WO2023283785A1 - Method for processing signal, and receiver - Google Patents

Method for processing signal, and receiver Download PDF

Info

Publication number
WO2023283785A1
WO2023283785A1 PCT/CN2021/105842 CN2021105842W WO2023283785A1 WO 2023283785 A1 WO2023283785 A1 WO 2023283785A1 CN 2021105842 W CN2021105842 W CN 2021105842W WO 2023283785 A1 WO2023283785 A1 WO 2023283785A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
receiver
decoder
online training
received signal
Prior art date
Application number
PCT/CN2021/105842
Other languages
French (fr)
Chinese (zh)
Inventor
肖寒
Original Assignee
Oppo广东移动通信有限公司
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 Oppo广东移动通信有限公司 filed Critical Oppo广东移动通信有限公司
Priority to CN202180095411.1A priority Critical patent/CN116982300A/en
Priority to PCT/CN2021/105842 priority patent/WO2023283785A1/en
Publication of WO2023283785A1 publication Critical patent/WO2023283785A1/en

Links

Images

Classifications

    • 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 AI decoder In order to improve the decoding accuracy of the AI decoder, it is necessary to pre-train the AI decoder based on the preset training set before deploying the AI decoder to the receiver (that is, go online to the receiver), also known as "offline train".
  • the actual communication system is more complicated, and the training data in the training set cannot cover all the situations, and there will be a big difference with the received signal actually received by the receiver. If the AI decoder is trained offline only based on the training set, it will As a result, the generalization ability of the trained AI decoder is poor.
  • Fig. 6 is a schematic diagram of an autoencoder-based CSI feedback system.
  • the pilot symbols are inserted into the modulation symbols to form a signal to be transmitted, wherein the pilot symbols can be used for channel estimation and symbol detection by the receiver.
  • the above-mentioned signal is carried on a channel and transmitted to a receiver. Wherein, during the transmission process of the signal through the channel, noise is usually superimposed.
  • Pooling layer 430 because it is often necessary to reduce the number of training parameters, it is often necessary to periodically introduce a pooling layer after the convolutional layer, for example, it can be a layer of convolutional layer followed by a layer of pooling layer as shown in Figure 4 , can also be a multi-layer convolutional layer followed by one or more pooling layers. In signal processing, the sole purpose of pooling layers is to reduce the spatial size of the extracted information.
  • the CNN model In order to minimize the loss function, the CNN model needs to be trained.
  • the CNN model may be trained using a backpropagation algorithm (BP).
  • the training process of BP consists of forward propagation process and back propagation process.
  • the input data In the process of forward propagation (the propagation from 410 to 450 in Fig. 4 is forward propagation), the input data is input into the above layers of the CNN model, processed layer by layer and transmitted to the output layer. If the result output at the output layer is quite different from the ideal result, the above loss function is minimized as the optimization goal, and transferred to backpropagation (as shown in Fig.
  • the partial derivative of the optimization target to the weight of each neuron constitutes the gradient of the optimization target to the weight vector, which is used as the basis for modifying the model weight.
  • the training process of CNN is completed in the weight modification process. When the above error reaches the expected value, the training process of CNN ends.
  • the CNN shown in Figure 4 is only an example of a convolutional neural network.
  • the convolutional neural network can also exist in the form of other network models, which are not discussed in this embodiment of the present application. limited.
  • Network equipment and terminal equipment can be deployed on land, including indoors or outdoors, hand-held or vehicle-mounted; they can also be deployed on water; they can also be deployed on aircraft, balloons and satellites in the air. In this embodiment of the application, there is no limitation on the scenarios where network devices and terminal devices are located.
  • the AI decoder after online training will be able to use the received signal When decoding, it can also have a high accuracy rate, which is conducive to improving the generalization ability of the AI decoder.
  • the indication information of the above online training may indicate the time of the online training in various ways.
  • the indication information of online training may indicate the online training period to instruct the receiver to periodically perform online training for the AI decoder.
  • the above online training indication information may also directly indicate the start time and/or end time of the online training.
  • the indication information of the online training may indicate the start time of the online training in a display or implicit manner.
  • the instruction information of the online training may directly carry the start time.
  • the receiver can start from the time-domain unit where the indication information for online training is transmitted, and offset the preset time-domain unit to obtain the shifted time-domain unit, which is the starting point start time, where the time domain unit may be, for example, a time slot, a subframe, and the like.
  • the online training start time and the online training end time may be carried in the indication information of the online training, and correspondingly, the receiver performs the online training within the time period indicated by the start time and the end time.
  • only the online training start time may be carried in the online training indication information, correspondingly, after receiving the online training indication information, the receiver performs an online training at the start time.
  • the receiver directly uses the aforementioned online trained AI decoder to decode the received signal received in the current transmission period.
  • the foregoing first random noise may be pre-stored by the receiver, or may be pre-generated by the receiver, which is not limited in this embodiment of the present application.
  • the indication information of the online training further includes second indication information, where the second indication information is used to indicate the size of the training data used in the first type of transmission period.
  • the received signal is a received signal of a pilot signal transmitted by the transmitter
  • the processing unit 1720 may also be configured to input the received signal into the AI decoder for decoding, perform channel estimation, and obtain the The decoded signal, the decoded signal includes estimated first channel information; and according to the first channel information and first random noise, the pilot signal stored in the receiver is processed to obtain the restored signal , wherein the pilot signal transmitted by the transmitter is the same as the pilot signal stored by the receiver.
  • 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 application 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.

Landscapes

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

Abstract

Provided are a method for processing a signal and a receiver. The method comprises: a receiver receives a wireless signal transmitted by a transmitter, and obtains a received signal; the receiver inputs the received signal into an Artificial Intelligence (AI) decoder for decoding, and obtains a decoded signal; the receiver, according to the decoded signal, generates a recovery signal of the received signal; the receiver, according to the difference between the recovery signal and the received signal, trains the AI decoder online. Since both the received signal and the recovery signal can be acquired by the receiver, the method can implement an online training process of the AI decoder so as to improve the generalization ability of the AI decoder.

