model in data transmission can adapt to the measured value of the reference signal corresponding to the change in the link environment, and achieve good noise reduction.
  • Noise effect function This function may be implemented by hardware, or may be implemented by executing corresponding software on the hardware.
  • the hardware or software includes one or more modules corresponding to the above functions.
  • the fourth aspect of the embodiments of the present invention provides a network device with a noise reduction model based on migration training, so that the noise reduction model in data transmission can adapt to the measured value of the reference signal that changes in the corresponding link environment, and achieve good noise reduction.
  • Noise effect function This function may be implemented by hardware, or may be implemented by executing corresponding software on the hardware.
  • the hardware or software includes one or more modules corresponding to the above functions.
  • a terminal device including: a memory storing executable program codes; a transceiver and a processor coupled to the memory; the processor and the transceiver are used to execute the implementation of the present invention Example of the method described in the first aspect.
  • Another aspect of the embodiments of the present invention provides a network device, including: a memory storing executable program codes; a processor coupled to the memory; the processor is used to execute the method described in the second aspect of the embodiments of the present invention method.
  • Still another aspect of the embodiments of the present invention provides a computer-readable storage medium, including instructions, which, when run on a computer, cause the computer to execute the method described in the first aspect or the second aspect of the present invention.
  • Another aspect of the embodiments of the present invention provides a chip, the chip is coupled with the memory in the terminal device, so that the chip calls the program instructions stored in the memory during operation, so that the terminal device executes the The method described in the first aspect or the second aspect of the invention.
  • the terminal device acquires a noise reduction model based on transfer learning; the terminal device performs noise reduction processing according to the noise reduction model.
  • the terminal device proposes a noise reduction model based on migration training, so that the noise reduction model in data transmission can adapt to the measured value of the reference signal that changes in the corresponding link environment, and achieves a good noise reduction effect.
  • Fig. 1 is a schematic diagram of a process from sending to receiving from a source in an implementation
  • Fig. 2 is a schematic diagram of neuron structure
  • Fig. 3 is a schematic diagram of neural network
  • Fig. 4 is a schematic diagram of the basic structure of a convolutional neural network
  • Fig. 5 is a schematic diagram of the transfer learning process
  • FIG. 6A is a system architecture diagram of a communication system applied in an embodiment of the present invention.
  • FIG. 6B is a schematic diagram of an embodiment of a noise reduction method based on transfer learning in the embodiment of the present application.
  • FIG. 7 is a schematic diagram of another embodiment of a noise reduction method based on transfer learning in the embodiment of the present application.
  • FIG. 8A is a schematic diagram of a wireless communication system receiver including a noise reduction model in an embodiment of the present application
  • Fig. 8B is a schematic diagram of the fully connected denoising model in the embodiment of the present application.
  • FIG. 8C is a schematic diagram of migration training in the embodiment of the present application.
  • FIG. 8D is a schematic diagram of another embodiment of the noise reduction method based on transfer learning in the embodiment of the present application.
  • FIG. 9 is a schematic diagram of another embodiment of a noise reduction method based on transfer learning in the embodiment of the present application.
  • FIG. 10 is a schematic diagram of another embodiment of the noise reduction method based on transfer learning in the embodiment of the present application.
  • FIG. 11 is a schematic diagram of transfer learning and updating of the noise reduction model of the terminal device on the network device side in the embodiment of the present application;
  • FIG. 12 is a schematic diagram of an embodiment of a terminal device in the embodiment of the present application.
  • FIG. 13 is a schematic diagram of an embodiment of a network device in the embodiment of the present application.
  • FIG. 14 is a schematic diagram of another embodiment of a terminal device in an embodiment of the present invention.
  • FIG. 15 is a schematic diagram of another embodiment of a network device in the embodiment of the present application.
  • the basic workflow is that the transmitter performs operations such as coding, modulation, and encryption on the information source at the sending end to form the sending information to be transmitted.
  • the sending information to be transmitted is transmitted to the receiving end through the wireless space, and the receiving end performs operations such as decoding, decryption and demodulation on the received receiving information, and finally recovers the source information, as shown in Fig.
  • the encoding, modulation, encryption, decoding, demodulation, decryption and other operations of the sending end and receiving end are controllable, but the channel conditions and noise conditions in the wireless space environment are uncontrollable, complex and varied. changing.
  • the channel conditions and noise conditions in the wireless space environment are uncontrollable, complex and varied. changing.
  • For the interference noise in the wireless space there is a corresponding lack of necessary processing solutions, and the recovery of the signal source under different signal-to-noise ratios will show a large difference.
  • a neural network is an operational model composed of multiple neuron nodes connected to each other, in which the connection between nodes represents the weighted value from the input signal to the output signal, called weight; each node performs weighted summation of different input signals , and output through a specific activation function.
  • FIG 2 it is a schematic diagram of the neuron structure.
  • Figure 3 it is a schematic diagram of the neural network.
  • the neural network includes an input layer, a hidden layer, and an output layer. Through different connection methods, weights, and activation functions of multiple neurons, different outputs can be generated, and then the mapping relationship from input to output can be fitted.
  • Deep learning uses a deep neural network with multiple hidden layers, which greatly improves the ability of the network to learn features, and can fit complex nonlinear mappings from input to output, so it is widely used in the fields of speech and image processing.
  • deep learning also includes common basic structures such as convolutional neural network (CNN), recurrent neural network (Recurrent Neural Network, RNN).
  • CNN convolutional neural network
  • RNN Recurrent Neural Network
  • the basic structure of a convolutional neural network includes: an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer.
  • Figure 4 it is a schematic diagram of the basic structure of a convolutional neural network.
  • Each neuron of the convolution kernel in the convolution layer is locally connected to its input, and the local maximum or average feature of a certain layer is extracted by introducing a pooling layer, which effectively reduces the parameters of the network and mines local features. It enables the convolutional neural network to converge quickly and obtain excellent performance.
  • Transfer learning can use the similarity between data, tasks or models to apply the models and knowledge learned in the old field to the new field.
  • Figure 5 shows a schematic diagram of the transfer learning process.
  • the models A and B constructed by data set/task A and data set/task B can be fused through some migration methods, and then Apply the migration fusion model to the new data set/task C to complete the application on the data set/task C.
  • datasets A and B can be called the source domain of transfer learning
  • dataset C can be called the target domain of transfer learning.
  • the data set of the source domain is usually labeled, while the data set of the target domain is usually unlabeled, so the transfer learning is trained on the source domain, after obtaining the initial model, and evaluating the similarity between the target domain and the source domain by adding The degree of loss function or the way of adversarial migration can train the source domain model to be suitable for the target domain and complete the task on the target domain.
  • noise reduction networks based on artificial intelligence (AI) are trained and deployed in specific signal-to-noise ratio scenarios, and their generalization performance is not good, so they are subject to many limitations in practical applications. Therefore, it is of great significance to design a noise reduction network with generalization performance in multiple SNR scenarios.
  • AI artificial intelligence
  • the technical solution of the embodiment of the present application can be applied to various communication systems, such as: Global System of Mobile communication (Global System of Mobile communication, GSM) system, code division multiple access (Code Division Multiple Access, CDMA) system, broadband code division multiple access (Wideband Code Division Multiple Access, WCDMA) system, General Packet Radio Service (GPRS), Long Term Evolution (LTE) system, Advanced long term evolution (LTE-A) system , New Radio (NR) system, evolution system of NR system, LTE (LTE-based access to unlicensed spectrum, LTE-U) system on unlicensed spectrum, NR (NR-based access to unlicensed spectrum) on unlicensed spectrum unlicensed spectrum (NR-U) system, Non-Terrestrial Networks (NTN) system, Universal Mobile Telecommunications System (UMTS), Wireless Local Area Networks (WLAN), Wireless Fidelity (Wireless Fidelity, WiFi), fifth-generation communication (5th-Generation, 5G) system or other communication systems, etc.
  • GSM Global System of Mobile
  • D2D Device to Device
  • M2M Machine to Machine
  • MTC Machine Type Communication
  • V2V Vehicle to Vehicle
  • V2X Vehicle to everything
  • the communication system in the embodiment of the present application may be applied to a carrier aggregation (Carrier Aggregation, CA) scenario, may also be applied to a dual connectivity (Dual Connectivity, DC) scenario, and may also be applied to an independent (Standalone, SA) deployment Web scene.
  • Carrier Aggregation, CA Carrier Aggregation
  • DC Dual Connectivity
  • SA independent deployment Web scene
  • the communication system in the embodiment of the present application may be applied to an unlicensed spectrum, where the unlicensed spectrum may also be considered as a shared spectrum; or, the communication system in the embodiment of the present application may also be applied to a licensed spectrum, where, Licensed spectrum can also be considered as non-shared spectrum.