Description

信号处理的方法及接收机Signal processing method and receiver 技术领域technical field
本申请涉及通信领域，具体涉及一种信号处理的方法及接收机。The present application relates to the communication field, and in particular to a signal processing method and receiver.
背景技术Background technique
自编码器(auto encoder，AE)是一种将输入信号作为训练目标的人工智能(artificial intelligence，AI)模型，其包含的AI解码器的架构与通信***中的很多架构天然适配。例如，AI解码器可对应通信***的接收机。Autoencoder (auto encoder, AE) is an artificial intelligence (AI) model that takes the input signal as the training target, and the architecture of the AI decoder included in it is naturally adapted to many architectures in the communication system. For example, an AI decoder may correspond to a receiver of a communication system.
目前，为了提高AI解码器解码的准确性，需要在将AI解码器部署至接收机(即上线至接收机)之前，基于预设的训练集对AI解码器进行预先训练，又称“线下训练”。但是，实际的通信***情况比较复杂，训练集中的训练数据无法涵盖全部情况，会与接收机实际接收的接收信号存在较大的差异，如果仅基于训练集对AI解码器进行线下训练，会导致训练后的AI解码器的泛化能力较差。At present, in order to improve the decoding accuracy of the AI decoder, it is necessary to pre-train the AI decoder based on the preset training set before deploying the AI decoder to the receiver (that is, go online to the receiver), also known as "offline train". However, the actual communication system is more complicated, and the training data in the training set cannot cover all the situations, and there will be a big difference with the received signal actually received by the receiver. If the AI decoder is trained offline only based on the training set, it will As a result, the generalization ability of the trained AI decoder is poor.
发明内容Contents of the invention
本申请提供一种信号处理的方法及接收机，以对AI解码器进行在线训练，有利于提高AI解码器的泛化能力。The present application provides a signal processing method and a receiver for online training of an AI decoder, which is beneficial to improving the generalization ability of the AI decoder.
第一方面，提供一种信号处理的方法，包括：接收机对发射机发射的无线信号进行接收，得到接收信号；所述接收机将接收信号输入人工智能AI解码器进行解码，得到解码信号；所述接收机根据所述解码信号，生成所述接收信号的恢复信号；所述接收机根据所述恢复信号与所述接收信号之间的差异，对所述AI解码器进行在线训练。In the first aspect, a signal processing method is provided, including: a receiver receives a wireless signal transmitted by a transmitter to obtain a received signal; the receiver inputs the received signal to an artificial intelligence AI decoder for decoding to obtain a decoded signal; The receiver generates a restoration signal of the received signal according to the decoded signal; the receiver performs online training for the AI decoder according to a difference between the restoration signal and the received signal.
第二方面，提供一种接收机，包括：接收单元，用于对发射机发射的无线信号进行接收，得到接收信号；处理单元，用于将所述接收信号输入人工智能AI解码器进行解码，得到解码信号；所述处理单元，用于根据所述解码信号，生成所述接收信号的恢复信号；所述处理单元，用于根据所述恢复信号与所述接收信号之间的差异，对所述AI解码器进行在线训练。In a second aspect, a receiver is provided, including: a receiving unit configured to receive a wireless signal transmitted by a transmitter to obtain a received signal; a processing unit configured to input the received signal into an artificial intelligence AI decoder for decoding, Obtaining a decoded signal; the processing unit is configured to generate a recovery signal of the received signal according to the decoded signal; the processing unit is configured to process the recovered signal based on a difference between the recovered signal and the received signal The above AI decoder is trained online.
第三方面，提供一种接收机，包括存储器和处理器，所述存储器用于存储程序，所述处理器用于调用所述存储器中的程序，以执行如第一方面所述的方法。In a third aspect, a receiver is provided, including a memory and a processor, the memory is used to store a program, and the processor is used to invoke the program in the memory to execute the method as described in the first aspect.
第四方面，提供一种装置，包括处理器，用于从存储器中调用程序，以执行第一方面所述的方法。In a fourth aspect, an apparatus is provided, including a processor, configured to call a program from a memory to execute the method described in the first aspect.
第五方面，提供一种芯片，包括处理器，用于从存储器调用程序，使得安装有所述芯片的设备执行第一方面所述的方法。According to a fifth aspect, a chip is provided, including a processor, configured to call a program from a memory, so that a device installed with the chip executes the method described in the first aspect.
第六方面，提供一种计算机可读存储介质，其上存储有程序，所述程序使得计算机执行第一方面所述的方法。In a sixth aspect, a computer-readable storage medium is provided, on which a program is stored, and the program causes a computer to execute the method described in the first aspect.
第七方面，提供一种计算机程序产品，包括程序，所述程序使得计算机执行第一方面所述的方法。In a seventh aspect, a computer program product is provided, including a program, the program causes a computer to execute the method described in the first aspect.
第八方面，提供一种计算机程序，所述计算机程序使得计算机执行第一方面所述的方法。In an eighth aspect, a computer program is provided, the computer program causes a computer to execute the method described in the first aspect.
接收机使用AI解码器对接收信号解码，得到解码信号，再根据解码信号生成接收信号的恢复信号，并基于接收信号与恢复信号之间的差异，对AI解码器进行在线训练，由于接收信号和恢复信号都是可以被接收机获取的，因此，本申请实施例的方法可以实现AI解码器的在线训练过程，有利于提高AI解码器的泛化能力。The receiver uses the AI decoder to decode the received signal to obtain the decoded signal, and then generates the restored signal of the received signal according to the decoded signal, and based on the difference between the received signal and the restored signal, the AI decoder is trained online. All recovered signals can be acquired by the receiver. Therefore, the method in the embodiment of the present application can realize the online training process of the AI decoder, which is beneficial to improve the generalization ability of the AI decoder.
附图说明Description of drawings
图1是本申请实施例适用的无线通信***中传输信号的流程图。FIG. 1 is a flow chart of transmitting signals in a wireless communication system to which an embodiment of the present application is applicable.
图2是本申请实施例适用的信道估计及信号恢复的示意图。FIG. 2 is a schematic diagram of channel estimation and signal recovery applicable to the embodiments of the present application.
图3介绍本申请实施例适用的神经网络的结构图。FIG. 3 introduces a structural diagram of a neural network applicable to the embodiment of the present application.
图4介绍本申请实施例适用的CNN的结构图。FIG. 4 introduces a structure diagram of a CNN applicable to the embodiment of the present application.
图5是基于AI解码器进行信道估计的过程的示意图。Fig. 5 is a schematic diagram of a process of channel estimation based on an AI decoder.
图6是基于自编码器的CSI反馈***的示意图。Fig. 6 is a schematic diagram of an autoencoder-based CSI feedback system.
图7是基于AI解码器的接收机的示意图。Fig. 7 is a schematic diagram of an AI decoder based receiver.
图8是本申请实施例适用的无线通信***800。Fig. 8 is a wireless communication system 800 applicable to the embodiment of the present application.
图9是本申请实施例的信号处理的方法的流程图。FIG. 9 is a flowchart of a signal processing method according to an embodiment of the present application.
图10是本申请实施例的基于周期性在线训练的信号处理的方法示意图。FIG. 10 is a schematic diagram of a signal processing method based on periodic online training according to an embodiment of the present application.
图11是本申请实施例的基于非持续性在线训练的信号处理的方法示意图。FIG. 11 is a schematic diagram of a method of signal processing based on non-persistent online training according to an embodiment of the present application.
图12是本申请实施例的基于非周期在线训练的信号处理的方法示意图。FIG. 12 is a schematic diagram of a signal processing method based on aperiodic online training according to an embodiment of the present application.
图13是本申请实施例的信道估计过程中信号处理的方法示意图。Fig. 13 is a schematic diagram of a signal processing method in a channel estimation process according to an embodiment of the present application.
图14是自编码器在通信***中部署的示意图。Fig. 14 is a schematic diagram of an autoencoder deployed in a communication system.
图15是本申请实施例的CSI反馈过程中信号处理的方法示意图。FIG. 15 is a schematic diagram of a signal processing method in a CSI feedback process according to an embodiment of the present application.
图16是本申请实施例的数据传输过程中信号处理的方法示意图。FIG. 16 is a schematic diagram of a signal processing method during data transmission according to an embodiment of the present application.
图17是本申请实施例的接收机的示意性结构图。Fig. 17 is a schematic structural diagram of a receiver according to an embodiment of the present application.
图18是本申请实施例的用于信号处理的装置的示意性结构图。Fig. 18 is a schematic structural diagram of an apparatus for signal processing according to an embodiment of the present application.
具体实施方式detailed description
下面将结合附图，对本申请中的技术方案进行描述。为了便于理解本申请，下文先结合图1至图7介绍本申请实施例涉及的术语及通信过程。The technical solution in this application will be described below with reference to the accompanying drawings. In order to facilitate the understanding of the present application, terms and communication processes involved in the embodiments of the present application are introduced below with reference to FIG. 1 to FIG. 7 .
一、无线通信***中的信号传输过程1. The signal transmission process in the wireless communication system
图1是本申请实施例适用的无线通信***中传输信号的流程图。如图1所示，在无线通信***中的信号传输过程可以大致分为图1所示的多种信号处理过程S111至S118。图1所示的部分或全部信号处理过程都可以通过单独的AI模型实现，其具体的实现可以参见图5至图7的介绍。FIG. 1 is a flow chart of transmitting signals in a wireless communication system to which an embodiment of the present application is applicable. As shown in FIG. 1 , the signal transmission process in the wireless communication system can be roughly divided into various signal processing processes S111 to S118 shown in FIG. 1 . Part or all of the signal processing process shown in FIG. 1 can be realized through a separate AI model, and its specific implementation can refer to the introduction of FIG. 5 to FIG. 7 .
发射机在信道编码过程S111中对待传输的信息进行信道编码，得到编码后的码流。其中，待传输的信息可以是比特流的形式。In the channel coding process S111, the transmitter performs channel coding on the information to be transmitted to obtain coded streams. Wherein, the information to be transmitted may be in the form of a bit stream.
在调制过程S112中将码流调制为调制符号。In the modulation process S112, the code stream is modulated into modulation symbols.
在***导频过程S113中，将导频符号***上述调制符号，形成待发射的信号，其中，导频符号可以供接收机进行信道估计和符号检测。In the pilot insertion process S113, the pilot symbols are inserted into the modulation symbols to form a signal to be transmitted, wherein the pilot symbols can be used for channel estimation and symbol detection by the receiver.
在传输信号S114中，将上述信号承载在信道上发射至接收机。其中，信号在通过信道的传输过程中，通常会叠加噪声。In the transmission signal S114, the above-mentioned signal is carried on a channel and transmitted to a receiver. Wherein, during the transmission process of the signal through the channel, noise is usually superimposed.
在信道估计过程S115中，接收机可以基于导频信号进行信道估计，得到信道状态信息(channel state information-reference signal，CSI)，并通过反馈链路将CSI反馈给发射机，供发射机调整信道编码、调制、预编码等方式。In the channel estimation process S115, the receiver can perform channel estimation based on the pilot signal to obtain channel state information (channel state information-reference signal, CSI), and feed back the CSI to the transmitter through the feedback link for the transmitter to adjust the channel Coding, modulation, precoding, etc.
在符号检测过程S116中，对接收到的调制符号进行符号检测，得到检测结果。In the symbol detection process S116, symbol detection is performed on the received modulation symbols to obtain a detection result.
在解调过程S117中，基于检测结果对接收到的调制符号进行解调，得到码流。In the demodulation process S117, the received modulation symbol is demodulated based on the detection result to obtain a code stream.
在信道解码过程S118中，对码流进行解码，得到恢复后的信息，其中，恢复后的信息可以是比特流的形式。In the channel decoding process S118, the code stream is decoded to obtain restored information, wherein the restored information may be in the form of a bit stream.
应理解，图1所示的信号处理过程S111至S118仅仅示例性地列出了无线通信***中常见的信号处理过程，无线通信***中还可以包括资源映射、预编码、干扰消除、CSI测量等信号处理过程，这些信号处理过程也都可以通过单独的AI模型实现。为了简洁，本申请不再赘述。It should be understood that the signal processing procedures S111 to S118 shown in FIG. 1 are only an example of common signal processing procedures in a wireless communication system, and the wireless communication system may also include resource mapping, precoding, interference cancellation, CSI measurement, etc. The signal processing process, these signal processing processes can also be realized through a separate AI model. For the sake of brevity, this application will not repeat them.
二、信道估计2. Channel Estimation
由于无线信道环境的复杂性和时变性，在无线通信***(例如，上文介绍的无线通信***)中，接收机需要基于对信道的估计结果对接收的信号进行恢复。图2是本申请实施例适用的信道估计及信号恢复的示意图。Due to the complexity and time-varying nature of the wireless channel environment, in a wireless communication system (eg, the wireless communication system introduced above), the receiver needs to restore the received signal based on the channel estimation result. FIG. 2 is a schematic diagram of channel estimation and signal recovery applicable to the embodiments of the present application.
如图2所示，在步骤S210中，发射机在时频资源上除了发射数据信号外，还会发射一系列接收机已知的导频信号，如信道状态信息参考信号(channel state information-reference signal，CSI-RS)、解调参考信号(demodulation reference signal，DMRS)等。As shown in Figure 2, in step S210, in addition to transmitting data signals on time-frequency resources, the transmitter also transmits a series of pilot signals known to the receiver, such as channel state information-reference signal (CSIRS) signal, CSI-RS), demodulation reference signal (demodulation reference signal, DMRS), etc.
在步骤S211中，发射机通过信道向发射机发射上述数据信号和导频信号。In step S211, the transmitter transmits the above-mentioned data signal and pilot signal to the transmitter through a channel.
在步骤S212中，接收机接收到导频信号后可以进行信道估计。在一种可能的实现方式中，接收机可以基于预存的导频序列与接收到的导频序列，通过信道估计算法(例如，最小二乘(least squares method，LS)信道估计)，估计出传输导频信号的信道的信道信息。In step S212, the receiver may perform channel estimation after receiving the pilot signal. In a possible implementation, the receiver can estimate the transmission frequency based on the pre-stored pilot sequence and the received pilot sequence through a channel estimation algorithm (for example, least squares method (LS) channel estimation). Channel information of the channel of the pilot signal.
在步骤S213中，接收机可以根据传输导频序列的信道的信道信息，利用插值算法恢复出全时频资源上的信道信息，用于后续的CSI反馈或数据恢复等。In step S213, the receiver can restore the channel information on the full time-frequency resource by using an interpolation algorithm according to the channel information of the channel transmitting the pilot sequence, for subsequent CSI feedback or data recovery.
三、CSI反馈3. CSI Feedback
在无线通信***中，主要是利用基于码本的方案来实现信道特征的提取与反馈，即在接收机进行信道估计后，根据信道估计的结果按照某种优化准则，从预先设定的预编码码本中选择与当前信道最匹配的预编码矩阵，并通过空口的反馈链路，将预编码矩阵索引(precoding matrix index，PMI)信息反馈给发射机，供发射机实现预编码。在一些实现方式中，接收机还可以将测量得出的信道质量指示(channel quality indication，CQI)反馈给发射机，供发射机实现自适应调制编码等。In wireless communication systems, the codebook-based scheme is mainly used to realize the extraction and feedback of channel features, that is, after the receiver performs channel estimation, according to the channel estimation results according to a certain optimization criterion, the pre-set precoding The precoding matrix that best matches the current channel is selected in the codebook, and the precoding matrix index (precoding matrix index, PMI) information is fed back to the transmitter through the feedback link of the air interface for the transmitter to implement precoding. In some implementation manners, the receiver may also feed back the measured channel quality indication (CQI) to the transmitter for the transmitter to implement adaptive modulation and coding.
四、自编码器4. Autoencoder
自编码器是一种将输入信号作为训练目标的神经网络，其包含的AI编码器和/或AI解码器的架构与通信***中的很多架构天然适配。例如，AI编码器与AI解码器可分别对应无线通信***的发射机和接收机。又例如，AI编码器与AI解码器也可以分别对应CSI反馈过程中的信道压缩模块与解压缩模块。又例如，自编码器中的AI解码器也可以单独应用于信道估计过程中，用于接收机对信道信息的恢复。下文将结合图5至图7介绍，为了简洁，在此不在赘述。An autoencoder is a neural network that uses an input signal as a training target, and the architecture of the AI encoder and/or AI decoder it contains is naturally adapted to many architectures in communication systems. For example, an AI encoder and an AI decoder may respectively correspond to a transmitter and a receiver of a wireless communication system. For another example, the AI encoder and the AI decoder may also correspond to the channel compression module and the decompression module in the CSI feedback process respectively. For another example, the AI decoder in the self-encoder can also be applied in the process of channel estimation alone to recover channel information by the receiver. The following will be introduced in conjunction with FIG. 5 to FIG. 7 , and for the sake of brevity, details are not repeated here.
通常，在自编码器部署到通信***之前，可以基于训练集对自编码器进行训练。例如，当自编码器仅包含AI解码器时，可以基于训练集对AI解码器进行训练。当自编码器包含AI编码器时，可以基于训练集对AI编码器进行训练。当自编码器包含AI编码器和AI解码器时，可以对AI编码器和AI解码器进行联合训练。Typically, an autoencoder can be trained on a training set before it is deployed in a communication system. For example, when the autoencoder contains only the AI decoder, the AI decoder can be trained based on the training set. When the autoencoder includes an AI encoder, the AI encoder can be trained based on the training set. When the autoencoder contains an AI encoder and an AI decoder, the AI encoder and AI decoder can be jointly trained.
在一些实现方式中，由于AE是一种将输入信号作为训练目标的神经网络模型，因此，在通过损失函数表示自编码器的输入和输出之间的差异时，自编码器的训练目标可以理解为在损失函数最小的情况下，优化AI编码器和AI解码器的权重。In some implementations, since AE is a neural network model that takes the input signal as the training target, when the difference between the input and output of the autoencoder is represented by the loss function, the training target of the autoencoder can be understood In order to optimize the weights of AI encoder and AI decoder with the minimum loss function.
例如，一个包含AI编码器f(·)和AI解码器g(·)的自编码器表示为g(f(·))。原始信号s首先通过AI编码器f(·)编码后，AI编码器f(·)输出编码信号表示为q＝f(s)。在将编码信号输入AI解码器g(·)解码，AI解码器g(·)输出的解码信号表示为s`＝g(q)＝g(f(s))。在联合训练阶段，可将min_ {g,f}l(s,g(f(s)))作为训练目标，对AI编码器f(·)和AI解码器g(·)进行联合训练，其中，l(·)表示损失函数。 For example, an autoencoder consisting of an AI encoder f( ) and an AI decoder g( ) is denoted as g(f( )). After the original signal s is encoded by the AI encoder f(·), the output coded signal of the AI encoder f(·) is denoted as q=f(s). When the encoded signal is input to the AI decoder g(·) for decoding, the decoded signal output by the AI decoder g(·) is expressed as s`=g(q)=g(f(s)). In the joint training phase, min_ {g,f} l(s,g(f(s))) can be used as the training target to jointly train the AI encoder f( ) and the AI decoder g( ), where , l( ) represents the loss function.
五、神经网络5. Neural Network
近年来，以神经网络为代表的人工智能研究在很多领域都取得了非常大的成果，其也将在未来很长一段时间内在人们的生产生活中起到重要的作用。常见的神经网络有卷积神经网络(convolutional neural network，CNN)、循环神经网络(recurrent neural network，RNN)、深度神经网络(deep neural network，DNN)等。In recent years, artificial intelligence research represented by neural networks has achieved great results in many fields, and it will also play an important role in people's production and life for a long time in the future. Common neural networks include convolutional neural network (CNN), recurrent neural network (RNN), deep neural network (DNN), etc.
下文结合图3介绍本申请实施例适用的神经网络。图3所示的神经网络按照不同层的 位置划分可以分为三类：输入层310，隐藏层320和输出层330。一般来说，第一层是输入层310、最后一层是输出层330，第一层和最后一层之间的中间层都是隐藏层320。The neural network applicable to the embodiment of the present application is introduced below with reference to FIG. 3 . The neural network shown in FIG. 3 can be divided into three types according to the position of different layers: input layer 310, hidden layer 320 and output layer 330. Generally speaking, the first layer is the input layer 310 , the last layer is the output layer 330 , and the middle layer between the first layer and the last layer is the hidden layer 320 .
输入层310用于输入数据，其中，输入数据例如可以是接收机接收的接收信号。隐藏层320用于对输入数据进行处理，例如，对接收信号进行解压缩处理。输出层330用于输出处理后的输出数据，例如，输出解压后的信号。The input layer 310 is used for input data, wherein the input data may be, for example, received signals received by a receiver. The hidden layer 320 is used to process the input data, for example, to decompress the received signal. The output layer 330 is used to output processed output data, for example, output a decompressed signal.