  • the embodiments of the present application describe various embodiments in conjunction with network equipment and terminal equipment, wherein the terminal equipment may also be referred to as user equipment (User Equipment, UE), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device, etc.
  • user equipment User Equipment, UE
  • access terminal user unit
  • user station mobile station
  • mobile station mobile station
  • remote station remote terminal
  • mobile device user terminal
  • terminal wireless communication device
  • wireless communication device user agent or user device
  • the terminal device can be a station (STAION, ST) in the WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (Session Initiation Protocol, SIP) phone, a wireless local loop (Wireless Local Loop, WLL) station, a personal digital processing (Personal Digital Assistant, PDA) devices, handheld devices with wireless communication functions, computing devices or other processing devices connected to wireless modems, vehicle-mounted devices, wearable devices, next-generation communication systems such as terminal devices in NR networks, or future Terminal equipment in the evolved public land mobile network (Public Land Mobile Network, PLMN) network, etc.
  • STAION, ST Session Initiation Protocol
  • SIP Session Initiation Protocol
  • WLL Wireless Local Loop
  • PDA Personal Digital Assistant
  • the terminal device can be deployed on land, including indoor or outdoor, handheld, wearable or vehicle-mounted; it can also be deployed on water (such as ships, etc.); it can also be deployed in the air (such as aircraft, balloons and satellites) superior).
  • the terminal device may be a mobile phone (Mobile Phone), a tablet computer (Pad), a computer with a wireless transceiver function, a virtual reality (Virtual Reality, VR) terminal device, an augmented reality (Augmented Reality, AR) terminal Equipment, wireless terminal equipment in industrial control, wireless terminal equipment in self driving, wireless terminal equipment in remote medical, wireless terminal equipment in smart grid , wireless terminal equipment in transportation safety, wireless terminal equipment in smart city, or wireless terminal equipment in smart home.
  • a virtual reality (Virtual Reality, VR) terminal device an augmented reality (Augmented Reality, AR) terminal Equipment
  • wireless terminal equipment in industrial control wireless terminal equipment in self driving
  • wireless terminal equipment in remote medical wireless terminal equipment in smart grid
  • wireless terminal equipment in transportation safety wireless terminal equipment in smart city, or wireless terminal equipment in smart home.
  • the terminal device may also be a wearable device.
  • Wearable devices can also be called wearable smart devices, which is a general term for the application of wearable technology to intelligently design daily wear and develop wearable devices, such as glasses, gloves, watches, clothing and shoes.
  • a wearable device is a portable device that is worn directly on the body or integrated into the user's clothing or accessories. Wearable devices are not only a hardware device, but also achieve powerful functions through software support, data interaction, and cloud interaction.
  • Generalized wearable smart devices include full-featured, large-sized, complete or partial functions without relying on smart phones, such as smart watches or smart glasses, etc., and only focus on a certain type of application functions, and need to cooperate with other devices such as smart phones Use, such as various smart bracelets and smart jewelry for physical sign monitoring.
  • the network device may be a device for communicating with the mobile device, and the network device may be an access point (Access Point, AP) in WLAN, a base station (Base Transceiver Station, BTS) in GSM or CDMA , or a base station (NodeB, NB) in WCDMA, or an evolved base station (Evolutional Node B, eNB or eNodeB) in LTE, or a relay station or access point, or a vehicle-mounted device, a wearable device, and an NR network
  • BTS Base Transceiver Station
  • NodeB, NB base station
  • Evolutional Node B, eNB or eNodeB evolved base station
  • LTE Long Term Evolutional Node B, eNB or eNodeB
  • gNB network equipment in the network or the network equipment in the future evolved PLMN network or the network equipment in the NTN network, etc.
  • the network device may have a mobile feature, for example, the network device may be a mobile device.
  • the network equipment may be a satellite or a balloon station.
  • the satellite can be a low earth orbit (low earth orbit, LEO) satellite, a medium earth orbit (medium earth orbit, MEO) satellite, a geosynchronous earth orbit (geosynchronous earth orbit, GEO) satellite, a high elliptical orbit (High Elliptical Orbit, HEO) satellite. ) Satellite etc.
  • the network device may also be a base station installed on land, water, and other locations.
  • the network device may provide services for a cell, and the terminal device communicates with the network device through the transmission resources (for example, frequency domain resources, or spectrum resources) used by the cell, and the cell may be a network device ( For example, a cell corresponding to a base station), the cell may belong to a macro base station, or may belong to a base station corresponding to a small cell (Small cell), and the small cell here may include: a metro cell (Metro cell), a micro cell (Micro cell), a pico cell ( Pico cell), Femto cell, etc. These small cells have the characteristics of small coverage and low transmission power, and are suitable for providing high-speed data transmission services.
  • the transmission resources for example, frequency domain resources, or spectrum resources
  • the cell may be a network device (
  • the cell may belong to a macro base station, or may belong to a base station corresponding to a small cell (Small cell)
  • the small cell here may include: a metro cell (Metro cell), a micro cell (Micro
  • the communication system may include a network device, and the network device may be a device for communicating with a terminal device (or called a communication terminal, terminal).
  • a network device can provide communication coverage for a specific geographic area, and can communicate with terminal devices located within the coverage area.
  • Figure 6A exemplarily shows one network device and two terminal devices.
  • the communication system may include multiple network devices and each network device may include other numbers of terminal devices within the coverage area. Examples are not limited to this.
  • the communication system may further include other network entities such as a network controller and a mobility management entity, which is not limited in this embodiment of the present application.
  • the network equipment may further include access network equipment and core network equipment. That is, the wireless communication system also includes multiple core networks for communicating with access network devices.
  • the access network device may be a long-term evolution (long-term evolution, LTE) system, a next-generation (mobile communication system) (next radio, NR) system or an authorized auxiliary access long-term evolution (LAA- Evolved base station (evolutional node B, abbreviated as eNB or e-NodeB) macro base station, micro base station (also called “small base station”), pico base station, access point (access point, AP), Transmission point (transmission point, TP) or new generation base station (new generation Node B, gNodeB), etc.
  • LTE long-term evolution
  • NR next-generation
  • LAA- Evolved base station evolutional node B, abbreviated as eNB or e-NodeB
  • eNB next-generation
  • NR next-generation
  • a device with a communication function in the network/system in the embodiment of the present application may be referred to as a communication device.
  • the communication equipment may include network equipment and terminal equipment with communication functions, and the network equipment and terminal equipment may be the specific equipment described in the embodiments of the present invention, which will not be repeated here; communication
  • the device may also include other devices in the communication system, such as network controllers, mobility management entities and other network entities, which are not limited in this embodiment of the present application.
  • FIG. 6B it is a schematic diagram of an embodiment of a noise reduction method based on transfer learning in the embodiment of the present application, which may include:
  • the terminal device acquires a noise reduction model based on transfer learning.
  • the terminal device obtains a noise reduction model based on transfer learning, which may include:
  • the terminal device performs model training according to the data set to obtain a noise reduction model based on transfer learning; or,
  • the terminal device receives the noise reduction model based on transfer learning delivered by the network device.
  • the terminal device performs model training according to the data set to obtain a noise reduction model based on transfer learning, which may include: the terminal device obtains the source domain data set, the label corresponding to the source domain data set, and Target domain data set; the terminal device performs model training to obtain a noise reduction model according to the source domain data set, the label corresponding to the source domain data set, and the target domain data set.
  • the terminal device performs model training to obtain the noise reduction model according to the source domain dataset, the label corresponding to the source domain dataset, and the target domain dataset, which may include: set to determine the joint loss function; the terminal device performs model training according to the joint loss function to obtain a noise reduction model.
  • the terminal device determines the joint loss function according to the source domain data set, the label and the target domain data set, which may include: the terminal device determines the error loss function according to the source domain data set and the label; that is, by combining the source domain data set and The labels are optimized so that the optimization error loss function reaches convergence.
  • the adaptation loss function is determined; the terminal device determines the joint loss function according to the adaptation loss function and the error loss function.
  • the terminal device performs model training according to the data set to obtain a noise reduction model based on transfer learning, which may include: the terminal device receives the network device update instruction and the current reference signal measurement value according to the noise reduction model The target data set or a subset of the target data set is sent, and model training is performed according to the target data set or a subset of the target data set to obtain a noise reduction model.
  • the acquisition of the transfer learning-based noise reduction model by the terminal device may include: the terminal device receiving the noise reduction model sent by the network device according to the noise reduction model update instruction and the current reference signal measurement value.
  • the noise reduction model obtained by the terminal device may be a noise reduction model obtained by the terminal device through model training according to the source domain data set obtained by the terminal device, the label corresponding to the source domain data set, and the target domain data set; It can also be that the terminal device receives the noise reduction model delivered by the network device; it can also be that the terminal device performs model training according to the target data set or a subset of the target data set delivered by the network device, and the obtained noise reduction model can also be other
  • the denoising model obtained by the method is not specifically limited here.
  • the terminal device performs noise reduction processing according to the noise reduction model.
  • the terminal device performs noise reduction processing on the downlink according to the noise reduction model.
  • the terminal device obtains a noise reduction model based on transfer learning; the terminal device performs noise reduction processing according to the noise reduction model.
  • the terminal device proposes a noise reduction model based on migration training, so that the noise reduction model in data transmission can adapt to the measured value of the reference signal that changes in the corresponding link environment, and achieves a good noise reduction effect.
  • FIG. 7 it is a schematic diagram of another embodiment of the noise reduction method based on transfer learning in the embodiment of the present application, which may include:
  • the terminal device acquires the noise reduction model based on transfer learning, which may include but not limited to the following steps 701-703, as follows:
  • the terminal device acquires a source domain dataset, a label corresponding to the source domain dataset, and a target domain dataset.
  • the terminal device obtains the source domain data set, the label corresponding to the source domain data set, and the target domain data set, which may include: when the measured value of the current reference signal measured by the terminal device meets the preset condition, the terminal device Obtain the source domain dataset, the label corresponding to the source domain dataset (also referred to as label data), and the target domain dataset.
  • the current reference signal measurement value measured by the terminal device is a downlink current reference signal measurement value measured by the terminal device.
  • the current reference signal measurement value includes Reference Signal Received Power (Reference Signal Received Power, RSRP), Reference Signal Received Quality (Reference Signal Received Quality, RSRQ), Received Signal Strength Indicator (Received Signal Strength Indicator, RSSI), and At least one of the Signal-to Interference plus Noise Ratio (SINR).
  • RSRP Reference Signal Received Power
  • RSRQ Reference Signal Received Quality
  • RSSI Received Signal Strength Indicator
  • SINR Signal-to Interference plus Noise Ratio
  • the current measured value of the reference signal satisfies a preset condition
  • the preset condition can trigger an update of the noise reduction model.
  • the preset condition may be that the current signal measurement value is greater than a first preset threshold, or, the absolute value of the difference between the current signal measurement value and a signal measurement value adapted to a known noise reduction model is greater than a second preset threshold, or, the current The absolute value of the ratio of the signal measurement value to the signal measurement value adapted to the known noise reduction model satisfies a threshold range and the like.
  • this embodiment of the present application may be applied to a deep neural network, a recurrent neural network, or a convolutional neural network, or other neural networks.
  • FIG. 7 is an embodiment of a method for designing a downlink transmission migration noise reduction model (also called a noise reduction network).
  • This embodiment provides a method for a terminal device to design a noise reduction model by using transfer learning during downlink transmission.
  • FIG. 8A it is a schematic diagram of a wireless communication system receiver including a noise reduction model in the embodiment of the present application.
  • the input of the noise reduction model is the received information after the channel and the noise
  • the output is the received information after the noise reduction processing.
  • the noise reduction model can be implemented in various ways such as fully connected deep neural network (Deep Neural Network, DNN), CNN or RNN, which is not specifically limited in this embodiment.
  • DNN Deep Neural Network
  • CNN CNN
  • RNN RNN
  • the input is the original received information, that is, the source domain data set
  • the label is the received information without noise.
  • MSE mean square error between the output of the noise reduction model and the label
  • SNR Signal-to-noise Ratio
  • the existing noise reduction model can be retrained on the new data set.
  • this embodiment proposes to use transfer learning from the source domain to the target domain to adapt the noise reduction model in the downlink transmission process to the changing SNR.
  • FIG. 8B it is a schematic diagram of the fully connected denoising model in the embodiment of the present application.
  • the noise reduction model trained on the source domain dataset A can input the received data to be denoised during the deployment process, and output the denoised received data after the Nth layer .
  • the link quality changes, the signal-to-noise ratio changes greatly, or the terminal equipment moves to a new cell and the existing noise reduction model is not suitable, it is necessary to perform migration learning on the new data set for the existing noise reduction model.
  • FIG. 8C is a schematic diagram of migration training in the embodiment of the present application.
  • the optimization of the MSE loss function on the source domain data set A can make the noise reduction model converge on the source domain data set A.
  • the target domain dataset B and the source domain dataset A can be Achieve adaptation.
  • the goal of the migration learning training process is to minimize the joint loss function, so that the noise reduction model can learn the noise and channel features on the source domain data set A, so that the data set A and the data set B can pass through the adaptation layer Finally, good results are also obtained on the target domain dataset B.
  • the terminal device determines a joint loss function according to the source domain dataset, the label, and the target domain dataset.
  • the terminal device determines the joint loss function according to the source domain data set, the label and the target domain data set, which may include: the terminal device determines the error loss function according to the source domain data set and the label; that is, by combining the source domain data set and The labels are optimized so that the optimization error loss function reaches convergence.
  • the adaptation loss function is determined; the terminal device determines the joint loss function according to the adaptation loss function and the error loss function.
  • the terminal device determines the joint loss function according to the adaptation loss function and the error loss function, which may include: the terminal device determines the joint loss function according to the first formula;
  • the error loss function may include a mean square error loss function, or other error loss functions, which are not specifically limited here.
  • the terminal device performs model training according to the joint loss function to obtain a noise reduction model.
  • the training ends.
  • FIG. 8D it is a schematic diagram of another embodiment of the noise reduction method based on transfer learning in the embodiment of the present application.
  • the network device takes the base station as an example.
  • the terminal device measures at least one of RSRP, RSRQ, RSSI, and SINR on the downlink data resources delivered by the network device.
  • the preset conditions for model update for example: when the absolute value of the difference between the measured RSRP and the RSRP adapted to the known noise reduction model is greater than a certain preset threshold, the source domain dataset download of the noise reduction model is reported to the network device instruct.
  • the download of the source domain dataset A includes the received dataset to be denoised and the corresponding labels.
  • only a subset of the source domain dataset A can be downloaded for adaptation training; at the same time, the terminal device collects the target domain dataset B to be migrated for training.
  • the terminal device can obtain a joint loss function according to the first formula above, and then perform model training according to the joint loss function to obtain an updated noise reduction model. For example: the number of model training reaches the preset number, or after the joint loss function corresponding to the noise reduction model obtained through model training reaches the preset value, the training is completed, and the noise reduction model obtained at this time is the updated noise reduction model .
  • the terminal device performs noise reduction processing according to the noise reduction model.
  • the terminal device performs noise reduction processing on the downlink according to the noise reduction model.
  • the terminal device obtains the source domain data set, the label corresponding to the source domain data set, and the target domain data set; the terminal device determines the joint loss function according to the source domain data set, the label and the target domain data set; the terminal The device performs model training according to the joint loss function to obtain a noise reduction model; the terminal device performs noise reduction processing according to the noise reduction model.
  • the embodiment of the present application proposes a noise reduction model for migration training, so that the noise reduction model in downlink transmission can adapt to the changing reference signal measurement value in the downlink environment, and achieves a good noise reduction effect.
  • FIG. 9 it is a schematic diagram of another embodiment of the noise reduction method based on transfer learning in the embodiment of the present application, which may include:
  • the network device acquires a current reference signal measurement value.
  • the acquisition of the current reference signal measurement value by the network device may include: the current reference signal measurement value measured by the network device. It can be understood that the current reference signal measurement value measured by the network device is an uplink current reference signal measurement value measured by the network device.
  • the embodiment of the present application provides a method for designing a noise reduction model group by a network device (such as a base station) using migration learning during uplink transmission.
  • a network device such as a base station
  • different terminal devices determine whether to perform transfer learning based on their own downlink quality measurement information such as RSRP/RSRQ/RSSI/SINR.
  • the noise reduction model is configured on the network device side.