如图3所示，神经网络包括多个层，每个层包括多个神经元，层与层之间的神经元可以是全连接的，也可以是部分连接的。对于连接的神经元而言，上一层的神经元的输出可以作为下一层的神经元的输入。As shown in Figure 3, the neural network includes multiple layers, each layer includes multiple neurons, and the neurons between layers can be fully connected or partially connected. For connected neurons, the output of neurons in the previous layer can be used as the input of neurons in the next layer.
随着神经网络研究的不断发展，近年来又提出了神经网络深度学习算法，在神经网络中引入较多的隐层，形成DNN，更多的隐含层让DNN更能够刻画现实世界中的复杂情形。理论上而言，参数越多的模型复杂度越高，“容量”也就越大，也就意味着它能完成更复杂的学习任务。这种神经网络模型广泛应用于模式识别、信号处理、优化组合、异常探测等方面。With the continuous development of neural network research, neural network deep learning algorithms have been proposed in recent years. More hidden layers are introduced into the neural network to form DNN. More hidden layers make DNN more able to describe the complexity of the real world. situation. Theoretically speaking, a model with more parameters has a higher complexity and a greater "capacity", which means that it can complete more complex learning tasks. This neural network model is widely used in pattern recognition, signal processing, optimization combination, anomaly detection and so on.
CNN是一种带有卷积结构的深度神经网络，其结构如图4所示，可以包括输入层410、卷积层420、池化层430、全连接层440、以及输出层450。CNN is a deep neural network with a convolutional structure. Its structure is shown in FIG.
每一个卷积层420可以包括很多个卷积算子，卷积算子也称为核，其作用可以看作是一个从输入信号中提取特定信息的过滤器，卷积算子本质上可以是一个权重矩阵，这个权重矩阵通常被预先定义。Each convolutional layer 420 can include many convolution operators, which are also called kernels, and their function can be regarded as a filter for extracting specific information from the input signal. The convolution operator can be essentially A weight matrix, which is usually predefined.
这些权重矩阵中的权重值在实际应用中需要经过大量的训练得到，通过训练得到的权重值形成的各个权重矩阵可以从输入信号中提取信息，从而帮助CNN进行正确的预测。The weight values in these weight matrices need to be obtained through a lot of training in practical applications, and each weight matrix formed by the weight values obtained through training can extract information from the input signal, thereby helping CNN to make correct predictions.
当CNN有多个卷积层的时候，初始的卷积层往往提取较多的一般特征，该一般特征也可以称之为低级别的特征；随着CNN深度的加深，越往后的卷积层提取到的特征越来越复杂。When CNN has multiple convolutional layers, the initial convolutional layer often extracts more general features, which can also be called low-level features; as the depth of CNN deepens, the later convolution The features extracted by the layers are getting more and more complex.
池化层430，由于常常需要减少训练参数的数量，因此卷积层之后常常需要周期性的引入池化层，例如，可以是图4所示的一层卷积层后面跟一层池化层，也可以是多层卷积层后面接一层或多层池化层。在信号处理过程中，池化层的唯一目的就是减少提取的信息的空间大小。Pooling layer 430, because it is often necessary to reduce the number of training parameters, it is often necessary to periodically introduce a pooling layer after the convolutional layer, for example, it can be a layer of convolutional layer followed by a layer of pooling layer as shown in Figure 4 , can also be a multi-layer convolutional layer followed by one or more pooling layers. In signal processing, the sole purpose of pooling layers is to reduce the spatial size of the extracted information.
全连接层440，在经过卷积层420、池化层430的处理后，CNN还不足以输出所需要的输出信息。因为如前所述，卷积层420、池化层430只会提取特征，并减少输入数据带来的参数。然而为了生成最终的输出信息(例如，发射端发射的原始信息的比特流)，CNN还需要利用全连接层440。通常，全连接层440中可以包括多个隐含层，该多层隐含层中所包含的参数可以根据具体的任务类型的相关训练数据进行预先训练得到，例如，该任务类型可以包括对接收机接收的数据信号进行解码，又例如，该任务类型还可以包括基于接收机接收的导频信号进行信道估计。The fully connected layer 440, after being processed by the convolutional layer 420 and the pooling layer 430, CNN is not enough to output the required output information. Because as mentioned above, the convolutional layer 420 and the pooling layer 430 only extract features and reduce the parameters brought by the input data. However, in order to generate the final output information (eg, the bit stream of the original information transmitted by the transmitter), the CNN also needs to utilize the fully connected layer 440 . Generally, the fully connected layer 440 may include a plurality of hidden layers, and the parameters contained in the multi-layer hidden layers may be pre-trained according to relevant training data of a specific task type, for example, the task type may include receiving For another example, the task type may also include performing channel estimation based on the pilot signal received by the receiver.
在全连接层440中的多层隐含层之后，也就是整个CNN的最后层为输出层450，用于输出结果。通常，该输出层450设置有损失函数(例如,类似分类交叉熵的损失函数)，用于计算预测误差，或者说用于评价CNN模型输出的结果(又称预测值)与理想结果(又称真实值)之间的差异程度。After the multi-layer hidden layers in the fully connected layer 440, that is, the last layer of the entire CNN is the output layer 450 for outputting results. Usually, the output layer 450 is provided with a loss function (for example, a loss function similar to classification cross entropy), which is used to calculate the prediction error, or to evaluate the result (also called predicted value) output by the CNN model and the ideal result (also called The degree of difference between the true value).
为了使损失函数最小化，需要对CNN模型进行训练。在一些实现方式中，可以使用反向传播算法(backpropagation algorithm，BP)对CNN模型进行训练。BP的训练过程由正向传播过程和反向传播过程组成。在正向传播(如图4由410至450的传播为正向传播)过程中，输入数据输入CNN模型的上述各层，经过逐层处理并传向输出层。如果在输出层输出的结果与理想结果差异较大，则将上述损失函数最小化作为优化目标，转入反向传播(如图4由450至410的传播为反向传播)，逐层求出优化目标对各神经元权值的偏导数，构成优化目标对权值向量的梯量，作为修改模型权重的依据，CNN的训练过程在权重修改过程中完成。当上述误差达到所期望值时，CNN的训练过程结束。In order to minimize the loss function, the CNN model needs to be trained. In some implementations, the CNN model may be trained using a backpropagation algorithm (BP). The training process of BP consists of forward propagation process and back propagation process. In the process of forward propagation (the propagation from 410 to 450 in Fig. 4 is forward propagation), the input data is input into the above layers of the CNN model, processed layer by layer and transmitted to the output layer. If the result output at the output layer is quite different from the ideal result, the above loss function is minimized as the optimization goal, and transferred to backpropagation (as shown in Fig. The partial derivative of the optimization target to the weight of each neuron constitutes the gradient of the optimization target to the weight vector, which is used as the basis for modifying the model weight. The training process of CNN is completed in the weight modification process. When the above error reaches the expected value, the training process of CNN ends.
需要说明的是，如图4所示的CNN仅作为一种卷积神经网络的示例，在具体的应用中，卷积神经网络还可以以其他网络模型的形式存在，本申请实施例对此不作限定。It should be noted that the CNN shown in Figure 4 is only an example of a convolutional neural network. In a specific application, the convolutional neural network can also exist in the form of other network models, which are not discussed in this embodiment of the present application. limited.
上文结合图1至图4介绍了无线通信***信号传输过程、信道估计、CSI反馈及自编码器，下文结合图5至图7介绍基于自编码器的信号传输过程、信道估计、CSI反馈。The above describes the signal transmission process, channel estimation, CSI feedback and autoencoder of the wireless communication system in conjunction with Figures 1 to 4. The following describes the signal transmission process, channel estimation, and CSI feedback based on the autoencoder in conjunction with Figures 5 to 7.
六、基于AI解码器的信道估计6. Channel estimation based on AI decoder
基于AI解码器的信道估计旨在利用AI解码器对接收机接收的导频信号进行处理，来实现信道估计。图5示出了基于AI解码器进行信道估计的过程。参见图5，将接收机500接收的导频信号作为AI解码器510的输入，相应地，AI解码器510对输入的导频信号进行处理来输出信道信息。另外，在一些实现方式中，除了导频信号之外还可以增加其他辅助信息来提升AI解码器输出信道信息的准确性。例如，还可以对AI解码器510输入接收机500预存的导频信号的原始序列，接收机500接收导频信号的能量水平，传输导频信号时的传输时延或者传输导频信号时噪声等。The channel estimation based on the AI decoder aims to use the AI decoder to process the pilot signal received by the receiver to realize the channel estimation. Fig. 5 shows the process of channel estimation based on AI decoder. Referring to FIG. 5 , the pilot signal received by the receiver 500 is used as the input of the AI decoder 510 , and accordingly, the AI decoder 510 processes the input pilot signal to output channel information. In addition, in some implementation manners, in addition to the pilot signal, other auxiliary information may be added to improve the accuracy of the channel information output by the AI decoder. For example, the original sequence of the pilot signal pre-stored by the receiver 500 can also be input to the AI decoder 510, the energy level of the pilot signal received by the receiver 500, the transmission delay when transmitting the pilot signal or the noise when transmitting the pilot signal, etc. .
七、基于自编码器的CSI反馈7. CSI Feedback Based on Autoencoder
自编码器中的AI编码器可以对携带CSI的接收信号进行特征提取，自编码器中的AI解码器可以在接收机处尽可能还原发射机压缩反馈的CSI，如此，便可以在不影响CSI传输准确性的同时节约反馈CSI的通信开销。The AI encoder in the autoencoder can perform feature extraction on the received signal carrying CSI, and the AI decoder in the autoencoder can restore the CSI compressed and fed back by the transmitter as much as possible at the receiver, so that it can be used without affecting the CSI While transmitting accuracy, the communication overhead of feedback CSI is saved.
图6示出了基于自编码器的CSI反馈***。如图6所示，整个反馈***中包含自编码器包含AI编码器611及AI解码器621部分，其中，AI编码器611部署在发射机处610，AI解码器621部署在接收机620处。发射机610将待传输的CSI通过AI编码器611进行压缩编码，得到压缩后的CSI。再将压缩后的CSI通过反馈链路反馈给接收机620，接收机620通过AI解码器621对压缩后的CSI进行解码，得到恢复后的CSI。Figure 6 shows an autoencoder based CSI feedback system. As shown in FIG. 6 , the entire feedback system includes an autoencoder including an AI encoder 611 and an AI decoder 621 , wherein the AI encoder 611 is deployed at the transmitter 610 , and the AI decoder 621 is deployed at the receiver 620 . The transmitter 610 compresses and encodes the CSI to be transmitted through the AI encoder 611 to obtain the compressed CSI. Then, the compressed CSI is fed back to the receiver 620 through the feedback link, and the receiver 620 decodes the compressed CSI through the AI decoder 621 to obtain the restored CSI.
八、基于AI解码器的接收机8. Receiver based on AI decoder
通过在接收机设计中引入AI解码器，并利用AI解码器来实现接收机内部对信号的处理过程(例如，解调制，解压缩等)，来提高接收机的性能。图7是基于AI解码器的接收机的示意图。图7所示的接收机700中，AI解码器710的输入为接收机接收的接收信号，输出为解码信号。The performance of the receiver is improved by introducing an AI decoder into the design of the receiver, and using the AI decoder to implement signal processing (for example, demodulation, decompression, etc.) inside the receiver. Fig. 7 is a schematic diagram of an AI decoder based receiver. In the receiver 700 shown in FIG. 7 , the input of the AI decoder 710 is a received signal received by the receiver, and the output is a decoded signal.
从上文的介绍可以看出，基于自编码器的模块化通信***设计是通信***发展的趋势，它可以很好地利用传统通信***模型的先验结构，同时还可以针对自编码器中的AI编码器和/或AI解码器进行灵活地调整和训练。From the above introduction, it can be seen that the design of modular communication system based on autoencoder is the development trend of communication system. It can make good use of the prior structure of the traditional communication system model, and can also target the AI encoders and/or AI decoders can be flexibly adjusted and trained.
目前，在AI解码器上线(即部署至接收机)之前，会基于预设的训练集采用线下训练的方式，对AI解码器进行训练。但是，训练的效果并不理想，由于实际的通信***情况比较复杂，而训练集中的训练数据无法涵盖全部情况，这就使得训练后的AI解码器仅在处理与训练集的训练数据具有相似特征的信号时，可以输出较为准确的解码信号。当接收机实际接收的接收信号与训练数据差异较大时，AI解码器则无法较为准确的为接收信号进行解码，也就是说，线下训练后的AI解码器的泛化能力较差，其中AI解码器的泛化能力用于描述AI解码器对与训练数据差异较大的其他信号进行解码时的准确率。Currently, before the AI decoder goes online (that is, deployed to the receiver), the AI decoder will be trained based on the preset training set using offline training. However, the effect of the training is not ideal, because the actual communication system is more complicated, and the training data in the training set cannot cover all the situations, which makes the trained AI decoder only deal with the training data with similar characteristics to the training set When the signal is low, it can output a more accurate decoded signal. When the received signal actually received by the receiver is quite different from the training data, the AI decoder cannot decode the received signal more accurately, that is to say, the generalization ability of the AI decoder after offline training is poor, among which The generalization ability of the AI decoder is used to describe the accuracy of the AI decoder when decoding other signals that are quite different from the training data.
本申请提供一种信号处理方案，以实现AI解码器的在线训练过程。为了便于理解本申请，下文先结合图8介绍本申请实施例适用的无线通信***，再结合图9介绍本申请实施例的方法。This application provides a signal processing solution to realize the online training process of AI decoder. In order to facilitate the understanding of the present application, the wireless communication system applicable to the embodiment of the present application is firstly introduced below in conjunction with FIG. 8 , and then the method of the embodiment of the present application is introduced in conjunction with FIG. 9 .
图8是本申请实施例适用的无线通信***800。该无线通信***800可以包括网络设备810。网络设备810可以是与终端设备820通信的设备。网络设备810可以为特定的地理区域提供通信覆盖，并且可以与位于该覆盖区域内的终端设备820进行通信。Fig. 8 is a wireless communication system 800 applicable to the embodiment of the present application. The wireless communication system 800 may include a network device 810 . The network device 810 may be a device that communicates with the terminal device 820 . The network device 810 may provide communication coverage for a specific geographical area, and may communicate with the terminal device 820 located within the coverage area.
图8示例性地示出了一个网络设备810和两个终端设备820，可选地，该无线通信***800可以包括多个网络设备并且每个网络设备的覆盖范围内可以包括其它数量的终端设备，本申请实施例对此不做限定。Fig. 8 exemplarily shows a network device 810 and two terminal devices 820, optionally, the wireless communication system 800 may include multiple network devices and each network device may include other numbers of terminal devices within the coverage area , which is not limited in this embodiment of the present application.
可选地，该无线通信***800还可以包括网络控制器、移动管理实体等其他网络实体， 本申请实施例对此不作限定。Optionally, the wireless communication system 800 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.
可选地，终端设备820之间也可以直接通信，例如，两个终端设备820之间可以通过设备到设备(device-to-device，D2D)链路之间通信。Optionally, the terminal devices 820 may also communicate directly, for example, two terminal devices 820 may communicate through a device-to-device (device-to-device, D2D) link.
应理解，本申请实施例的技术方案可以应用于各种通信***，例如：第五代(5th generation，5G)***或新无线(new radio，NR)、长期演进(long term evolution，LTE)***、LTE频分双工(frequency division duplex，FDD)***、LTE时分双工(time division duplex，TDD)等。本申请提供的技术方案还可以应用于未来的通信***，如第六代移动通信***，又如卫星通信***，等等。It should be understood that the technical solutions of the embodiments of the present application can be applied to various communication systems, for example: the fifth generation (5th generation, 5G) system or new radio (new radio, NR), long term evolution (long term evolution, LTE) system , LTE frequency division duplex (frequency division duplex, FDD) system, LTE time division duplex (time division duplex, TDD), etc. The technical solutions provided in this application can also be applied to future communication systems, such as the sixth generation mobile communication system, and satellite communication systems, and so on.
本申请实施例中的终端设备也可以称为用户设备(user equipment，UE)、接入终端、用户单元、用户站、移动站、移动台(mobile station，MS)、移动终端(mobile Terminal，MT)、远方站、远程终端、移动设备、用户终端、终端、无线通信设备、用户代理或用户装置。本申请实施例中的终端设备可以是指向用户提供语音和/或数据连通性的设备，可以用于连接人、物和机，例如具有无线连接功能的手持式设备、车载设备等。本申请的实施例中的终端设备可以是手机(mobile phone)、平板电脑(Pad)、笔记本电脑、掌上电脑、移动互联网设备(mobile internet device，MID)、可穿戴设备，虚拟现实(virtual reality，VR)设备、增强现实(augmented reality，AR)设备、工业控制(industrial control)中的无线终端、无人驾驶(self driving)中的无线终端、远程手术(remote medical surgery)中的无线终端、智能电网(smart grid)中的无线终端、运输安全(transportation safety)中的无线终端、智慧城市(smart city)中的无线终端、智慧家庭(smart home)中的无线终端等。可选地，UE可以用于充当基站。例如，UE可以充当调度实体，其在V2X或D2D等中的UE之间提供侧行链路信号。比如，蜂窝电话和汽车利用侧行链路信号彼此通信。蜂窝电话和智能家居设备之间通信，而无需通过基站中继通信信号。The terminal equipment in the embodiment of the present application may also be referred to as user equipment (user equipment, UE), access terminal, subscriber unit, subscriber station, mobile station, mobile station (mobile station, MS), mobile terminal (mobile Terminal, MT) ), remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent, or user device. The terminal device in the embodiment of the present application may be a device that provides voice and/or data connectivity to users, and can be used to connect people, objects and machines, such as handheld devices with wireless connection functions, vehicle-mounted devices, and the like. The terminal device in the embodiment of the present application can be mobile phone (mobile phone), tablet computer (Pad), notebook computer, palmtop computer, mobile internet device (mobile internet device, MID), wearable device, virtual reality (virtual reality, VR) equipment, augmented reality (augmented reality, AR) equipment, wireless terminals in industrial control, wireless terminals in self driving, wireless terminals in remote medical surgery, smart Wireless terminals in smart grid, wireless terminals in transportation safety, wireless terminals in smart city, wireless terminals in smart home, etc. Optionally, UE can be used to act as a base station. For example, a UE may act as a scheduling entity that provides sidelink signals between UEs in V2X or D2D, etc. For example, a cell phone and an automobile communicate with each other using sidelink signals. Communication between cellular phones and smart home devices without relaying communication signals through base stations.
本申请实施例中的网络设备可以是用于与终端设备通信的设备，该网络设备也可以称为接入网设备或无线接入网设备，如网络设备可以是基站。本申请实施例中的网络设备可以是指将终端设备接入到无线网络的无线接入网(radio access network，RAN)节点(或设备)。基站可以广义的覆盖如下中的各种名称，或与如下名称进行替换，比如：节点B(NodeB)、演进型基站(evolved NodeB，eNB)、下一代基站(next generation NodeB，gNB)、中继站、接入点、传输点(transmitting and receiving point，TRP)、发射点(transmitting point，TP)、主站MeNB、辅站SeNB、多制式无线(MSR)节点、家庭基站、网络控制器、接入节点、无线节点、接入点(access piont，AP)、传输节点、收发节点、基带单元(base band unit，BBU)、射频拉远单元(Remote Radio Unit，RRU)、有源天线单元(active antenna unit，AAU)、射频头(remote radio head，RRH)、中心单元(central unit，CU)、分布式单元(distributed unit，DU)、定位节点等。基站可以是宏基站、微基站、中继节点、施主节点或类似物，或其组合。基站还可以指用于设置于前述设备或装置内的通信模块、调制解调器或芯片。基站还可以是移动交换中心以及设备到设备D2D、车辆外联(vehicle-to-everything，V2X)、机器到机器(machine-to-machine，M2M)通信中承担基站功能的设备、6G网络中的网络侧设备、未来的通信***中承担基站功能的设备等。基站可以支持相同或不同接入技术的网络。本申请的实施例对网络设备所采用的具体技术和具体设备形态不做限定。The network device in this embodiment of the present application may be a device for communicating with a terminal device, and the network device may also be called an access network device or a wireless access network device, for example, the network device may be a base station. The network device in this embodiment of the present application may refer to a radio access network (radio access network, RAN) node (or device) that connects a terminal device to a wireless network. The base station can broadly cover various names in the following, or replace with the following names, such as: Node B (NodeB), evolved base station (evolved NodeB, eNB), next generation base station (next generation NodeB, gNB), relay station, Access point, transmission point (transmitting and receiving point, TRP), transmission point (transmitting point, TP), primary station MeNB, secondary station SeNB, multi-standard wireless (MSR) node, home base station, network controller, access node , wireless node, access point (access piont, AP), transmission node, transceiver node, base band unit (base band unit, BBU), remote radio unit (Remote Radio Unit, RRU), active antenna unit (active antenna unit) , AAU), radio head (remote radio head, RRH), central unit (central unit, CU), distributed unit (distributed unit, DU), positioning nodes, etc. A base station may be a macro base station, a micro base station, a relay node, a donor node, or the like, or a combination thereof. A base station may also refer to a communication module, a modem or a chip configured in the aforementioned equipment or device. The base station can also be a mobile switching center, a device that undertakes the function of a base station in D2D, vehicle-to-everything (V2X), machine-to-machine (M2M) communication, and a device in a 6G network. Network-side equipment, equipment that assumes base station functions in future communication systems, etc. Base stations can support networks of the same or different access technologies. The embodiment of the present application does not limit the specific technology and specific device form adopted by the network device.
基站可以是固定的，也可以是移动的。例如，直升机或无人机可以被配置成充当移动基站，一个或多个小区可以根据该移动基站的位置移动。在其他示例中，直升机或无人机可以被配置成用作与另一基站通信的设备。Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move according to the location of the mobile base station. In other examples, a helicopter or drone may be configured to serve as a device in communication with another base station.
在一些部署中，本申请实施例中的网络设备可以是指CU或者DU，或者，网络设备包括CU和DU。gNB还可以包括AAU。In some deployments, the network device in this embodiment of the present application may refer to a CU or a DU, or, the network device includes a CU and a DU. A gNB may also include an AAU.
网络设备和终端设备可以部署在陆地上，包括室内或室外、手持或车载；也可以部署在水面上；还可以部署在空中的飞机、气球和卫星上。本申请实施例中对网络设备和终端 设备所处的场景不做限定。Network equipment and terminal equipment can be deployed on land, including indoors or outdoors, hand-held or vehicle-mounted; they can also be deployed on water; they can also be deployed on aircraft, balloons and satellites in the air. In this embodiment of the application, there is no limitation on the scenarios where network devices and terminal devices are located.
应理解，本申请中涉及到的通信设备，可以为网络设备，或者也可以为终端设备。例如，第一通信设备为网络设备，第二通信设备为终端设备。又如，第一通信设备为终端设备，第二通信设备为网络设备。又如，第一通信设备和第二通信设备均为网络设备，或者均为终端设备。It should be understood that the communication device mentioned in this application may be a network device, or may also be a terminal device. For example, the first communication device is a network device, and the second communication device is a terminal device. In another example, the first communication device is a terminal device, and the second communication device is a network device. In another example, both the first communication device and the second communication device are network devices, or both are terminal devices.
还应理解，本申请中的通信设备的全部或部分功能也可以通过在硬件上运行的软件功能来实现，或者通过平台(例如云平台)上实例化的虚拟化功能来实现。It should also be understood that all or part of the functions of the communication device in this application may also be realized by software functions running on hardware, or by virtualization functions instantiated on a platform (such as a cloud platform).
图9是本申请实施例的信号处理的方法的流程图。图9所示的方法包括步骤S910至步骤S940。应理解，图9所示的接收机可以是上文介绍的终端设备，相应地，发射机也可以是上文介绍的终端设备，此时，接收机和发射机之间可以通过D2D链路通信。当然，接收机为终端设备时，发射机也可以是网络设备。FIG. 9 is a flowchart of a signal processing method according to an embodiment of the present application. The method shown in FIG. 9 includes steps S910 to S940. It should be understood that the receiver shown in FIG. 9 may be the terminal device described above, and correspondingly, the transmitter may also be the terminal device described above. In this case, the receiver and the transmitter may communicate through the D2D link . Certainly, when the receiver is a terminal device, the transmitter may also be a network device.
在步骤S910，接收机对发射机发射的无线信号进行接收，得到接收信号。In step S910, the receiver receives the wireless signal transmitted by the transmitter to obtain a received signal.
在步骤S920，接收机将接收信号输入AI解码器进行解码，得到解码信号。In step S920, the receiver inputs the received signal into the AI decoder for decoding to obtain a decoded signal.
在步骤S930，接收机根据解码信号，生成接收信号的恢复信号。In step S930, the receiver generates a recovery signal of the received signal according to the decoded signal.
在本申请实施例中，上述生成恢复信号的方式不作限定。在一些实施例中，可以通过对解码信号进行简单的信号处理，以得到恢复信号。在另一些实施例中，可以将解码信号输入AI编码器，以生成接收信号的恢复信号。下文将结合图14至图16具体介绍，为了简洁，在此不再赘述。In this embodiment of the present application, the foregoing manner of generating the recovery signal is not limited. In some embodiments, the restored signal can be obtained by performing simple signal processing on the decoded signal. In other embodiments, the decoded signal may be input to an AI encoder to generate a recovered signal of the received signal. The following will introduce in detail in conjunction with FIG. 14 to FIG. 16 , and for the sake of brevity, details will not be repeated here.
在步骤S940，接收机根据恢复信号与接收信号之间的差异，对AI解码器进行在线训练。其中，恢复信号与接收信号之间的差异可以通过恢复信号与接收信号之间的误差表示。In step S940, the receiver performs online training for the AI decoder according to the difference between the restored signal and the received signal. Wherein, the difference between the restored signal and the received signal can be represented by an error between the restored signal and the received signal.
在一些实现方式中，上述步骤S940可以包括接收机将恢复信号与接收信号之间的差异最小化作为优化目标，更新AI解码器的权重，或者说对AI解码器的权重进行优化，以对AI解码器进行在线训练。In some implementations, the above step S940 may include the receiver taking the minimization of the difference between the recovered signal and the received signal as an optimization goal, updating the weights of the AI decoder, or optimizing the weights of the AI decoder to optimize the AI The decoder is trained online.
上述AI解码器为上文中介绍的神经网络模型时，上述AI解码器的权重可以理解为神经网络模型中的每层模型对输入数据进行处理时使用的权重。例如，参见图4，AI解码器的权重可以包括输入层410、卷积层420、池化层430、全连接层440以及输出层450对各层的输入进行处理使用的权重。When the aforementioned AI decoder is the neural network model introduced above, the weight of the aforementioned AI decoder can be understood as the weight used by each layer model in the neural network model to process the input data. For example, referring to FIG. 4 , the weights of the AI decoder may include weights used by the input layer 410 , the convolutional layer 420 , the pooling layer 430 , the fully connected layer 440 and the output layer 450 to process the input of each layer.
上述更新AI解码器中的权重可以理解为，更新AI解码器的全部的权重，或者更新AI解码器的部分的权重。The aforementioned updating of the weights in the AI decoder can be understood as updating all the weights of the AI decoder, or updating the weights of some parts of the AI decoder.
上述恢复信号与接收信号之间的差异，可以通过恢复信号与接收信号之间的误差表示。在一些实现方式中，上述恢复信号与接收信号之间的误差可以通过损失函数计算，或者说，损失函数表示恢复信号与接收信号之间的误差。The above-mentioned difference between the restored signal and the received signal can be represented by an error between the restored signal and the received signal. In some implementation manners, the error between the restored signal and the received signal may be calculated by using a loss function, or in other words, the loss function represents the error between the restored signal and the received signal.
通常，为了使损失函数最小化，需要对AI解码器进行在线训练。例如，可以使用BP对AI解码器进行训练。BP的训练过程可以由正向传播过程和反向传播过程组成。以AI解码器为图4所示的模型为例介绍，在正向传播(如图4由410至450的传播为正向传播)过程中，接收信号输入AI解码器模型的上述各层，经过逐层处理并传向输出层，相应地输出层输出解码信号，接收机基于解码信号得到接收信号的恢复信号。如果接收信号与恢复信号之间差异较大，则将上述损失函数最小化作为优化目标，转入反向传播(如图4由450至410的传播为反向传播)，逐层求出优化目标对各神经元权值的偏导数，构成优化目标对权值向量的梯量，作为修改AI解码器的权重的依据，AI解码器在线训练过程在权重修改过程中完成。当上述误差达到所期望值时，AI解码器的在线训练过程结束。当然，本申请对AI解码器的在线训练过程也可以使用其他已知的训练原理，本申请实施例对此不作限定。Usually, an AI decoder needs to be trained online in order to minimize the loss function. For example, AI decoders can be trained using BP. The training process of BP can be composed of forward propagation process and back propagation process. Taking the AI decoder as the model shown in Figure 4 as an example, in the process of forward propagation (the propagation from 410 to 450 in Figure 4 is forward propagation), the received signal is input to the above-mentioned layers of the AI decoder model, and passed through It is processed layer by layer and transmitted to the output layer, and the output layer outputs the decoded signal accordingly, and the receiver obtains the recovery signal of the received signal based on the decoded signal. If there is a large difference between the received signal and the restored signal, the above loss function is minimized as the optimization goal, and transferred to backpropagation (as shown in Figure 4, the propagation from 450 to 410 is backpropagation), and the optimization goal is calculated layer by layer The partial derivative of the weight of each neuron constitutes the gradient of the optimization target to the weight vector, which is used as the basis for modifying the weight of the AI decoder. The online training process of the AI decoder is completed during the weight modification process. When the above error reaches the expected value, the online training process of the AI decoder ends. Certainly, other known training principles may also be used in the online training process of the AI decoder in the present application, which is not limited in this embodiment of the present application.
在本申请实施例中，接收机使用AI解码器对接收信号解码，得到解码信号，再根据解码信号生成接收信号的恢复信号，并基于接收信号与恢复信号之间的差异，对AI解码器进行在线训练，由于接收信号和恢复信号都是可以被接收机获取的，因此，本申请实施 例的方法可以实现AI解码器的在线训练过程，有利于提高AI解码器的泛化能力。或者说，在AI解码器上线之后，利用接收机实际接收的接收信号对AI解码器进行在线训练，这样，即使接收信号与训练数据有较大差异，在线训练后的AI解码器在对接收信号进行解码时，也能有较高的准确率，有利于提高AI解码器的泛化能力。In the embodiment of the present application, the receiver uses the AI decoder to decode the received signal to obtain the decoded signal, and then generates the restored signal of the received signal according to the decoded signal, and based on the difference between the received signal and the restored signal, the AI decoder performs For online training, since both the received signal and the restored signal can be obtained by the receiver, the method in the embodiment of the present application can realize the online training process of the AI decoder, which is beneficial to improve the generalization ability of the AI decoder. In other words, after the AI decoder goes online, the AI decoder is trained online using the received signal actually received by the receiver. In this way, even if the received signal is quite different from the training data, the AI decoder after online training will be able to use the received signal When decoding, it can also have a high accuracy rate, which is conducive to improving the generalization ability of the AI decoder.
本申请实施例的方案相对于基于发射机待发射的原始信号与接收机接收的接收信号之间的差异，对AI解码器进行训练的方案而言，避免了由于接收机和发射机分布式部署，原始信号无法被接收机准确获知，而导致AI解码器无法在接收机中进行在线训练的问题。Compared with the scheme of training the AI decoder based on the difference between the original signal to be transmitted by the transmitter and the received signal received by the receiver, the scheme of the embodiment of the present application avoids the distributed deployment of the receiver and the transmitter. , the original signal cannot be accurately known by the receiver, which leads to the problem that the AI decoder cannot be trained online in the receiver.
在线训练的时间可以由网络设备为接收机配置。在一些实现方式中，网络设备可以向接收机发送在线训练的指示信息，其中，在线训练的指示信息用于指示AI解码器进行在线训练的时间，(为了便于理解，下文将AI解码器进行在线训练的时间称为“在线训练时间”)。当然，接收机也可以按照通信协议规定的在线训练时间进行在线训练。本申请实施例对此不作具体限定。The time for online training can be configured by the network device for the receiver. In some implementations, the network device may send online training instruction information to the receiver, where the online training instruction information is used to indicate the time for the AI decoder to perform online training, (for ease of understanding, the AI decoder will be online The training time is called "online training time"). Of course, the receiver can also perform online training according to the online training time stipulated in the communication protocol. This embodiment of the present application does not specifically limit it.
上述在线训练的指示信息可以通过多种方式指示在线训练的时间。例如，在线训练的指示信息可以通过指示在线训练周期，指示接收机周期性地为AI解码器进行在线训练。又例如，上述在线训练的指示信息还可以直接指示在线训练的起始时间和/或结束时间。The indication information of the above online training may indicate the time of the online training in various ways. For example, the indication information of online training may indicate the online training period to instruct the receiver to periodically perform online training for the AI decoder. For another example, the above online training indication information may also directly indicate the start time and/or end time of the online training.
在一些实现方式中，上述在线训练的指示信息可以通过无线资源控制(radio resource control，RRC)信令传输，在另一些实现方式中，上述在线训练的指示信息还可以通过控制信道传输，本申请实施例对此不作限定。In some implementation manners, the indication information of the above-mentioned online training may be transmitted through radio resource control (RRC) signaling, and in other implementation manners, the indication information of the above-mentioned online training may also be transmitted through a control channel. The embodiment does not limit this.
基于在线训练的指示方式的不同，AI解码器的在线训练配置方式可以分为周期性在线训练、非持续性在线训练以及非周期性在线训练。下文结合图10至图12介绍以上三种在线训练配置方式。Based on the different instruction methods of online training, the online training configuration methods of AI decoders can be divided into periodic online training, non-persistent online training, and aperiodic online training. The above three online training configuration methods are introduced below in conjunction with Fig. 10 to Fig. 12 .
在线训练配置方式一、周期性在线训练。即接收机基于配置的在线训练周期，周期性地对AI解码器进行在线训练。Online training configuration method 1. Periodic online training. That is, the receiver periodically performs online training for the AI decoder based on the configured online training period.
由于在线训练需要基于发射机和接收机之间实际传输的信号(即上文的接收信号)进行，因此，为了简化在线训练周期的配置方式，在一些实现方式中，上述在线训练周期可以包含多个实际传输的信号的传输周期。相应地，接收机可以在这多个传输周期中任意的传输周期中进行在线训练。通常，为了保证AI解码器解码的准确性，可以在多个传输周期中的第一个传输周期进行在线训练，这样，在第一个传输周期之后的几个传输周期中可以使用训练后的AI解码器解码。Since the online training needs to be performed based on the signal actually transmitted between the transmitter and the receiver (that is, the received signal above), in order to simplify the configuration of the online training period, in some implementations, the above online training period can include multiple The transmission period of the actually transmitted signal. Correspondingly, the receiver can perform online training in any transmission period among the multiple transmission periods. Usually, in order to ensure the accuracy of AI decoder decoding, online training can be performed in the first transmission cycle of multiple transmission cycles, so that the trained AI can be used in several transmission cycles after the first transmission cycle The decoder decodes.
如上文所述，接收机可以在多个传输周期中的任意传输周期进行在线训练，也就是说，在线训练周期包括第一类型的传输周期和第二类型的传输周期，其中，在第一类型的传输周期中，接收机可以对AI解码器进行在线训练，并利用训练后的AI解码器对当前传输周期接收的接收信号进行解码。在第二类型的传输周期中，接收机可以直接采用AI解码器对接收信号解码，而无需对AI解码器进行在线训练。通常，为了提高在线训练的准确性，第二类型的传输周期还可以用于收集接收信号作为在线训练的训练数据。As mentioned above, the receiver can perform online training in any of the multiple transmission periods, that is to say, the online training period includes the first type of transmission period and the second type of transmission period, wherein, in the first type In the transmission period of , the receiver can perform online training on the AI decoder, and use the trained AI decoder to decode the received signal received in the current transmission period. In the second type of transmission period, the receiver can directly use the AI decoder to decode the received signal without performing online training on the AI decoder. Usually, in order to improve the accuracy of online training, the second type of transmission cycle can also be used to collect received signals as training data for online training.
需要说明的是，上述第一类型的传输周期和第二类型的传输周期在在线训练周期中的排布方式可以是基于通信协议预先配置的。当然，也可以是网络设备通过信令指示的，本申请实施例对此不作限定。It should be noted that, the arrangement manner of the first type of transmission period and the second type of transmission period in the online training period may be pre-configured based on the communication protocol. Of course, it may also be indicated by the network device through signaling, which is not limited in this embodiment of the present application.
另外，本申请实施例对上述在线训练周期包含的传输周期的数量不作具体限定。在一些实现方式中，在线训练周期可以包含较少的传输周期，相应地，在线训练的频率会较高，这种在线训练周期的配置适用于变化比较频繁的通信场景，例如，高铁等。在另一些实施例中，在线训练周期可以包含较多的传输周期，相应地，在线训练的频率会较低，有利于降低在线训练所需的计算资源。这种在线训练周期的配置适用于变化不太频繁的通信场景，例如，写字楼等。In addition, the embodiment of the present application does not specifically limit the number of transmission periods included in the online training period. In some implementation manners, the online training period may include fewer transmission periods, and accordingly, the frequency of online training will be higher. This configuration of the online training period is suitable for communication scenarios with frequent changes, such as high-speed rail. In some other embodiments, the online training period may include more transmission periods, and accordingly, the frequency of the online training will be lower, which is beneficial to reduce the computing resources required for the online training. This online training cycle configuration is suitable for communication scenarios that change less frequently, for example, office buildings and the like.
下文以图10所示的周期性在线训练为例，介绍本申请实施例的信号处理的方法。如图10所示，在线训练周期r 1包括4个传输周期t 1，且4个传输周期t 1中的第一个传输周 期作为第一类型的传输周期，其他3个传输周期作为第二类型的传输周期。 The following takes the periodic online training shown in FIG. 10 as an example to introduce the signal processing method of the embodiment of the present application. As shown in Figure 10, the online training period r 1 includes 4 transmission periods t 1 , and the first transmission period of the 4 transmission periods t 1 is used as the first type of transmission period, and the other 3 transmission periods are used as the second type transmission cycle.
在第一类型的传输周期中，接收机需要基于训练数据对AI解码器进行在线训练，再利用在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。In the first type of transmission cycle, the receiver needs to perform online training on the AI decoder based on the training data, and then use the AI decoder after the online training to decode the received signal received in the current transmission cycle.
在第二类型的传输周期中，接收机可以直接采用上述在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。同时，接收机可以收集当前传输周期接收到的接收信号作为后续在线训练的训练数据。In the second type of transmission period, the receiver may directly use the aforementioned online trained AI decoder to decode the received signal received in the current transmission period. At the same time, the receiver can collect received signals received in the current transmission cycle as training data for subsequent online training.
在线训练配置方式二、非持续性在线训练。即，在周期性在线训练的基础上，可以通过在线训练的指示信息指示在线训练的起始时间和结束时间，以提高配置在线训练的灵活性。Online training configuration method 2. Non-continuous online training. That is, on the basis of periodic online training, the online training instruction information may indicate the start time and end time of online training, so as to improve the flexibility of configuring online training.
在该种训练方式中，周期性在线训练可以参见上文关于在线训练配置方式一的介绍，为了简洁，下文重点介绍指示在线训练的起始时间和结束时间的方式。In this training mode, periodic online training can refer to the introduction of online training configuration mode 1 above. For the sake of brevity, the following will focus on the way to indicate the start time and end time of online training.