  • the network device needs to receive uplink transmission data from different terminal devices, and the link conditions of different terminal devices vary widely, multiple different noise reduction models are configured on the network device. The complexity is relatively high, therefore, the embodiment of the present application adopts the configuration method of the noise reduction model group for the configuration of the noise reduction model for uplink transmission.
  • this embodiment provides a method for configuring pre-grouping of noise reduction models of network devices during uplink transmission.
  • the network device configures the computing resources of K noise reduction models, and the parameter set of the kth noise reduction model is denoted as ⁇ k .
  • the parameter configuration and training target of the noise reduction model can be obtained by the second formula:
  • the second formula is:
  • the network device measures the uplink reference signal to obtain the measured value of the current reference signal, such as the signal-to-noise ratio of the current reference signal.
  • X' represents the received data after noise reduction
  • X represents the noise-free label data
  • ⁇ SNR k ⁇ represents the signal-to-noise ratio interval centered on SNR k with a width of ⁇ , which can be expressed as
  • the signal-to-noise ratio intervals ⁇ SNR k ⁇ corresponding to each noise reduction model do not overlap.
  • the noise reduction model group can be trained and deployed offline, and no online update of the noise reduction model group is required. Chemical.
  • the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the current reference signal measurement value and the preset reference signal measurement value interval set, and/or, the target data set, the noise reduction model or the target data set is used for performing Noise reduction processing.
  • the noise reduction model is a noise reduction model related to the uplink.
  • the current reference signal measurement value belongs to the target reference signal measurement value interval in the preset reference signal measurement value interval set
  • the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the interval set of the current reference signal measurement value and the preset reference signal measurement value, which may include: the network device determines that the current reference signal measurement value belongs to the preset reference signal In the case of the target reference signal measurement value interval in the measurement value interval set, the noise reduction model corresponding to the current reference signal measurement value is acquired according to the target reference signal measurement value interval. It can be understood that the network device pre-stores noise reduction models corresponding to different reference signal measurement value intervals, and/or target data sets.
  • the network device acquires the noise reduction model corresponding to the current reference signal measurement value according to the target reference signal measurement value interval, which may include:
  • the network device searches for the target noise reduction model corresponding to the target reference signal measurement value interval, and uses it as the noise reduction model corresponding to the current reference signal measurement value. It can be understood that if the network device pre-stores noise reduction models corresponding to different reference signal measurement value intervals, the network device can directly search for the target noise reduction model corresponding to the target reference signal measurement value interval as an updated noise reduction model. and / or,
  • the network device searches for the target data set corresponding to the target reference signal measurement value range, performs model training according to the target data set or a subset of the target data set, and obtains the noise reduction model corresponding to the current reference signal measurement value. It can be understood that if the network device pre-stores target data sets corresponding to different reference signal measurement value intervals, the network device can perform model training according to the target data set or a subset of the target data set, and the obtained target noise reduction model is used as Updated denoising model.
  • the current reference signal measurement value does not belong to any reference signal measurement value interval in the preset reference signal measurement value interval set
  • the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the current reference signal measurement value and the preset reference signal measurement value interval set, which may include:
  • the network device searches for the target reference signal measurement value interval corresponding to the current reference signal measurement value closest
  • the target noise reduction model of is used as the noise reduction model corresponding to the measured value of the current reference signal.
  • the network device searches for the target reference signal measurement value interval corresponding to the current reference signal measurement value closest to The target data set, model training is performed according to the target data set or a subset of the target data set, and the noise reduction model corresponding to the measured value of the current reference signal is obtained.
  • the network device performs model training according to the target data set or a subset of the target data set to obtain a noise reduction model corresponding to the current reference signal measurement value, which may include: the network device according to the target data set or a subset of the target data set, The target noise reduction model corresponding to the target reference signal measurement value range closest to the current reference signal measurement value is adjusted to obtain the noise reduction model corresponding to the current reference signal measurement value.
  • the network device after the network device completes the configuration of the noise reduction model group, when the network device measures that the current reference signal SNR of the uplink signal of a certain terminal device is in the signal-to-noise ratio interval ⁇ SNR k ⁇ , it can further target the The noise reduction model of the uplink performs transfer learning based on data weights. Specifically, since a large number of offline training datasets containing labels are stored in the base station, the base station does not need to adopt the unlabeled migration method of the target domain dataset.
  • the base station can select a part of the sub-dataset that is closest to the measured uplink SNR, for example: select a sub-dataset composed of data that is no more than 3dB away from the measured uplink SNR, and use the converged noise reduction model ⁇ k Migration and fine-tuning are performed on this sub-dataset to make the noise reduction model more suitable for the SNR represented by the sub-dataset, and to obtain an improvement in noise reduction performance.
  • the network device performs noise reduction processing on the uplink according to the noise reduction model.
  • the network device obtains the current reference signal measurement value; the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the current reference signal measurement value and the preset reference signal measurement value interval set, and/or, The target dataset, denoising model or target dataset is used for denoising.
  • transfer learning is applied to the noise reduction model of the wireless communication system, and a transfer learning design method of the noise reduction model group is proposed for the uplink network, so as to improve the applicability of the uplink noise reduction model.
  • FIG. 10 it is a schematic diagram of another embodiment of the noise reduction method based on transfer learning in the embodiment of the present application, which may include:
  • the terminal device acquires the noise reduction model based on transfer learning, which may include but not limited to the following steps 1001-1003, as follows:
  • the terminal device reports a noise reduction model update instruction and the measured value of the current reference signal.
  • the current reference signal measurement value measured by the terminal device is a downlink current reference signal measurement value measured by the terminal device.
  • the current reference signal measurement value includes at least one of reference signal received power, reference signal received quality, received signal strength indication, and signal-to-noise ratio.
  • the preset condition may be that the current signal measurement value is greater than the first preset threshold, or the absolute value of the difference between the current signal measurement value and the signal measurement value adapted to the known noise reduction model is greater than the second preset threshold , or, the absolute value of the ratio of the current signal measurement value to the signal measurement value adapted to the known noise reduction model satisfies a threshold range and the like.
  • this embodiment of the present application may be applied to a deep neural network, a recurrent neural network, or a convolutional neural network, or other neural networks.
  • the terminal device reporting the noise reduction model update instruction and the current reference signal measurement value may include: the terminal device reports the noise reduction model update instruction and the current reference signal measurement value through an uplink control instruction.
  • the network device acquires a current reference signal measurement value and a noise reduction model update instruction.
  • acquiring the current reference signal measurement value and the noise reduction model update instruction by the network device may include: the network device receiving the current reference signal measurement value and the noise reduction model update instruction reported by the terminal device.
  • the network device receiving the current reference signal measurement value and the noise reduction model update instruction reported by the terminal device may include: the network device receives the current reference signal measurement value and the noise reduction model update instruction reported by the terminal device through an uplink control instruction.
  • the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the current reference signal measurement value and the preset reference signal measurement value interval set, and/or, the target data set, which is used for model training, and obtains Noise reduction model.
  • the network device can perform model training according to the target data set or a subset of the target data set to obtain a noise reduction model, and then deliver the noise reduction model to the terminal device. It may also be that the network device sends the target data set or a subset of the target data set to the terminal device, and the terminal device performs model training according to the target data set or the subset of the target data set to obtain a noise reduction model.
  • the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the current reference signal measurement value and the preset reference signal measurement value interval set, and/or, the target data set can refer to the implementation shown in FIG. 9 The related description of step 902 in the example will not be repeated here.