在线训练的指示信息可以通过显示或隐示的方式指示在线训练的起始时间。例如，在线训练的指示信息可以直接携带起始时间。又例如，接收机可以以传输在线训练的指示信息所在的时域单元为起点，偏移预设的时域单元后得到偏移后的时域单元，这偏移后的时域单元便为起始时间，其中时域单元例如可以是时隙、子帧等。The indication information of the online training may indicate the start time of the online training in a display or implicit manner. For example, the instruction information of the online training may directly carry the start time. For another example, the receiver can start from the time-domain unit where the indication information for online training is transmitted, and offset the preset time-domain unit to obtain the shifted time-domain unit, which is the starting point start time, where the time domain unit may be, for example, a time slot, a subframe, and the like.
相应地，在线训练的指示信息也可以通过显示或隐示的方式指示在线训练的结束时间。例如，在线训练的指示信息可以直接携带结束时间。又例如，在线训练的指示信息可以通过指示在线训练周期的总周期数指示结束时间。Correspondingly, the indication information of the online training may also indicate the end time of the online training in a display or implicit manner. For example, the indication information of the online training may directly carry the end time. For another example, the indication information of the online training may indicate the end time by indicating the total number of periods of the online training period.
下文以图11所示的非持续性在线训练为例，介绍本申请实施例的信号处理的方法。如图11所示，假设在线训练指示信息指示在线训练的起始时间为第i个传输周期t i的起始时间，在线训练的总周期数为3，且在线训练周期r 2包括2个传输周期，则在线训练周期r 2的起始时间为第i个传输周期t i的起始时间，在线训练周期r 2的结束时间为第i+6个传输周期t i+6的结束时间。 The following takes the non-persistent online training shown in FIG. 11 as an example to introduce the signal processing method of the embodiment of the present application. As shown in Figure 11, assume that the online training indication information indicates that the start time of the online training is the start time of the i-th transmission cycle t i , the total number of cycles of the online training is 3, and the online training cycle r 2 includes 2 transmission period, the start time of the online training period r2 is the start time of the i-th transmission period t i , and the end time of the online training period r2 is the end time of the i+6th transmission period t i+6 .
每个在线训练周期r 2中的第一个传输周期为第一类型的传输周期，在第一类型的传输周期中，接收机需要基于训练数据对AI解码器进行在线训练，在利用在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。 The first transmission period in each online training period r 2 is the first type of transmission period. In the first type of transmission period, the receiver needs to perform online training on the AI decoder based on the training data. After using the online training The AI decoder decodes the received signal received in the current transmission cycle.
每个在线训练周期r 2中的第二个传输周期为第二类型的传输周期，在第二类型的传输周期中，接收机可以直接采用上述在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。同时，接收机可以收集当前传输周期接收到的接收信号作为后续在线训练的训练数据。 The second transmission period in each online training period r 2 is the second type of transmission period. In the second type of transmission period, the receiver can directly use the above-mentioned AI decoder after online training to receive the current transmission period The received signal is decoded. At the same time, the receiver can collect received signals received in the current transmission cycle as training data for subsequent online training.
在线训练配置方式三、非周期性在线训练。即每次的在线训练都需要通过在线训练的指示信息来进行。Online training configuration method 3. Aperiodic online training. That is, each online training needs to be performed through the instruction information of the online training.
在一些实现方式中，可以在在线训练的指示信息中携带在线训练的起始时间和在线训练的结束时间，相应地，接收机在起始时间与结束时间指示的时间段内进行在线训练。在另一些实现方式中，也可以仅在在线训练的指示信息中携带在线训练的起始时间，相应地，接收机在接收在线训练的指示信息后，便在起始时间进行一次在线训练。In some implementation manners, the online training start time and the online training end time may be carried in the indication information of the online training, and correspondingly, the receiver performs the online training within the time period indicated by the start time and the end time. In some other implementation manners, only the online training start time may be carried in the online training indication information, correspondingly, after receiving the online training indication information, the receiver performs an online training at the start time.
上述在线训练的指示信息可以通过显示或隐示的方式指示在线训练的起始时间。例如，在线训练的指示信息可以直接携带起始时间。又例如，接收机可以以传输在线训练的指示信息所在的时隙为起点，偏移预设的时隙后得到偏移后的时隙，这偏移后的时隙便为起始时间。The indication information of the above-mentioned online training may indicate the start time of the online training in a display or implicit manner. For example, the instruction information of the online training may directly carry the start time. For another example, the receiver may start from the time slot where the instruction information for online training is transmitted, and obtain a shifted time slot after shifting the preset time slot, and the shifted time slot is the start time.
上述在线训练的指示信息可以通过显示或隐示的方式指示在线训练的结束时间。例如，在线训练的指示信息可以直接携带结束时间。又例如，接收机可以以传输在线训练的指示信息所在的时隙为起点，偏移预设的时隙后得到偏移后的时隙，这偏移后的时隙便为结束时间。The indication information of the above-mentioned online training may indicate the end time of the online training in a display or implicit manner. For example, the indication information of the online training may directly carry the end time. For another example, the receiver may start from the time slot where the indication information of the online training is transmitted, shift the preset time slot to obtain the shifted time slot, and the shifted time slot is the end time.
下文以图12所示的非周期性在线训练为例，介绍本申请实施例的信号处理的方法。如图12所示，假设在线训练的指示信息 1指示在传输周期t i进行在线训练，在线训练的指 示信息 2指示在传输周期t i+4进行在线训练进行在线训练。 The following uses the aperiodic online training shown in FIG. 12 as an example to introduce the signal processing method of the embodiment of the present application. As shown in FIG. 12 , it is assumed that online training instruction information 1 indicates that online training is performed in a transmission period t i , and online training instruction information 2 indicates that online training is performed in a transmission period t i+4 for online training.
当接收机接收到在线训练的指示信息 1后，在传输周期t i，接收机需要基于训练数据对AI解码器进行在线训练，在利用在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。 When the receiver receives the instruction information 1 of online training, in the transmission cycle t i , the receiver needs to perform online training on the AI decoder based on the training data, and use the AI decoder after the online training to analyze the received data in the current transmission cycle The received signal is decoded.
在传输周期t i+1到传输周期t i+3，接收机直接使用上述在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。 From the transmission period t i+1 to the transmission period t i+3 , the receiver directly uses the aforementioned online trained AI decoder to decode the received signal received in the current transmission period.
当接收机接收到在线训练的指示信息 2后，在传输周期t i+4，接收机继续基于训练数据对AI解码器进行在线训练，在利用在线训练后的AI解码器，对当前传输周期接收到的接收信号进行解码。 After the receiver receives the instruction information 2 of online training, in the transmission cycle t i+4 , the receiver continues to perform online training on the AI decoder based on the training data, and uses the AI decoder after the online training to receive The received signal is decoded.
在上文介绍的在线训练过程中，接收机可以收集接收信号作为在线训练的训练数据，但是在一些情况下，接收机收集的接收信号的数量不足以对AI解码器进行在线训练，此时，可以额外选取接收机预存的训练集中的部分或全部训练数据进行在线训练。在一些实现方式中，预存的训练集可以是线下训练AI解码器使用的数据集。In the online training process described above, the receiver can collect received signals as training data for online training, but in some cases, the number of received signals collected by the receiver is not enough for online training of the AI decoder. At this time, Part or all of the training data in the training set pre-stored in the receiver can be additionally selected for online training. In some implementation manners, the pre-stored training set may be a data set used for offline training of the AI decoder.
需要说明的是，额外选择的训练集中的训练数据越少，使用接收机收集的接收信号的数量越多，则在线训练后的AI解码器的泛化性能越高。It should be noted that the less training data in the additionally selected training set and the greater the number of received signals collected using the receiver, the higher the generalization performance of the AI decoder after online training.
在一些实现方式中，在线训练使用的训练数据的大小可以是预先规定的。在另一些实施例中，在线训练使用的训练数据的大小还可以是由网络设备指示。例如，网络设备可以通过在线训练指示信息指示第一类型的传输周期所使用的训练数据的大小。又例如，网络设备可以通过其他信息指示第一类型的传输周期所使用的训练数据的大小。In some implementations, the size of the training data used for online training may be predetermined. In some other embodiments, the size of the training data used for the online training may also be indicated by the network device. For example, the network device may indicate the size of the training data used in the first type of transmission period through the online training indication information. For another example, the network device may use other information to indicate the size of the training data used in the first type of transmission period.
上述在线训练使用的训练数据的大小可以理解为每次在线训练使用全部训练数据的总量。在一些实现方式中，如果接收机收集的接收信号的数量不足以对AI解码器进行在线训练，上述在线训练使用的训练数据的大小包括接收机收集的接收信号的数量与训练集中选择的训练数据的数量之和。在另一些实现方式中，如果接收机收集的接收信号的数量充足，上述在线训练使用的训练数据的大小可以仅包括作为训练数据的接收信号的数量。The size of the training data used in the aforementioned online training can be understood as the total amount of all training data used in each online training. In some implementations, if the number of received signals collected by the receiver is not enough for online training of the AI decoder, the size of the training data used in the above-mentioned online training includes the number of received signals collected by the receiver and the training data selected in the training set sum of the quantities. In some other implementation manners, if the number of received signals collected by the receiver is sufficient, the size of the training data used for the above online training may only include the number of received signals used as training data.
需要说明的是，上述网络设备指示训练数据的大小的方案可以与上文介绍的三种在线配置方式中的任一种结合使用。例如，与在线训练配置方式一结合使用时，上述在线训练指示信息可以包括第一指示信息指示AI解码器进行在线训练的在线训练周期，上述在线训练指示信息还可以包括第二指示信息指示第一类型的传输周期所使用的训练数据的大小。It should be noted that, the above solution of indicating the size of the training data by the network device can be used in combination with any of the three online configuration methods introduced above. For example, when used in combination with online training configuration mode one, the online training instruction information may include the first instruction information indicating the online training period for the AI decoder to perform online training, and the online training instruction information may also include the second instruction information indicating the first The size of the training data used by the type of transfer cycle.
上文结合图10至图12介绍了本申请实施例的信号处理的方法，下文结合图13以AI解码器在信道估计过程中的应用为例，介绍信号处理的方法。需要说明的是，下文介绍的在线训练过程可以与上文中任意一种在线训练配置方式结合。为了简洁，下文不再赘述。The signal processing method of the embodiment of the present application is introduced above with reference to FIG. 10 to FIG. 12 , and the signal processing method is introduced below with reference to FIG. 13 by taking the application of the AI decoder in the channel estimation process as an example. It should be noted that the online training process described below can be combined with any of the online training configuration methods above. For the sake of brevity, no further details are given below.
在信道估计过程中，上述接收信号包含发射机发射的导频信号，上述步骤S920包括接收机将接收信号输入AI解码器进行解码，以进行信道估计，得到解码信号，解码信号包含估计出的第一信道信息，上述步骤S930包括接收机根据第一信道信息和第一随机噪声对接收机存储的导频信号进行处理，得到恢复信号。In the channel estimation process, the above-mentioned received signal includes the pilot signal transmitted by the transmitter. The above-mentioned step S920 includes that the receiver inputs the received signal into the AI decoder for decoding, so as to perform channel estimation and obtain a decoded signal. The decoded signal includes the estimated first A channel information, the above step S930 includes the receiver processing the pilot signal stored in the receiver according to the first channel information and the first random noise to obtain a restored signal.
通常，可以配置发射机发射的导频信号与接收机存储的导频信号相同，以便于接收机基于接收信号与存储的导频信号比较，以进行信道估计，得到估计出的信道信息。Generally, the pilot signal transmitted by the transmitter can be configured to be the same as the pilot signal stored by the receiver, so that the receiver can perform channel estimation based on the comparison between the received signal and the stored pilot signal, and obtain estimated channel information.
上述第一随机噪声可以是接收机预存的，也可以是接收机预先生成的，本申请实施例对此不作限定。The foregoing first random noise may be pre-stored by the receiver, or may be pre-generated by the receiver, which is not limited in this embodiment of the present application.
为了提高AI解码器估计出的信道信息的准确性，在一些实现方式中，还可以将接收机存储的导频信号与接收信号一起输入AI解码器。当然，也可以仅向AI解码器输入接收信号，本申请实施例对此不作限定。In order to improve the accuracy of the channel information estimated by the AI decoder, in some implementation manners, the pilot signal stored in the receiver may also be input into the AI decoder together with the received signal. Of course, it is also possible to only input the received signal to the AI decoder, which is not limited in this embodiment of the present application.
另外，由于获取实际通信环境中的噪声难度较大，在上述得到恢复信号的过程中，接收机可以基于预存的多种不同的随机噪声对导频信号进行处理，得到多个恢复信号。再从多个恢复信号中选择与接收信号最为相似的恢复信号，根据基于相似的恢复信号与接收信 号之间的差异，对AI解码器在线训练。In addition, because it is difficult to obtain the noise in the actual communication environment, in the process of obtaining the restored signal, the receiver can process the pilot signal based on various pre-stored random noises to obtain multiple restored signals. Then select the recovery signal that is most similar to the received signal from multiple recovered signals, and train the AI decoder online based on the difference between the similar recovery signal and the received signal.
下文结合图13，介绍本申请实施例的信道估计过程中信号处理的方法。The following describes the signal processing method in the channel estimation process of the embodiment of the present application with reference to FIG. 13 .
如图13所示，发射机1310将导频信号p通过信道H传输，在传输过程中，由于实际信道中噪声N的存在，因此，接收机1320收集的接收信号Y p可以表示为Y p＝H*p+N。 As shown in FIG. 13 , the transmitter 1310 transmits the pilot signal p through the channel H. During the transmission process, due to the existence of noise N in the actual channel, the received signal Y p collected by the receiver 1320 can be expressed as Y p = H*p+N.
接收机1320将收集的接收信号Y p输入AI解码器1321，得到信道信息H’。然后，接收机1320基于信道信息H’以及随机噪声N’对接收机1320存储的导频信号p进行处理，得到恢复信号Y p’，即Y p’＝H’*p+N’。最后，接收机1320基于恢复信号Y p’与接收信号Y p之间的差异，对AI解码器1321进行在线训练。 The receiver 1320 inputs the collected received signal Y p into the AI decoder 1321 to obtain channel information H'. Then, the receiver 1320 processes the pilot signal p stored in the receiver 1320 based on the channel information H' and the random noise N' to obtain a restored signal Y p ', that is, Y p '=H'*p+N'. Finally, the receiver 1320 performs online training for the AI decoder 1321 based on the difference between the recovered signal Yp ' and the received signal Yp .
如前文介绍，自编码器还可以包括AI编码器。对于包含AI编码器以及AI解码器的自编码器而言，通常采用联合训练的方式对AI编码器以及AI解码器进行训练。即基于AI编码器的输入信号与AI解码器的输出信号之间的差异，对AI编码器以及AI解码器形成的自编码器进行联合训练。As introduced earlier, autoencoders can also include AI encoders. For an autoencoder that includes an AI encoder and an AI decoder, the AI encoder and the AI decoder are usually trained in a joint training manner. That is, based on the difference between the input signal of the AI encoder and the output signal of the AI decoder, joint training is performed on the self-encoder formed by the AI encoder and the AI decoder.
同样，上线后的AI编码器以及AI解码器分别被部署在接收机和发射机中，由于接收机和发射机的分布式部署，AI编码器的输入无法在接收机处获知，导致上述联合训练只能在线下基于训练集进行。如图14所示，AI编码器1411部署在发射机1410处，用于对原始信号s编码，得到编码信号q。发射机1410通过无线链路将编码信号q发射至接收机1420。相应地，接收机1420将接收到的编码信号q输入AI解码器1421解码，得到解码信号s’。在线训练过程中，需要基于原始信号s与解码信号s’之间的差异进行，而原始信号s无法在发射机1410和接收机1420之间的通信链路准确传输，因此，无法以在线训练的方式，对AI编码器1411和AI解码器1421进行联合训练，其中，联合训练可以理解为将AI编码器和AI解码器作为一个整体，对AI编码器和AI解码器中的权重进行更新。Similarly, the online AI encoder and AI decoder are deployed in the receiver and transmitter respectively. Due to the distributed deployment of the receiver and transmitter, the input of the AI encoder cannot be known at the receiver, resulting in the joint training It can only be done offline based on the training set. As shown in FIG. 14, the AI encoder 1411 is deployed at the transmitter 1410, and is used to encode the original signal s to obtain an encoded signal q. The transmitter 1410 transmits the encoded signal q to the receiver 1420 via a wireless link. Correspondingly, the receiver 1420 inputs the received coded signal q to the AI decoder 1421 for decoding to obtain a decoded signal s'. In the online training process, it needs to be based on the difference between the original signal s and the decoded signal s', and the original signal s cannot be accurately transmitted in the communication link between the transmitter 1410 and the receiver 1420, so the online training cannot In this way, the AI encoder 1411 and the AI decoder 1421 are jointly trained, wherein the joint training can be understood as taking the AI encoder and the AI decoder as a whole, and updating the weights in the AI encoder and the AI decoder.
为了提高自编码器的泛化能力。本申请实施例还提供一种信号处理方法，以在线训练的方式，对AI编码器和AI解码器进行联合训练。通过在接收机中配置一个AI编码器，这样，可以对配置的AI编码器与接收机原本的AI解码器进行联合训练，即基于AI解码器的输入与AI编码器的输出之间的差异，对AI编码器和AI解码器进行联合训练。也就是说，上述步骤S930包括：接收机将解码信号输入AI编码器，生成恢复信号。In order to improve the generalization ability of the autoencoder. The embodiment of the present application also provides a signal processing method for jointly training an AI encoder and an AI decoder in an online training manner. By configuring an AI encoder in the receiver, the configured AI encoder and the original AI decoder of the receiver can be jointly trained, that is, based on the difference between the input of the AI decoder and the output of the AI encoder, Joint training of AI encoder and AI decoder. That is to say, the above step S930 includes: the receiver inputs the decoded signal to the AI encoder to generate a restored signal.
在一些实现方式中，为了提高联合训练的准确性，上述在接收机配置的AI编码器可以与发射机中的AI编码器具有相同的模型结构，和/或，相同的模型参数，例如，相同的模型权重等。当然，如果联合训练的目标是为了重点提高AI解码器的泛化性能，可以在接收机中配置一个较为简单的AI编码器。或者，还可以在联合训练的过程中冻结AI编码器的参数，仅训练AI解码器的参数。In some implementations, in order to improve the accuracy of joint training, the AI encoder configured in the receiver may have the same model structure as the AI encoder in the transmitter, and/or the same model parameters, for example, the same model weights etc. Of course, if the goal of joint training is to focus on improving the generalization performance of the AI decoder, a simpler AI encoder can be configured in the receiver. Alternatively, it is also possible to freeze the parameters of the AI encoder during joint training and only train the parameters of the AI decoder.
如果希望采用联合训练的方式，整体提高发射机中AI编码器和接收机中AI解码器的泛化性能，可以以在线训练的方式，在接收机中对AI编码器和AI解码器进行联合训练，并在联合训练结束后，将接收机中训练后的AI编码器的模型参数反馈给发射机。If you want to use joint training to improve the generalization performance of the AI encoder in the transmitter and the AI decoder in the receiver as a whole, you can jointly train the AI encoder and AI decoder in the receiver in the form of online training , and after the joint training ends, the model parameters of the trained AI encoder in the receiver are fed back to the transmitter.
在一些实现方式中，通过损失函数将接收机的自编码器的训练目标(或优化目标)表示为：min_ {g’}l`(q,f’(g’(q)))，其中，l`(·)表示损失函数；q表示接收机接收的接收信号；g’(·)表示接收机中AI解码器；f’(·)表示接收机中AI编码器。 In some implementations, the training objective (or optimization objective) of the receiver's autoencoder is represented by a loss function as: min_ {g'} l`(q,f'(g'(q))), where, l`(·) represents the loss function; q represents the received signal received by the receiver; g'(·) represents the AI decoder in the receiver; f'(·) represents the AI encoder in the receiver.
下文结合图15和图16以AI编码器、AI解码器形成的自编码器，在CSI反馈过程中和数据传输过程中的应用为例，介绍以联合训练的方式对自编码器进行在线训练过程。需要说明的是，下文介绍的在线训练过程可以与上文中任意一种在线训练配置方式结合。为了简洁，下文不再赘述。In the following, combined with Figure 15 and Figure 16, the application of the self-encoder formed by AI encoder and AI decoder in the process of CSI feedback and data transmission is taken as an example to introduce the online training process of the self-encoder in the way of joint training . It should be noted that the online training process described below can be combined with any of the online training configuration methods above. For the sake of brevity, no further details are given below.
在CSI反馈过程中，上述接收信号包含发射机发射的CSI，该CSI为压缩后的CSI，上述步骤S920包括接收机将CSI输入AI解码器进行解码，得到解码信号，解码信号包含恢复后的信道信息；上述步骤S930包括接收机将解码后的CSI输入AI编码器进行编码，以恢复CSI，并生成恢复信号，恢复信号包含恢复后的CSI。In the CSI feedback process, the above-mentioned received signal includes the CSI transmitted by the transmitter, and the CSI is compressed CSI. The above step S920 includes the receiver inputting the CSI into the AI decoder for decoding to obtain a decoded signal. The decoded signal includes the recovered channel Information; the above step S930 includes that the receiver inputs the decoded CSI into an AI encoder for encoding to restore the CSI and generate a restored signal, the restored signal includes the restored CSI.
图15是本申请实施例的CSI反馈过程中信号处理的方法示意图。如图15所示，发射 机1510将待传输的CSI输入AI编码器1511编码，得到编码后的CSI(又称“压缩后的CSI”)，发射机1510将压缩后的CSI发射至接收机1520。相应地，接收机1520将包含压缩后的CSI的接收信号输入AI解码器1521解码，得到解码后的CSI。接收机1520再将解码后的CSI输入AI编码器1522编码，得到恢复后的CSI。接收机1520基于恢复后的CSI与压缩后的CSI之间的差异，对接收机1520中的AI编码器1522和AI解码器1521进行联合训练。FIG. 15 is a schematic diagram of a signal processing method in a CSI feedback process according to an embodiment of the present application. As shown in Figure 15, the transmitter 1510 inputs the CSI to be transmitted into the AI encoder 1511 for encoding to obtain the encoded CSI (also called "compressed CSI"), and the transmitter 1510 transmits the compressed CSI to the receiver 1520 . Correspondingly, the receiver 1520 inputs the received signal including the compressed CSI to the AI decoder 1521 for decoding, and obtains the decoded CSI. The receiver 1520 then inputs the decoded CSI to the AI encoder 1522 for encoding to obtain the restored CSI. The receiver 1520 jointly trains the AI encoder 1522 and the AI decoder 1521 in the receiver 1520 based on the difference between the restored CSI and the compressed CSI.
在数据传输过程中，上述接收信号包括发射机发射的第一数据信号，上述步骤S920包括接收机将第一数据信号输入AI解码器进行解码，以恢复发射机发射的数据，得到解码信号，解码信号包含恢复后的数据；上述接收机将解码信号输入AI编码器，生成恢复信号，包括：接收机将恢复后的数据输入AI编码器，得到第二数据信号；接收机根据估计出的第二信道信息以及第二随机噪声，对第二数据信号进行处理，得到恢复信号。In the data transmission process, the above-mentioned received signal includes the first data signal transmitted by the transmitter, and the above-mentioned step S920 includes the receiver inputting the first data signal into the AI decoder for decoding, so as to recover the data transmitted by the transmitter, obtain the decoded signal, and decode The signal contains recovered data; the above-mentioned receiver inputs the decoded signal into the AI encoder to generate the recovered signal, including: the receiver inputs the recovered data into the AI encoder to obtain a second data signal; the receiver uses the estimated second The channel information and the second random noise are used to process the second data signal to obtain a restored signal.
上述第二随机噪声可以是接收机预存的，也可以是接收机预先生成的，本申请实施例对此不作限定。The foregoing second random noise may be pre-stored by the receiver, or may be pre-generated by the receiver, which is not limited in this embodiment of the present application.
在一些实现方式中，可以在上述第一数据信号中***导频信号，这样接收机可以基于导频信号进行信道估计，以提高恢复信号与接收信号的相似程度。当然，也可以不进行信道估计，直接将第二数据信号输入AI模型得到恢复信号，本申请实施例对此不作限定。In some implementation manners, a pilot signal may be inserted into the first data signal, so that the receiver may perform channel estimation based on the pilot signal, so as to improve the similarity between the recovered signal and the received signal. Of course, channel estimation may not be performed, and the second data signal may be directly input into the AI model to obtain the restored signal, which is not limited in this embodiment of the present application.
图16是本申请实施例的数据传输过程中信号处理的方法示意图。如图16所示，发射机1610将待传输的数据输入AI编码器1611编码，得到第一数据信号。发射机1620通过无线信道H将第一数据信号发射至接收机1620，并且在发射的过程中，会在第一数据信号上叠加噪声N，形成接收信号Y。相应地，接收机1620在接收到接收信号Y后，将接收信号Y输入AI解码器1621解码，得到恢复后的数据。接收机1620再将恢复后的数据输入AI编码器1622编码，得到第二数据信号。接收机1620根据随机噪声N’以及估计得到的信道信息H’，对第二数据信号进行处理得到恢复信号Y’。接收机1620基于恢复信号Y’和接收信号Y之间的差异，对接收机1620中的AI编码器1622和AI解码器1621进行联合训练。FIG. 16 is a schematic diagram of a signal processing method during data transmission according to an embodiment of the present application. As shown in FIG. 16 , the transmitter 1610 inputs the data to be transmitted into the AI encoder 1611 for encoding to obtain a first data signal. The transmitter 1620 transmits the first data signal to the receiver 1620 through the wireless channel H, and during the transmission process, noise N will be superimposed on the first data signal to form a received signal Y. Correspondingly, after receiving the received signal Y, the receiver 1620 inputs the received signal Y to the AI decoder 1621 for decoding to obtain restored data. The receiver 1620 then inputs the restored data into the AI encoder 1622 for encoding to obtain a second data signal. The receiver 1620 processes the second data signal to obtain a restored signal Y' according to the random noise N' and the estimated channel information H'. The receiver 1620 jointly trains the AI encoder 1622 and the AI decoder 1621 in the receiver 1620 based on the difference between the recovered signal Y' and the received signal Y.
当然，在数据传输过程中，也可以仅使用AI解码器，而不使用AI编码器，此时，发射机执行的信号处理流程可以与传统的信号处理流程相似，例如，对待发射的信号进行编码、调制等。相应地，接收机对AI解码器输出的解码信号也可以执行与传统的信号处理流程相似的信号处理方式，例如，对待发射的信号进行编码、调制等。在这种情况下，可以理解为仅对AI解码器在线训练，具体可以参见上文关于AI解码器的在线训练过程，为了简洁，在此不再赘述。Of course, in the process of data transmission, only the AI decoder can be used instead of the AI encoder. At this time, the signal processing flow performed by the transmitter can be similar to the traditional signal processing flow, for example, encoding the signal to be transmitted , Modulation, etc. Correspondingly, the receiver may also perform a signal processing method similar to a traditional signal processing flow on the decoded signal output by the AI decoder, for example, encoding and modulating the signal to be transmitted. In this case, it can be understood that only the AI decoder is trained online. For details, please refer to the online training process of the AI decoder above. For the sake of brevity, details are not repeated here.
上文结合图1至图16，详细描述了本申请的方法实施例，下面结合图17至图18，详细描述本申请的装置实施例。应理解，方法实施例的描述与装置实施例的描述相互对应，因此，未详细描述的部分可以参见前面方法实施例。The method embodiment of the present application is described in detail above with reference to FIG. 1 to FIG. 16 , and the device embodiment of the present application is described in detail below in conjunction with FIG. 17 to FIG. 18 . It should be understood that the descriptions of the method embodiments correspond to the descriptions of the device embodiments, therefore, for parts not described in detail, reference may be made to the foregoing method embodiments.
图17是本申请实施例的接收机的示意性结构图。图17所示的接收机1700包括：接收单元1710和处理单元1720。Fig. 17 is a schematic structural diagram of a receiver according to an embodiment of the present application. The receiver 1700 shown in FIG. 17 includes: a receiving unit 1710 and a processing unit 1720 .
接收单元1710可用于对发射机发射的无线信号进行接收，得到接收信号。The receiving unit 1710 may be configured to receive the wireless signal transmitted by the transmitter to obtain a received signal.
处理单元1720可用于将所述接收信号输入AI解码器对所述接收信号进行解码，得到解码信号；根据所述解码信号，生成所述接收信号的恢复信号；以及根据所述恢复信号与所述接收信号之间的差异，对所述AI解码器进行在线训练。The processing unit 1720 may be configured to input the received signal into an AI decoder to decode the received signal to obtain a decoded signal; generate a recovery signal of the received signal according to the decoded signal; and generate a recovery signal of the received signal according to the recovery signal and the The difference between the received signals is used for online training of the AI decoder.
可选地，所述接收单元1710还可用于接收网络设备发送的在线训练的指示信息，所述在线训练的指示信息用于指示所述AI解码器进行所述在线训练的时间。Optionally, the receiving unit 1710 is further configured to receive online training instruction information sent by a network device, where the online training instruction information is used to indicate the time for the AI decoder to perform the online training.
可选地，所述在线训练的指示信息包括第一指示信息，所述第一指示信息用于指示AI解码器进行在线训练的在线训练周期，所述在线训练周期包括多个传输周期，所述在线训练周期包括第一类型的传输周期和第二类型的传输周期，所述第一类型的传输周期用于对所述AI解码器进行在线训练，所述第二类型的传输周期用于收集在线训练的训练数据。Optionally, the instruction information of the online training includes first instruction information, the first instruction information is used to instruct the AI decoder to perform an online training period for online training, the online training period includes a plurality of transmission periods, the The online training period includes a first type of transmission period and a second type of transmission period, the first type of transmission period is used for online training of the AI decoder, and the second type of transmission period is used to collect online Training data for training.
可选地，所述在线训练的指示信息还包括第二指示信息，所述第二指示信息用于指示所述第一类型的传输周期所使用的训练数据的大小。Optionally, the indication information of the online training further includes second indication information, where the second indication information is used to indicate the size of the training data used in the first type of transmission period.
可选地，所述在线训练的时间包括所述在线训练的起始时间和/或结束时间。Optionally, the time of the online training includes the start time and/or end time of the online training.
可选地，所述接收信号为所述发射机发射的导频信号的接收信号，所述处理单元1720还可用于将所述接收信号输入所述AI解码器进行解码，进行信道估计，得到所述解码信号，所述解码信号包含估计出的第一信道信息；以及根据所述第一信道信息和第一随机噪声对所述接收机存储的所述导频信号进行处理，得到所述恢复信号，其中，所述发射机发射的导频信号与所述接收机存储的导频信号相同。Optionally, the received signal is a received signal of a pilot signal transmitted by the transmitter, and the processing unit 1720 may also be configured to input the received signal into the AI decoder for decoding, perform channel estimation, and obtain the The decoded signal, the decoded signal includes estimated first channel information; and according to the first channel information and first random noise, the pilot signal stored in the receiver is processed to obtain the restored signal , wherein the pilot signal transmitted by the transmitter is the same as the pilot signal stored by the receiver.
可选地，所述处理单元还可用于将所述接收机存储的导频信号、所述接收信号输入所述AI解码器，进行信道估计，得到所述解码信号。Optionally, the processing unit is further configured to input the pilot signal stored in the receiver and the received signal into the AI decoder for channel estimation to obtain the decoded signal.
可选地，所述处理单元1720可具体用于将所述解码信号输入AI编码器，生成所述恢复信号。Optionally, the processing unit 1720 may be specifically configured to input the decoded signal into an AI encoder to generate the restored signal.
可选地，所述接收信号包含所述发射机发射的压缩后的CSI，所述CSI为压缩后的CSI，所述处理单元1720还可用于将所述CSI输入所述AI解码器进行解码，得到所述解码信号，所述解码信号包含恢复后的信道信息；以及将所述恢复后的信道信息输入所述AI编码器进行编码，恢复所述CSI，并生成所述恢复信号，所述恢复信号包含恢复后的所述CSI。Optionally, the received signal includes compressed CSI transmitted by the transmitter, where the CSI is compressed CSI, and the processing unit 1720 is further configured to input the CSI into the AI decoder for decoding, Obtain the decoded signal, the decoded signal includes restored channel information; and input the restored channel information into the AI encoder for encoding, restore the CSI, and generate the restored signal, the restored The signal contains the recovered CSI.
可选地，所述接收信号包括所述发射机发射的第一数据信号，所述处理单元1720还可用于将所述第一数据信号输入所述AI解码器进行解码，恢复所述发射机发射的数据，得到所述解码信号，所述解码信号包含恢复后的所述数据；将所述恢复后的数据输入所述AI编码器，得到第二数据信号；以及根据估计出的第二信道信息和第二随机噪声对所述第二数据信号进行处理，得到所述恢复信号。Optionally, the received signal includes a first data signal transmitted by the transmitter, and the processing unit 1720 is further configured to input the first data signal into the AI decoder for decoding, and recover the signal transmitted by the transmitter. data to obtain the decoded signal, the decoded signal includes the restored data; input the restored data into the AI encoder to obtain a second data signal; and according to the estimated second channel information and second random noise to process the second data signal to obtain the restored signal.
可选地，所述处理单元1720还可用于将所述差异最小化作为优化目标，更新所述AI解码器的权重，对所述AI解码器进行所述在线训练。Optionally, the processing unit 1720 is further configured to take the minimization of the difference as an optimization goal, update the weight of the AI decoder, and perform the online training on the AI decoder.
可选地，所述恢复信号与所述接收信号之间的所述差异是通过所述恢复信号与所述接收信号之间的误差表示的。Optionally, the difference between the restored signal and the received signal is represented by an error between the restored signal and the received signal.
图18是本申请实施例的用于信号处理的装置的示意性结构图。图18中的虚线表示该单元或模块为可选的。该装置1800可用于实现上述方法实施例中描述的方法。装置1800可以是芯片、终端设备或网络设备。Fig. 18 is a schematic structural diagram of an apparatus for signal processing according to an embodiment of the present application. The dashed line in Figure 18 indicates that the unit or module is optional. The apparatus 1800 may be used to implement the methods described in the foregoing method embodiments. Apparatus 1800 may be a chip, a terminal device or a network device.
装置1800可以包括一个或多个处理器1810。该处理器1810可支持装置1800实现前文方法实施例所描述的方法。该处理器1810可以是通用处理器或者专用处理器。例如，该处理器可以为中央处理单元(central processing unit，CPU)。或者，该处理器还可以是其他通用处理器、数字信号处理器(digital signal processor，DSP)、专用集成电路(application specific integrated circuit，ASIC)、现成可编程门阵列(field programmable gate array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。Apparatus 1800 may include one or more processors 1810 . The processor 1810 may support the device 1800 to implement the methods described in the foregoing method embodiments. The processor 1810 may be a general purpose processor or a special purpose processor. For example, the processor may be a central processing unit (central processing unit, CPU). Alternatively, the processor can also be other general-purpose processors, digital signal processors (digital signal processors, DSPs), application specific integrated circuits (application specific integrated circuits, ASICs), off-the-shelf programmable gate arrays (field programmable gate arrays, FPGAs) Or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.
装置1800还可以包括一个或多个存储器1820。存储器1820上存储有程序，该程序可以被处理器1810执行，使得处理器1810执行前文方法实施例所描述的方法。存储器1820可以独立于处理器1810也可以集成在处理器1810中。Apparatus 1800 may also include one or more memories 1820 . A program is stored in the memory 1820, and the program can be executed by the processor 1810, so that the processor 1810 executes the methods described in the foregoing method embodiments. The memory 1820 may be independent from the processor 1810 or may be integrated in the processor 1810 .
装置1800还可以包括收发器1830。处理器1810可以通过收发器1830与其他设备或芯片进行通信。例如，处理器1810可以通过收发器1830与其他设备或芯片进行数据收发。The apparatus 1800 may also include a transceiver 1830 . The processor 1810 can communicate with other devices or chips through the transceiver 1830 . For example, the processor 1810 may send and receive data with other devices or chips through the transceiver 1830 .
本申请实施例还提供一种计算机可读存储介质，用于存储程序。该计算机可读存储介质可应用于本申请实施例提供的终端或网络设备中，并且该程序使得计算机执行本申请各个实施例中的由终端或网络设备执行的方法。The embodiment of the present application also provides a computer-readable storage medium for storing programs. The computer-readable storage medium can be applied to the terminal or the network device provided in the embodiments of the present application, and the program enables the computer to execute the methods performed by the terminal or the network device in the various embodiments of the present application.
本申请实施例还提供一种计算机程序产品。该计算机程序产品包括程序。该计算机程序产品可应用于本申请实施例提供的终端或网络设备中，并且该程序使得计算机执行本申 请各个实施例中的由终端或网络设备执行的方法。The embodiment of the present application also provides a computer program product. The computer program product includes programs. The computer program product can be applied to the terminal or the network device provided in the embodiments of the present application, and the program enables the computer to execute the methods performed by the terminal or the network device in the various embodiments of the present application.
本申请实施例还提供一种计算机程序。该计算机程序可应用于本申请实施例提供的终端或网络设备中，并且该计算机程序使得计算机执行本申请各个实施例中的由终端或网络设备执行的方法。The embodiment of the present application also provides a computer program. The computer program can be applied to the terminal or the network device provided in the embodiments of the present application, and the computer program enables the computer to execute the methods performed by the terminal or the network device in the various embodiments of the present application.
应理解，本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中字符“/”，一般表示前后关联对象是一种“或”的关系。It should be understood that the term "and/or" in this article is only an association relationship describing associated objects, which means that there may be three relationships, for example, A and/or B may mean: A exists alone, and A and B exist at the same time , there are three cases of B alone. In addition, the character "/" in this article generally indicates that the contextual objects are an "or" relationship.
应理解，在本申请的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。It should be understood that, in various embodiments of the present application, the sequence numbers of the above-mentioned processes do not mean the order of execution, and the execution order of the processes should be determined by their functions and internal logic, and should not be used in the embodiments of the present application. The implementation process constitutes any limitation.
在本申请所提供的几个实施例中，应该理解到，所揭露的***、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个***，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods may be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or May be integrated into another system, or some features may be ignored, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit.
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(digital subscriber line，DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够读取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，数字通用光盘(digital video disc，DVD))或者半导体介质(例如，固态硬盘(solid state disk，SSD))等。In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof. 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 application 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 (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be read 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, a magnetic tape), an optical medium (for example, a digital versatile disc (digital video disc, DVD)) or a semiconductor medium (for example, a solid state disk (solid state disk, SSD) )Wait.
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。The above is only a specific implementation of the application, but the scope of protection of the application is not limited thereto. Anyone familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the application. Should be covered within the protection scope of this application. Therefore, the protection scope of the present application should be determined by the protection scope of the claims.