  • the terminal device acquires a noise reduction model based on transfer learning.
  • the noise reduction model is a noise reduction model related to the downlink.
  • the terminal device obtains a noise reduction model based on transfer learning, which may include:
  • the network device sends the noise reduction model to the terminal device according to the update instruction of the noise reduction model, and the noise reduction model is used for the terminal device to perform noise reduction processing on the downlink; A denoising model for ; or,
  • the network device sends the target data set or a subset of the target data set to the terminal device according to the update instruction of the noise reduction model, and the target data set or a subset of the target data set is used for model training by the terminal device to obtain a noise reduction model ;
  • the terminal device receives the target data set or a subset of the target data set sent by the network device according to the update instruction of the noise reduction model, performs model training according to the target data set or the subset of the target data set, and obtains the noise reduction model.
  • the network device sends the noise reduction model to the terminal device according to the noise reduction model update instruction, which may include: the network device sends the noise reduction model to the terminal device through a downlink control instruction according to the noise reduction model update instruction; or,
  • the network device sends the target data set or a subset of the target data set to the terminal device according to the update instruction of the noise reduction model, which may include: the network device sends the target data set or the subset of the target data set through the downlink control instruction according to the update instruction of the noise reduction model.
  • the subset is sent to the terminal device.
  • the terminal device performs noise reduction processing according to the noise reduction model.
  • the terminal device performs noise reduction processing on the downlink according to the noise reduction model.
  • the embodiment of the present application provides a method for designing and downloading a migration noise reduction model on the terminal device side at the network device side during the downlink transmission process.
  • FIG. 11 it is a schematic diagram of performing transfer learning and updating of the noise reduction model of the terminal device on the network device side in the embodiment of the present application.
  • the terminal device measures at least one of RSRP, RSRQ, RSSI, and SINR on the downlink data resources issued by the network device.
  • a preset condition that needs to trigger the update of the noise reduction model for example:
  • the network device selects a target data set or a subset of the target data set matching the RSRP for migration training.
  • the terminal device downloads the updated noise reduction model.
  • the model update indication report may be carried by an uplink control indicator (Uplink Control Indicator, UCI), or carried by other uplink indication signals.
  • UCI Uplink Control Indicator
  • the terminal device when the measured value of the current reference signal measured by the terminal device satisfies the preset condition, the terminal device reports the update indication of the noise reduction model and the measured value of the current reference signal; the network device obtains the measured value of the current reference signal and Noise reduction model update indication; the network device obtains the noise reduction model corresponding to the current reference signal measurement value according to the current reference signal measurement value and the preset reference signal measurement value interval set, and/or, the target data set, the target data set is used for Model training to obtain a noise reduction model; the terminal device obtains a noise reduction model based on transfer learning.
  • the terminal device receives the noise reduction model sent by the network device according to the noise reduction model update instruction; or, the terminal device receives the target data set or a subset of the target data set sent by the network device according to the noise reduction model update instruction, and according to the target data set Or a subset of the target data set for model training to obtain a noise reduction model.
  • the embodiment of the present application applies transfer learning to the noise reduction model of the wireless communication system, proposes a transfer learning design method of the noise reduction model group for the downlink network, improves the applicability of the downlink noise reduction model, and reduces the computational complexity of the terminal equipment.
  • transfer learning is applied to the noise reduction model of the wireless communication system.
  • a noise reduction model for migration training the noise reduction model during downlink and uplink transmission can adapt to the changing signal-noise in the link environment than to achieve a better noise reduction effect.
  • a label-free migration learning method is proposed for the downlink noise reduction model to improve the performance of the noise reduction model in the scene of changing signal-to-noise ratio;
  • a migration learning method for the noise reduction model group is proposed for the uplink network to improve the applicability of the uplink noise reduction model ;
  • a migration training method for network equipment to the downlink noise reduction model is proposed to reduce the computational complexity of terminal equipment.
  • the proposal of this application does not limit the specific implementation method of the noise reduction model, but mainly protects the design method of using transfer learning to train the downlink and uplink noise reduction models.
  • FIG. 12 it is a schematic diagram of an embodiment of the terminal device in the embodiment of the present application, which may include:
  • An acquisition module 1201 configured to acquire a noise reduction model based on migration learning
  • a processing module 1202 configured to perform noise reduction processing according to the noise reduction model.
  • the processing module 1202 is specifically configured to perform model training according to the data set to obtain a noise reduction model based on migration learning; or,
  • the obtaining module 1201 is specifically configured to receive the noise reduction model based on migration learning delivered by the network device.
  • the acquiring module 1201 is specifically configured to acquire the source domain dataset, the label corresponding to the source domain dataset, and the target domain dataset;
  • the processing module 1202 is specifically configured to acquire the source domain dataset, the tag corresponding to the target domain dataset;
  • the label corresponding to the source domain data set and the target domain data set are subjected to model training to obtain a noise reduction model.
  • the processing module 1202 is specifically configured to determine a joint loss function according to the source domain data set, the label and the target domain data set; perform model training according to the joint loss function to obtain a noise reduction model.
  • the obtaining module 1201 is specifically configured to obtain the source domain data set, the label corresponding to the source domain data set, and the target domain dataset.
  • the processing module 1202 is specifically configured to determine an error loss function according to the source domain dataset and the label; determine an adaptation loss function according to the source domain dataset and the target domain dataset; The adaptation loss function and the error loss function determine a joint loss function.
  • the processing module 1202 is specifically configured to determine a joint loss function according to the first formula
  • L joint L 1 + ⁇ L 2 ;
  • L joint is the joint loss function, L 1 is the error loss function, L 2 is the adaptation loss function, and
  • ⁇ is the network device or all Describe the weight parameters configured by the terminal device.
  • the processing module 1202 is further configured to end the training when the number of times of the model training reaches a preset number, and/or after the joint loss function corresponding to the noise reduction model obtained by performing the model training reaches a preset value .
  • the obtaining module 1201 is configured to report the noise reduction model update instruction and the current reference signal measurement value to the network device.
  • the obtaining module 1201 is specifically configured to receive the noise reduction model sent by the network device according to the noise reduction model update instruction and the current reference signal measurement value; or,
  • the obtaining module 1201 is configured to receive the target data set or a subset of the target data set sent by the network device according to the noise reduction model update indication and the current reference signal measurement value; the processing module 1202 is configured to Perform model training on the target data set or a subset of the target data set to obtain the noise reduction model.
  • the obtaining module 1201 is specifically configured to report a noise reduction model update instruction and a current reference signal measurement value to the network device when the current reference signal measurement value measured by the terminal device meets a preset condition.
  • the obtaining module 1201 is specifically configured to report the noise reduction model update instruction and the current reference signal measurement value to the network device through an uplink control instruction.
  • the current reference signal measurement value includes at least one of reference signal received power, reference signal received quality, received signal strength indication, and signal-to-noise ratio.
  • the method is applied to a deep neural network, a recurrent neural network, or a convolutional neural network.
  • FIG. 13 it is a schematic diagram of an embodiment of a network device in the embodiment of the present application, which may include:
  • An acquisition module 1301, configured to acquire the measured value of the current reference signal
  • a processing module 1302 configured to acquire a noise reduction model corresponding to the current reference signal measurement value, or, a target data set, the noise reduction model or The target data set is used for noise reduction processing.
  • the processing module 1302 is configured to obtain the target reference signal measurement value interval according to the target reference signal measurement value interval when the current reference signal measurement value belongs to the target reference signal measurement value interval set in the preset reference signal measurement value interval set. Describe the noise reduction model corresponding to the measured value of the current reference signal.
  • the processing module 1302 is configured to search for a target noise reduction model corresponding to the target reference signal measurement value interval as the noise reduction model corresponding to the current reference signal measurement value; or,
  • the processing module 1302 is configured to search for the target data set corresponding to the target reference signal measurement value interval, perform model training according to the target data set or a subset of the target data set, and obtain the current The noise reduction model corresponding to the measured value of the reference signal.
  • the processing module 1302 is configured to, if the current reference signal measurement value does not belong to any reference signal measurement value interval in the preset reference signal measurement value interval set, find the maximum value of the current reference signal measurement value.
  • the target noise reduction model corresponding to the close target reference signal measurement value interval is used as the noise reduction model corresponding to the current reference signal measurement value; and/or, searching for the target reference signal measurement value interval corresponding to the closest target reference signal measurement value performing model training according to the target data set or a subset of the target data set to obtain a noise reduction model corresponding to the measured value of the current reference signal.
  • the processing module 1302 is configured to perform, according to the target data set or a subset of the target data set, the target noise reduction model corresponding to the target reference signal measurement value interval closest to the current reference signal measurement value Adjust to obtain the noise reduction model corresponding to the measured value of the current reference signal.
  • the noise reduction model is a noise reduction model related to the uplink.
  • the processing module 1302 is further configured to perform noise reduction processing on the uplink according to the noise reduction model.
  • the noise reduction model is a noise reduction model related to the downlink.
  • the obtaining module 1301 is configured to receive the current reference signal measurement value reported by the terminal device.
  • the obtaining module 1301 is specifically configured to receive the current reference signal measurement value reported by the terminal device through an uplink control instruction.
  • the acquiring module 1301 is further configured to receive an update indication of the noise reduction model reported by the terminal device.
  • the obtaining module 1301 is specifically configured to receive the noise reduction model update instruction reported by the terminal device through the uplink control instruction.
  • the obtaining module 1301 is specifically configured to deliver the noise reduction model to the terminal device according to the update instruction of the noise reduction model, and the noise reduction model is used by the terminal device to perform the downlink Noise reduction processing on ; or,
  • the acquisition module 1301 is specifically configured to deliver the target data set or a subset of the target data set to the terminal device according to the update instruction of the noise reduction model, and the target data set or the target data set The subset is used for the terminal device to perform model training to obtain the noise reduction model.
  • the obtaining module 1301 is specifically configured to deliver the noise reduction model to the terminal device through a downlink control instruction according to the noise reduction model update instruction; or,
  • the obtaining module 1301 is specifically configured to send the target data set or a subset of the target data set to the terminal device through the downlink control instruction according to the noise reduction model update instruction.
  • the current reference signal measurement value includes at least one of reference signal received power, reference signal received quality, received signal strength indication, and signal-to-noise ratio.
  • the network device is applied to a deep neural network, a recurrent neural network, or a convolutional neural network.
  • this embodiment of the present application further provides one or more types of terminal devices.
  • the terminal device in this embodiment of the present application may implement any implementation manner in the foregoing methods.
  • FIG. 14 it is a schematic diagram of another embodiment of a terminal device in an embodiment of the present invention.
  • the terminal device is described by taking a mobile phone as an example, and may include: a radio frequency (radio frequency, RF) circuit 1410, a memory 1420, an input unit 1430, Display unit 1440, sensor 1450, audio circuit 1460, wireless fidelity (wireless fidelity, WiFi) module 1470, processor 1480, and power supply 1490 and other components.
  • RF radio frequency
  • the radio frequency circuit 1410 includes a receiver 1414 and a transmitter 1412 .
  • the structure of the mobile phone shown in FIG. 14 does not constitute a limitation to the mobile phone, and may include more or less components than shown in the figure, or combine some components, or arrange different components.
  • the RF circuit 1410 can be used for sending and receiving information or receiving and sending signals during a call. In particular, after receiving the downlink information from the base station, it is processed by the processor 1480; in addition, the designed uplink data is sent to the base station.
  • the RF circuit 1410 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (low noise amplifier, LNA), a duplexer, and the like.
  • RF circuitry 1410 may also communicate with networks and other devices via wireless communications.
  • the above wireless communication can use any communication standard or protocol, including but not limited to global system of mobile communication (global system of mobile communication, GSM), general packet radio service (general packet radio service, GPRS), code division multiple access (code division multiple access) multiple access (CDMA), wideband code division multiple access (WCDMA), long term evolution (LTE), e-mail, short message service (short messaging service, SMS), etc.
  • GSM global system of mobile communication
  • GPRS general packet radio service
  • code division multiple access code division multiple access
  • WCDMA wideband code division multiple access
  • LTE long term evolution
  • e-mail short message service
  • SMS short message service
  • the memory 1420 can be used to store software programs and modules, and the processor 1480 executes various functional applications and data processing of the mobile phone by running the software programs and modules stored in the memory 1420 .
  • Memory 1420 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, at least one application program required by a function (such as a sound playback function, an image playback function, etc.); Data created by the use of mobile phones (such as audio data, phonebook, etc.), etc.
  • the memory 1420 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage devices.
  • the input unit 1430 can be used to receive input numbers or character information, and generate key signal input related to user settings and function control of the mobile phone.
  • the input unit 1430 may include a touch panel 1431 and other input devices 1432 .
  • the touch panel 1431 also referred to as a touch screen, can collect touch operations of the user on or near it (for example, the user uses any suitable object or accessory such as a finger or a stylus on the touch panel 1431 or near the touch panel 1431). operation), and drive the corresponding connection device according to the preset program.
  • the touch panel 1431 may include two parts, a touch detection device and a touch controller.
  • the touch detection device detects the user's touch orientation, and detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into contact coordinates, and sends it to the to the processor 1480, and can receive and execute commands sent by the processor 1480.
  • the touch panel 1431 can be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave.
  • the input unit 1430 may also include other input devices 1432 .
  • other input devices 1432 may include but not limited to one or more of a physical keyboard, function keys (such as volume control keys, switch keys, etc.), trackball, mouse, joystick, and the like.
  • the display unit 1440 may be used to display information input by or provided to the user and various menus of the mobile phone.
  • the display unit 1440 may include a display panel 1441.
  • the display panel 1441 may be configured in the form of a liquid crystal display (liquid crystal display, LCD) or an organic light-emitting diode (OLED).
  • the touch panel 1431 can cover the display panel 1441, and when the touch panel 1431 detects a touch operation on or near it, it sends it to the processor 1480 to determine the type of the touch event, and then the processor 1480 determines the type of the touch event according to the The type provides a corresponding visual output on the display panel 1441 .
  • the touch panel 1431 and the display panel 1441 are used as two independent components to realize the input and input functions of the mobile phone, in some embodiments, the touch panel 1431 and the display panel 1441 can be integrated to form a mobile phone. Realize the input and output functions of the mobile phone.
  • the handset may also include at least one sensor 1450, such as a light sensor, motion sensor, and other sensors.
  • the light sensor may include an ambient light sensor and a proximity sensor, wherein the ambient light sensor may adjust the brightness of the display panel 1441 according to the brightness of the ambient light, and the proximity sensor may turn off the display panel 1441 and/or when the mobile phone is moved to the ear. or backlight.
  • the accelerometer sensor can detect the magnitude of acceleration in various directions (generally three axes), and can detect the magnitude and direction of gravity when it is stationary, and can be used to identify the application of mobile phone posture (such as horizontal and vertical screen switching, related Games, magnetometer attitude calibration), vibration recognition related functions (such as pedometer, tap), etc.; as for other sensors such as gyroscope, barometer, hygrometer, thermometer, infrared sensor, etc. repeat.
  • mobile phone posture such as horizontal and vertical screen switching, related Games, magnetometer attitude calibration
  • vibration recognition related functions such as pedometer, tap
  • other sensors such as gyroscope, barometer, hygrometer, thermometer, infrared sensor, etc. repeat.
  • the audio circuit 1460, the speaker 1461, and the microphone 1462 can provide an audio interface between the user and the mobile phone.
  • the audio circuit 1460 can transmit the electrical signal converted from the received audio data to the speaker 1461, and the speaker 1461 converts it into an audio signal for output; After being received, it is converted into audio data, and then the audio data is processed by the output processor 1480, and then sent to another mobile phone through the RF circuit 1410, or the audio data is output to the memory 1420 for further processing.
  • WiFi is a short-distance wireless transmission technology.
  • the mobile phone can help users send and receive emails, browse web pages, and access streaming media through the WiFi module 1470. It provides users with wireless broadband Internet access.