Claims (30)

  1. 一种信号处理的方法，其特征在于，包括：A signal processing method, characterized in that, comprising:
    接收机对发射机发射的无线信号进行接收，得到接收信号；The receiver receives the wireless signal transmitted by the transmitter to obtain the received signal;
    所述接收机将所述接收信号输入人工智能AI解码器进行解码，得到解码信号；The receiver inputs the received signal into an artificial intelligence AI decoder for decoding to obtain a decoded signal;
    所述接收机根据所述解码信号，生成所述接收信号的恢复信号；The receiver generates a recovery signal of the received signal according to the decoded signal;
    所述接收机根据所述恢复信号与所述接收信号之间的差异，对所述AI解码器进行在线训练。The receiver performs online training for the AI decoder according to the difference between the restored signal and the received signal.
  2. 如权利要求1所述的方法，其特征在于，所述方法还包括：The method of claim 1, further comprising:
    所述接收机接收网络设备发送的在线训练的指示信息，所述在线训练的指示信息用于指示所述AI解码器进行所述在线训练的时间。The receiver receives the instruction information of the online training sent by the network device, and the instruction information of the online training is used to indicate the time when the AI decoder performs the online training.
  3. 如权利要求2所述的方法，其特征在于，所述在线训练的指示信息包括第一指示信息，所述第一指示信息用于指示所述AI解码器进行所述在线训练的在线训练周期，所述在线训练周期包括多个传输周期，所述在线训练周期包括第一类型的传输周期和第二类型的传输周期，所述第一类型的传输周期用于对所述AI解码器进行在线训练，所述第二类型的传输周期用于收集在线训练的训练数据。The method according to claim 2, wherein the instruction information of the online training includes first instruction information, and the first instruction information is used to instruct the AI decoder to perform the online training period of the online training, The online training period includes a plurality of transmission periods, the online training period includes a first type of transmission period and a second type of transmission period, and the first type of transmission period is used for online training of the AI decoder , the second type of transmission period is used to collect training data for online training.
  4. 如权利要求3所述的方法，其特征在于，所述在线训练的指示信息还包括第二指示信息，所述第二指示信息用于指示所述第一类型的传输周期所使用的训练数据的大小。The method according to claim 3, wherein the indication information of the online training further includes second indication information, and the second indication information is used to indicate the training data used in the first type of transmission cycle. size.
  5. 如权利要求2-4中任一项所述的方法，其特征在于，所述在线训练的时间包括所述在线训练的起始时间和/或结束时间。The method according to any one of claims 2-4, wherein the time of the online training includes a start time and/or an end time of the online training.
  6. 如权利要求1-5中任一项所述的方法，其特征在于，所述接收信号为所述发射机发射的导频信号的接收信号，The method according to any one of claims 1-5, wherein the received signal is a received signal of a pilot signal transmitted by the transmitter,
    所述接收机将所述接收信号输入AI解码器进行解码，得到解码信号，包括：The receiver inputs the received signal into an AI decoder for decoding to obtain a decoded signal, including:
    所述接收机将所述接收信号输入所述AI解码器进行解码，进行信道估计，得到所述解码信号，所述解码信号包含估计出的第一信道信息；The receiver inputs the received signal into the AI decoder for decoding, performs channel estimation, and obtains the decoded signal, and the decoded signal includes the estimated first channel information;
    所述接收机根据所述解码信号，生成所述接收信号的恢复信号，包括：The receiver generates a recovery signal of the received signal according to the decoded signal, including:
    所述接收机根据所述第一信道信息和第一随机噪声，对所述接收机存储的导频信号进行处理，得到所述恢复信号，其中，所述发射机发射的导频信号与所述接收机存储的导频信号相同。The receiver processes the pilot signal stored in the receiver according to the first channel information and the first random noise to obtain the restored signal, wherein the pilot signal transmitted by the transmitter is the same as the The receiver stores the same pilot signal.
  7. 如权利要求6所述的方法，其特征在于，所述接收机将所述接收信号输入所述AI解码器进行解码，以进行信道估计，得到所述解码信号，包括：The method according to claim 6, wherein the receiver inputs the received signal into the AI decoder for decoding to perform channel estimation to obtain the decoded signal, comprising:
    所述接收机将所述接收机存储的导频信号、所述接收信号输入所述AI解码器，进行信道估计，得到所述解码信号。The receiver inputs the pilot signal stored in the receiver and the received signal into the AI decoder, performs channel estimation, and obtains the decoded signal.
  8. 如权利要求1-5中任一项所述的方法，其特征在于，所述接收机根据所述解码信号，生成所述接收信号的恢复信号，包括：The method according to any one of claims 1-5, wherein the receiver generates a restoration signal of the received signal according to the decoded signal, comprising:
    所述接收机将所述解码信号输入AI编码器，生成所述恢复信号。The receiver inputs the decoded signal into an AI encoder to generate the restored signal.
  9. 如权利要求8所述的方法，其特征在于，所述接收信号包含所述发射机发射的信道状态信息CSI，所述CSI为压缩后的CSI，The method according to claim 8, wherein the received signal includes channel state information (CSI) transmitted by the transmitter, and the CSI is compressed CSI,
    所述接收机将所述接收信号输入AI解码器进行解码，得到解码信号，包括：The receiver inputs the received signal into an AI decoder for decoding to obtain a decoded signal, including:
    所述接收机将所述CSI输入所述AI解码器进行解码，得到所述解码信号，所述解码信号包含恢复后的信道信息；The receiver inputs the CSI into the AI decoder for decoding to obtain the decoded signal, and the decoded signal includes restored channel information;
    所述接收机将所述解码信号输入AI编码器，生成所述恢复信号，包括：The receiver inputs the decoded signal into an AI encoder to generate the restored signal, including:
    所述接收机将所述解码后的CSI输入所述AI编码器进行编码，恢复所述CSI，并生成所述恢复信号，所述恢复信号包含恢复后的所述CSI。The receiver inputs the decoded CSI to the AI encoder for encoding, restores the CSI, and generates the restored signal, where the restored signal includes the restored CSI.
  10. 如权利要求8所述的方法，其特征在于，所述接收信号包括所述发射机发射的第一数据信号，The method of claim 8, wherein said received signal comprises a first data signal transmitted by said transmitter,
    所述接收机将所述接收信号输入AI解码器进行解码，得到解码信号，包括：The receiver inputs the received signal into an AI decoder for decoding to obtain a decoded signal, including:
    所述接收机将所述第一数据信号输入所述AI解码器进行解码，恢复所述发射机发射的数据，得到所述解码信号，所述解码信号包含恢复后的所述数据；The receiver inputs the first data signal into the AI decoder for decoding, recovers the data transmitted by the transmitter, and obtains the decoded signal, and the decoded signal includes the recovered data;
    所述接收机将所述解码信号输入AI编码器，生成所述恢复信号，包括：The receiver inputs the decoded signal into an AI encoder to generate the restored signal, including:
    所述接收机将所述恢复后的数据输入所述AI编码器，得到第二数据信号；The receiver inputs the recovered data into the AI encoder to obtain a second data signal;
    所述接收机根据估计出的第二信道信息和第二随机噪声，对所述第二数据信号进行处理，得到所述恢复信号。The receiver processes the second data signal according to the estimated second channel information and second random noise to obtain the restored signal.
  11. 如权利要求1-10中任一项所述的方法，其特征在于，所述接收机根据所述恢复信号与所述接收信号之间的差异，对所述AI解码器进行在线训练，包括：The method according to any one of claims 1-10, wherein the receiver performs online training on the AI decoder according to the difference between the restored signal and the received signal, including:
    所述接收机将所述差异最小化作为优化目标，更新所述AI解码器的权重，对所述AI解码器进行所述在线训练。The receiver takes the minimization of the difference as an optimization goal, updates the weights of the AI decoder, and performs the online training on the AI decoder.
  12. 如权利要求1-11中任一项所述的方法，其特征在于，所述恢复信号与所述接收信号之间的所述差异是通过所述恢复信号与所述接收信号之间的误差表示的。The method according to any one of claims 1-11, wherein the difference between the recovered signal and the received signal is represented by an error between the recovered signal and the received signal of.
  13. 一种接收机，其特征在于，包括：A receiver, characterized in that it comprises:
    接收单元，用于对发射机发射的无线信号进行接收，得到接收信号；The receiving unit is used to receive the wireless signal transmitted by the transmitter to obtain the received signal;
    处理单元，用于将所述接收信号输入人工智能AI解码器进行解码，得到解码信号；A processing unit, configured to input the received signal into an artificial intelligence AI decoder for decoding to obtain a decoded signal;
    所述处理单元，用于根据所述解码信号，生成所述接收信号的恢复信号；The processing unit is configured to generate a recovery signal of the received signal according to the decoded signal;
    所述处理单元，用于根据所述恢复信号与所述接收信号之间的差异，对所述AI解码器进行在线训练。The processing unit is configured to perform online training on the AI decoder according to the difference between the restored signal and the received signal.
  14. 如权利要求13所述的接收机，其特征在于，所述接收单元，还用于：The receiver according to claim 13, wherein the receiving unit is further used for:
    接收网络设备发送的在线训练的指示信息，所述在线训练的指示信息用于指示所述AI解码器进行所述在线训练的时间。Receive online training instruction information sent by the network device, where the online training instruction information is used to indicate the time when the AI decoder performs the online training.
  15. 如权利要求14所述的接收机，其特征在于，所述在线训练的指示信息包括第一指示信息，所述第一指示信息用于指示所述AI解码器进行所述在线训练的在线训练周期，所述在线训练周期包括多个传输周期，所述在线训练周期包括第一类型的传输周期和第二类型的传输周期，所述第一类型的传输周期用于对所述AI解码器进行在线训练，所述第二类型的传输周期用于收集在线训练的训练数据。The receiver according to claim 14, wherein the instruction information of the online training includes first instruction information, and the first instruction information is used to instruct the AI decoder to perform the online training period of the online training , the online training period includes a plurality of transmission periods, the online training period includes a first type of transmission period and a second type of transmission period, and the first type of transmission period is used to perform online training on the AI decoder training, the second type of transmission period is used to collect training data for online training.
  16. 如权利要求15所述的接收机，其特征在于，所述在线训练的指示信息还包括第二指示信息，所述第二指示信息用于指示所述第一类型的传输周期所使用的训练数据的大小。The receiver according to claim 15, wherein the indication information of the online training further includes second indication information, and the second indication information is used to indicate the training data used in the transmission period of the first type the size of.
  17. 如权利要求14-16中任一项所述的接收机，其特征在于，所述在线训练的时间包括所述在线训练的起始时间和/或结束时间。The receiver according to any one of claims 14-16, wherein the time of the online training includes a start time and/or an end time of the online training.
  18. 如权利要求13-17中任一项所述的接收机，其特征在于，所述接收信号为所述发射机发射的导频信号的接收信号，所述处理单元，还用于：The receiver according to any one of claims 13-17, wherein the received signal is a received signal of a pilot signal transmitted by the transmitter, and the processing unit is further configured to:
    将所述接收信号输入所述AI解码器进行解码，进行信道估计，得到所述解码信号，所述解码信号包含估计出的信道信息；以及inputting the received signal into the AI decoder for decoding, and performing channel estimation to obtain the decoded signal, the decoded signal including estimated channel information; and
    根据所述第一信道信息和第一随机噪声对所述接收机存储的导频信号进行处理，得到所述恢复信号，其中，所述发射机发射的导频信号与所述接收机存储的导频信号相同。Process the pilot signal stored by the receiver according to the first channel information and the first random noise to obtain the restored signal, wherein the pilot signal transmitted by the transmitter is the same as the pilot signal stored by the receiver The frequency signal is the same.
  19. 如权利要求18所述的接收机，其特征在于，所述处理单元，还用于：The receiver according to claim 18, wherein the processing unit is further configured to:
    将所述接收机存储的导频信号、所述接收信号输入所述AI解码器，进行信道估计，得到所述解码信号。Inputting the pilot signal stored in the receiver and the received signal into the AI decoder for channel estimation to obtain the decoded signal.
  20. 如权利要求13-17中任一项所述的接收机，其特征在于，A receiver as claimed in any one of claims 13-17, characterized in that,
    所述处理单元，还用于将所述解码信号输入AI编码器，生成所述恢复信号。The processing unit is further configured to input the decoded signal into an AI encoder to generate the restored signal.
  21. 如权利要求20所述的接收机，其特征在于，所述接收信号包含所述发射机发射的信道状态信息CSI，所述CSI为压缩后的CSI，所述处理单元，还用于：The receiver according to claim 20, wherein the received signal includes channel state information CSI transmitted by the transmitter, the CSI is compressed CSI, and the processing unit is further configured to:
    将所述CSI输入所述AI解码器进行解码，得到所述解码信号，所述解码信号包含恢 复后的信道信息；以及Inputting the CSI into the AI decoder for decoding to obtain the decoded signal, the decoded signal including restored channel information; and
    将所述恢复后的信道信息输入所述AI编码器进行编码，恢复所述CSI，并生成所述恢复信号，所述恢复信号包含恢复后的所述CSI。Inputting the restored channel information into the AI encoder for encoding, restoring the CSI, and generating the restored signal, where the restored signal includes the restored CSI.
  22. 如权利要求20所述的接收机，其特征在于，所述接收信号包括所述发射机发射的第一数据信号，所述处理单元，还用于：The receiver according to claim 20, wherein the received signal comprises the first data signal transmitted by the transmitter, and the processing unit is further configured to:
    将所述第一数据信号输入所述AI解码器进行解码，恢复所述发射机发射的数据，得到所述解码信号，所述解码信号包含恢复后的所述数据；inputting the first data signal into the AI decoder for decoding, recovering the data transmitted by the transmitter, and obtaining the decoded signal, the decoded signal including the recovered data;
    将所述恢复后的数据输入所述AI编码器，得到第二数据信号；以及inputting the recovered data into the AI encoder to obtain a second data signal; and
    根据估计出的第二信道信息和第二随机噪声对所述第二数据信号进行处理，得到所述恢复信号。Process the second data signal according to the estimated second channel information and second random noise to obtain the restored signal.
  23. 如权利要求13-22中任一项所述的接收机，其特征在于，所述处理单元，还用于：The receiver according to any one of claims 13-22, wherein the processing unit is further configured to:
    将所述差异最小化作为优化目标，更新所述AI解码器的权重，对所述AI解码器进行所述在线训练。Taking the minimization of the difference as an optimization goal, updating the weights of the AI decoder, and performing the online training on the AI decoder.
  24. 如权利要求13-23中任一项所述的接收机，其特征在于，所述恢复信号与所述接收信号之间的所述差异是通过所述恢复信号与所述接收信号之间的误差表示的。A receiver according to any one of claims 13-23, wherein said difference between said recovered signal and said received signal is via an error between said recovered signal and said received signal Expressed.
  25. 一种接收机，其特征在于，包括存储器和处理器，所述存储器用于存储程序，所述处理器用于调用所述存储器中的程序，以执行如权利要求1-12中任一项所述的方法。A receiver, characterized by comprising a memory and a processor, the memory is used to store a program, and the processor is used to call the program in the memory to execute the program described in any one of claims 1-12 Methods.
  26. 一种装置，其特征在于，包括处理器，用于从存储器中调用程序，以执行如权利要求1-12中任一项所述的方法。An apparatus, characterized by comprising a processor, configured to call a program from a memory to execute the method according to any one of claims 1-12.
  27. 一种芯片，其特征在于，包括处理器，用于从存储器调用程序，使得安装有所述芯片的设备执行如权利要求1-12中任一项所述的方法。A chip, characterized by comprising a processor, configured to call a program from a memory, so that a device installed with the chip executes the method according to any one of claims 1-12.
  28. 一种计算机可读存储介质，其特征在于，其上存储有程序，所述程序使得计算机执行如权利要求1-12中任一项所述的方法。A computer-readable storage medium, characterized in that a program is stored thereon, and the program causes a computer to execute the method according to any one of claims 1-12.
  29. 一种计算机程序产品，其特征在于，包括程序，所述程序使得计算机执行如权利要求1-12中任一项所述的方法。A computer program product, characterized by comprising a program, the program causes a computer to execute the method according to any one of claims 1-12.
  30. 一种计算机程序，其特征在于，所述计算机程序使得计算机执行如权利要求1-12中任一项所述的方法。A computer program, characterized in that the computer program causes a computer to execute the method according to any one of claims 1-12.
PCT/CN2021/105842 2021-07-12 2021-07-12 Method for processing signal, and receiver WO2023283785A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202180095411.1A CN116982300A (en) 2021-07-12 2021-07-12 Signal processing method and receiver
PCT/CN2021/105842 WO2023283785A1 (en) 2021-07-12 2021-07-12 Method for processing signal, and receiver

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/105842 WO2023283785A1 (en) 2021-07-12 2021-07-12 Method for processing signal, and receiver

Publications (1)

Publication Number Publication Date
WO2023283785A1 true WO2023283785A1 (en) 2023-01-19

Family

ID=84919805

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/105842 WO2023283785A1 (en) 2021-07-12 2021-07-12 Method for processing signal, and receiver

Country Status (2)

Country Link
CN (1) CN116982300A (en)
WO (1) WO2023283785A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108566257A (en) * 2018-04-27 2018-09-21 电子科技大学 A kind of signal recovery method based on reverse transmittance nerve network
CN109743268A (en) * 2018-12-06 2019-05-10 东南大学 Millimeter wave channel estimation and compression method based on deep neural network
WO2020035684A1 (en) * 2018-08-15 2020-02-20 Imperial College Of Science, Technology And Medicine Joint source channel coding of information sources using neural networks
CN112183736A (en) * 2019-07-05 2021-01-05 三星电子株式会社 Artificial intelligence processor and method for executing neural network operation

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108566257A (en) * 2018-04-27 2018-09-21 电子科技大学 A kind of signal recovery method based on reverse transmittance nerve network
WO2020035684A1 (en) * 2018-08-15 2020-02-20 Imperial College Of Science, Technology And Medicine Joint source channel coding of information sources using neural networks
CN109743268A (en) * 2018-12-06 2019-05-10 东南大学 Millimeter wave channel estimation and compression method based on deep neural network
CN112183736A (en) * 2019-07-05 2021-01-05 三星电子株式会社 Artificial intelligence processor and method for executing neural network operation

Also Published As

Publication number Publication date
CN116982300A (en) 2023-10-31

Similar Documents

Publication Publication Date Title
CN112054863B (en) Communication method and device
WO2021217519A1 (en) Method and apparatus for adjusting neural network
EP4228347A1 (en) Communication method and communication apparatus
US11956031B2 (en) Communication of measurement results in coordinated multipoint
US20230342593A1 (en) Neural network training method and related apparatus
WO2023125660A1 (en) Communication method and device
WO2023283785A1 (en) Method for processing signal, and receiver
WO2023116808A1 (en) Communication method and apparatus
WO2022001822A1 (en) Method and device for acquiring neural network
WO2022222116A1 (en) Channel recovery method and receiving end device
WO2023070675A1 (en) Data processing method and apparatus
CN114492784A (en) Neural network testing method and device
WO2023115254A1 (en) Data processing method and device
WO2019200971A1 (en) Polar code decoding method and communication device
WO2023060503A1 (en) Information processing method and apparatus, device, medium, chip, product, and program
WO2022233061A1 (en) Signal processing method, communication device, and communication system
WO2023015499A1 (en) Wireless communication method and device
WO2023016503A1 (en) Communication method and apparatus
WO2023004563A1 (en) Method for obtaining reference signal and communication devices
WO2022217556A1 (en) Data enhancement method, receiver and storage medium
WO2023137641A1 (en) Channel estimation method, channel estimation model training method, and communication device
WO2024037321A1 (en) Ai model training method and apparatus, and device and readable storage medium
WO2023097645A1 (en) Data acquisition method and apparatus, and device, medium, chip, product and program
WO2023185758A1 (en) Data transmission method and communication apparatus
WO2023179577A1 (en) Communication method and related apparatus

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: 21949581

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 202180095411.1

Country of ref document: CN

NENP Non-entry into the national phase

Ref country code: DE