  • Fig. 14 shows a WiFi module 1470, it can be understood that it is not an essential component of the mobile phone, and can be completely omitted as required without changing the essence of the invention.
  • the processor 1480 is the control center of the mobile phone. It uses various interfaces and lines to connect various parts of the entire mobile phone. By running or executing software programs and/or modules stored in the memory 1420, and calling data stored in the memory 1420, execution Various functions and processing data of the mobile phone, so as to monitor the mobile phone as a whole.
  • the processor 1480 may include one or more processing units; preferably, the processor 1480 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application programs, etc. , the modem processor mainly handles wireless communications. It can be understood that the foregoing modem processor may not be integrated into the processor 1480 .
  • the mobile phone also includes a power supply 1490 (such as a battery) for supplying power to various components.
  • a power supply 1490 (such as a battery) for supplying power to various components.
  • the power supply can be logically connected to the processor 1480 through the power management system, so as to realize functions such as managing charging, discharging, and power consumption management through the power management system.
  • the mobile phone may also include a camera, a Bluetooth module, etc., which will not be repeated here.
  • the processor 1480 is configured to obtain a noise reduction model based on transfer learning; perform noise reduction processing according to the noise reduction model.
  • the processor 1480 is specifically configured to receive the noise reduction model based on migration learning delivered by the network device.
  • the processor 1480 is specifically configured to acquire a source domain dataset, a label corresponding to the source domain dataset, and a target domain dataset; according to the source domain dataset, the label corresponding to the source domain dataset , and the target domain data set, perform model training to obtain a noise reduction model.
  • the processor 1480 is specifically configured to determine a joint loss function according to the source domain data set, the label, and the target domain data set; perform model training according to the joint loss function to obtain a noise reduction model.
  • the processor 1480 is specifically configured to acquire the source domain data set, the label corresponding to the source domain data set, and the target domain dataset.
  • the processor 1480 is specifically configured to determine an error loss function according to the source domain dataset and the label; determine an adaptation loss function according to the source domain dataset and the target domain dataset; The adaptation loss function and the error loss function determine a joint loss function.
  • the processor 1480 is specifically configured to determine a joint loss function according to the first formula
  • L joint L 1 + ⁇ L 2 ;
  • L joint is the joint loss function, L 1 is the error loss function, L 2 is the adaptation loss function, and
  • ⁇ is the network device or all Describe the weight parameters configured by the terminal device.
  • the processor 1480 is further configured to end the training when the number of times of model training reaches a preset number, and/or after the joint loss function corresponding to the noise reduction model obtained by performing the model training reaches a preset value .
  • the RF circuit 1410 is configured to report a noise reduction model update instruction and a current reference signal measurement value to the network device.
  • the RF circuit 1410 is specifically configured to receive the noise reduction model sent by the network device according to the noise reduction model update instruction and the current reference signal measurement value; or,
  • the RF circuit 1410 is configured to receive the target data set or a subset of the target data set sent by the network device according to the noise reduction model update indication and the current reference signal measurement value; the processor 1480 is configured to Perform model training on the target data set or a subset of the target data set to obtain the noise reduction model.
  • the RF circuit 1410 is specifically configured to report a noise reduction model update instruction and a current reference signal measurement value to the network device when the measured value of the current reference signal measured by the terminal device satisfies a preset condition value.
  • the RF circuit 1410 is specifically configured to report the noise reduction model update instruction and the current reference signal measurement value to the network device through an uplink control instruction.
  • the current reference signal measurement value includes at least one of reference signal received power, reference signal received quality, received signal strength indication, and signal-to-noise ratio.
  • the method is applied to a deep neural network, a recurrent neural network, or a convolutional neural network.
  • FIG. 15 it is a schematic diagram of another embodiment of the network device in the embodiment of the present application, which may include:
  • a memory 1501 storing executable program codes
  • a processor 1502 and a transceiver 1503 coupled to a memory 1501;
  • Processor 1502 configured to obtain the current reference signal measurement value; according to the current reference signal measurement value and the preset reference signal measurement value interval set, obtain a noise reduction model corresponding to the current reference signal measurement value, or a target data set , the noise reduction model or the target data set is used for noise reduction processing.
  • the processor 1502 is configured to obtain the target reference signal measurement value interval according to the target reference signal measurement value interval when the current reference signal measurement value belongs to the target reference signal measurement value interval set in the preset reference signal measurement value interval set. Describe the noise reduction model corresponding to the measured value of the current reference signal.
  • the processor 1502 is configured to search for a target noise reduction model corresponding to the target reference signal measurement value interval as the noise reduction model corresponding to the current reference signal measurement value; or,
  • the processor 1502 is configured to search for the target data set corresponding to the target reference signal measurement value interval, perform model training according to the target data set or a subset of the target data set, and obtain the current The noise reduction model corresponding to the measured value of the reference signal.
  • the processor 1502 is configured to, if the current reference signal measurement value does not belong to any reference signal measurement value interval in the preset reference signal measurement value interval set, search for the current reference signal measurement value that is the highest
  • the target noise reduction model corresponding to the close target reference signal measurement value interval is used as the noise reduction model corresponding to the current reference signal measurement value; and/or, searching for the target reference signal measurement value interval corresponding to the closest target reference signal measurement value performing model training according to the target data set or a subset of the target data set to obtain a noise reduction model corresponding to the measured value of the current reference signal.
  • the processor 1502 is configured to perform, according to the target data set or a subset of the target data set, the target noise reduction model corresponding to the target reference signal measurement value interval closest to the current reference signal measurement value Adjust to obtain the noise reduction model corresponding to the measured value of the current reference signal.
  • the noise reduction model is a noise reduction model related to the uplink.
  • the processor 1502 is further configured to perform noise reduction processing on the uplink according to the noise reduction model.
  • the noise reduction model is a noise reduction model related to the downlink.
  • the transceiver 1503 is configured to receive the current reference signal measurement value reported by the terminal device.
  • the transceiver 1503 is specifically configured to receive the current reference signal measurement value reported by the terminal device through an uplink control instruction.
  • the transceiver 1503 is further configured to receive the update instruction of the noise reduction model reported by the terminal device.
  • the transceiver 1503 is specifically configured to receive the noise reduction model update instruction reported by the terminal device through the uplink control instruction.
  • the transceiver 1503 is specifically configured to deliver the noise reduction model to the terminal device according to the noise reduction model update instruction, and the noise reduction model is used by the terminal device to perform the downlink Noise reduction processing on ; or,
  • the transceiver 1503 is specifically configured to deliver the target data set or a subset of the target data set to the terminal device according to the update instruction of the noise reduction model, and the target data set or a subset of the target data set The subset is used for the terminal device to perform model training to obtain the noise reduction model.
  • the transceiver 1503 is specifically configured to deliver the noise reduction model to the terminal device through a downlink control instruction according to the noise reduction model update instruction; or,
  • the transceiver 1503 is specifically configured to deliver the target data set or a subset of the target data set to the terminal device through the downlink control instruction according to the noise reduction model update instruction.
  • the current reference signal measurement value includes at least one of reference signal received power, reference signal received quality, received signal strength indication, and signal-to-noise ratio.
  • the network device is applied to a deep neural network, a recurrent neural network, or a convolutional neural network.
  • all or part of them may be implemented by software, hardware, firmware or any combination thereof.
  • software When implemented using software, it may be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present invention will be generated in whole or in part.
  • the computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website, computer, server, or data center Transmission to another website site, computer, server, or data center by wired (eg, coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that can be stored by a computer, or a data storage device such as a server or a data center integrated with one or more available media.
  • the available medium may be a magnetic medium (for example, a floppy disk, a hard disk, or a magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, a Solid State Disk (SSD)).
  • SSD Solid State Disk

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

Des modes de réalisation de la présente demande concernent un procédé de réduction de bruit basé sur l'apprentissage par transfert, un dispositif terminal, un dispositif de réseau et un support de stockage, qui sont utilisés pour présenter un modèle de réduction de bruit basé sur l'entraînement par transfert, permettant à un modèle de réduction de bruit dans une transmission de liaison descendante ou de liaison montante de s'adapter à une valeur de mesure de signal de référence qui change dans un environnement de liaison correspondant, ce qui permet d'obtenir de bons résultats de réduction de bruit. Les modes de réalisation de la présente demande peuvent comprendre les étapes suivantes : un dispositif terminal acquiert un modèle de réduction de bruit basé sur l'apprentissage par transfert ; et le dispositif terminal effectue un traitement de réduction de bruit en fonction du modèle de réduction de bruit.
PCT/CN2021/105462 2021-07-09 2021-07-09 Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage Ceased WO2023279366A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2021/105462 WO2023279366A1 (fr) 2021-07-09 2021-07-09 Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage
CN202180095342.4A CN116941185A (zh) 2021-07-09 2021-07-09 基于迁移学习的降噪方法、终端设备、网络设备及存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/105462 WO2023279366A1 (fr) 2021-07-09 2021-07-09 Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage

Publications (1)

Publication Number Publication Date
WO2023279366A1 true WO2023279366A1 (fr) 2023-01-12

Family

ID=84800227

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/105462 Ceased WO2023279366A1 (fr) 2021-07-09 2021-07-09 Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage

Country Status (2)

Country Link
CN (1) CN116941185A (fr)
WO (1) WO2023279366A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116250844A (zh) * 2023-03-03 2023-06-13 山东大学 基于条件生成对抗网络的心电信号降噪优化方法及系统
WO2025189861A1 (fr) * 2024-03-15 2025-09-18 华为技术有限公司 Procédé de communication et appareil associé

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118316900B (zh) * 2024-06-11 2024-10-11 广东省电信规划设计院有限公司 一种基于aigc的即时通信系统的噪声处理方法及装置

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080139155A1 (en) * 2006-12-07 2008-06-12 Olivier Boireau Techniques to reduce radio frequency noise
CN102543096A (zh) * 2011-12-26 2012-07-04 上海聚力传媒技术有限公司 对媒体文件播放过程中场景噪声进行抑制的方法与装置
CN106205631A (zh) * 2015-05-28 2016-12-07 三星电子株式会社 用于消除音频信号的噪声的方法及其电子装置
CN106961509A (zh) * 2017-04-25 2017-07-18 广东欧珀移动通信有限公司 通话参数处理方法、装置及电子设备
US20180033420A1 (en) * 2015-01-26 2018-02-01 Shenzhen Grandsun Electronic Co., Ltd Method and apparatus for controlling earphone noise reduction
CN109150775A (zh) * 2018-08-14 2019-01-04 西安交通大学 一种自适应噪声环境动态变化的鲁棒性在线信道状态信息估计方法
US20190057712A1 (en) * 2018-01-24 2019-02-21 Hisense Mobile Communications Technology Co., Ltd. Noise reduction method and electronic device
CN111402877A (zh) * 2020-03-17 2020-07-10 北京百度网讯科技有限公司 基于车载多音区的降噪方法、装置、设备和介质
CN111627455A (zh) * 2020-06-03 2020-09-04 腾讯科技（深圳）有限公司 一种音频数据降噪方法、装置以及计算机可读存储介质
CN111883164A (zh) * 2020-06-22 2020-11-03 北京达佳互联信息技术有限公司 模型训练方法、装置、电子设备及存储介质
CN112801218A (zh) * 2021-03-22 2021-05-14 中国人民解放军国防科技大学 一种基于降噪特征增强的多视角一维距离像融合识别方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111192599B (zh) * 2018-11-14 2022-11-22 中移（杭州）信息技术有限公司 一种降噪方法及装置
CN111883091B (zh) * 2020-07-09 2024-07-26 腾讯音乐娱乐科技（深圳）有限公司 音频降噪方法和音频降噪模型的训练方法

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080139155A1 (en) * 2006-12-07 2008-06-12 Olivier Boireau Techniques to reduce radio frequency noise
CN102543096A (zh) * 2011-12-26 2012-07-04 上海聚力传媒技术有限公司 对媒体文件播放过程中场景噪声进行抑制的方法与装置
US20180033420A1 (en) * 2015-01-26 2018-02-01 Shenzhen Grandsun Electronic Co., Ltd Method and apparatus for controlling earphone noise reduction
CN106205631A (zh) * 2015-05-28 2016-12-07 三星电子株式会社 用于消除音频信号的噪声的方法及其电子装置
CN106961509A (zh) * 2017-04-25 2017-07-18 广东欧珀移动通信有限公司 通话参数处理方法、装置及电子设备
US20190057712A1 (en) * 2018-01-24 2019-02-21 Hisense Mobile Communications Technology Co., Ltd. Noise reduction method and electronic device
CN109150775A (zh) * 2018-08-14 2019-01-04 西安交通大学 一种自适应噪声环境动态变化的鲁棒性在线信道状态信息估计方法
CN111402877A (zh) * 2020-03-17 2020-07-10 北京百度网讯科技有限公司 基于车载多音区的降噪方法、装置、设备和介质
CN111627455A (zh) * 2020-06-03 2020-09-04 腾讯科技（深圳）有限公司 一种音频数据降噪方法、装置以及计算机可读存储介质
CN111883164A (zh) * 2020-06-22 2020-11-03 北京达佳互联信息技术有限公司 模型训练方法、装置、电子设备及存储介质
CN112801218A (zh) * 2021-03-22 2021-05-14 中国人民解放军国防科技大学 一种基于降噪特征增强的多视角一维距离像融合识别方法

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116250844A (zh) * 2023-03-03 2023-06-13 山东大学 基于条件生成对抗网络的心电信号降噪优化方法及系统
CN116250844B (zh) * 2023-03-03 2024-04-26 山东大学 基于条件生成对抗网络的心电信号降噪优化方法及系统
WO2025189861A1 (fr) * 2024-03-15 2025-09-18 华为技术有限公司 Procédé de communication et appareil associé

Also Published As

Publication number Publication date
CN116941185A (zh) 2023-10-24

Similar Documents

Publication Publication Date Title
EP4422248A1 (fr) Procédé de demande de modèle, procédé de traitement de demande de modèle et dispositif associé
KR102753058B1 (ko) 코딩 방법, 디코딩 방법 및 디바이스
CN110401473A (zh) 动态调整发射功率的方法、移动终端及存储介质
WO2023279366A1 (fr) Procédé de réduction de bruit basé sur l'apprentissage par transfert, dispositif terminal, dispositif de réseau et support de stockage
US11342977B2 (en) Method and apparatus of fusing radio frequency and sensor measurements for beam management
US20240357436A1 (en) Communication method and communication apparatus
US20240357437A1 (en) Communication method and communication apparatus
US20240023062A1 (en) Apparatuses and methods for facilitating efficient paging in networks
US11800381B2 (en) Optimal device position for wireless communication
CN113225787B (zh) Wi-Fi扫描的方法、终端设备及存储介质
US11496908B2 (en) Apparatuses and methods for enhancing network coverage in accordance with predictions
WO2022116651A1 (fr) Procédé de sélection préférentielle d'une cellule endc, dispositif terminal et support de stockage
WO2022133727A1 (fr) Procédé de transmission répétée de pusch et dispositif de terminal
CN117835262A (zh) Ai模型的处理方法、装置及通信设备
CN110177169A (zh) 数据交互方法、第一终端、第二终端及计算机存储介质
CN116965116A (zh) 使用神经网络将蜂窝通信与传感器数据相结合的设备
US20220231709A1 (en) Apparatuses and methods for generating ad-hoc networks to extend coverage
CN110045903A (zh) 界面操作响应方法、移动终端、装置及计算机存储介质
CN109375979A (zh) 流量控制方法、终端及可读存储介质
WO2022252716A1 (fr) Procédé d'acquisition de paramètres de temporisateur, équipement terminal et support d'enregistrement
CN118055421A (zh) 波束预测方法、装置、终端、网络侧设备及存储介质
CN116982300A (zh) 信号处理的方法及接收机
WO2022233061A1 (fr) Procédé de traitement de signal, dispositif de communication et système de communication
WO2022217552A1 (fr) Procédé de traitement d'état de panneau, dispositif de communication et support de stockage
US20240107596A1 (en) Device-Driven Network Connection

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21948860

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 202180095342.4

Country of ref document: CN

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21948860

Country of ref document: EP

Kind code of ref document: A1