CN114257477B

CN114257477B - Signal detection method and related equipment

Info

Publication number: CN114257477B
Application number: CN202011011511.5A
Authority: CN
Inventors: 曾兴
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2020-09-23
Filing date: 2020-09-23
Publication date: 2023-07-18
Anticipated expiration: 2040-09-23
Also published as: CN114257477A; WO2022062868A1

Abstract

The embodiment of the application provides a signal detection method and related equipment, which are used for optimizing the processing process of bit soft information estimation and improving the communication efficiency. In the method, the statistical characteristics of the signal to be detected can be processed by utilizing the neural network and then the corresponding bit soft information is output, and compared with the comparison operation in the traditional max-log algorithm, the comparison operation is replaced by the more general and efficient multiplication and addition operation in the neural network processing process. In addition, in the method, a method of randomly sampling the received signal according to the filtering parameter of the received signal can also be used for obtaining a sampling result of discrete information for indicating the posterior probability of the signal to be detected, and compared with the implementation process of participating in calculation through multiple QR decomposition, the method of randomly sampling can reduce the complexity of the calculation process. Therefore, the signal processing process of the signal detection device is optimized, and the communication efficiency is improved.

Description

Signal detection method and related equipment

Technical Field

The present disclosure relates to the field of communications, and in particular, to a signal detection method and related device.

Background

MIMO (multiple-input and multiple-output) technology is a hotspot currently studied in the field of wireless communication, and MIMO technology is adopted in various novel mobile communication systems to improve the spectral efficiency of the system. The MIMO technology can increase the spatial dimension of data multiplexing, so that multiple data can be spatially multiplexed to the same time-frequency resource, and the same data can be transmitted by multiple antennas and/or received by multiple receiving antennas, thereby obtaining the spatial diversity gain.

In the transmission process of applying the MIMO technology, the signal transmitting end maps bit information to constellation symbols, modulates the constellation symbols into signals, and transmits the signals through the transmitting antennas. Correspondingly, the signal receiving end receives the signal through the receiving antenna, demodulates the signal into a searching path result of the constellation symbol, namely, discrete information of posterior probability of the constellation symbol is obtained through demodulation, and determines bit information in the searching path result. In general, in the signal receiving end, the process of demodulating the signal into the search path result of the constellation symbol is generally obtained through the calculation process of multiple QR operations, and in addition, the process of mapping the constellation symbol into the bit information is generally obtained through the approximate calculation of a max-log algorithm.

However, in the process of realizing bit soft information estimation at the signal receiving end, the computation process of multiple QR operations and the computation of the max-log algorithm are too complex, and irregular searching and comparing operations exist in the computation process, which is not beneficial to the efficient realization of a general processor and affects the communication efficiency.

Disclosure of Invention

The embodiment of the application provides a signal detection method and related equipment, which are used for optimizing the processing process of bit soft information estimation and improving the communication efficiency.

The first aspect of the embodiments of the present application provides a signal detection method, where the method is applied to a signal detection apparatus, where the signal detection apparatus may be a network device or a terminal device, or may be a network device or a component (such as a processor, a chip, or a chip system) of the terminal device. In the method, a signal detection device determines channel state information of a received signal after receiving the received signal, wherein the received signal comprises a signal to be detected; then, the device determines the filtering parameters of the received signal according to the channel state information; then, the device determines a target statistical characteristic of the signal to be detected according to the received signal and the filtering parameter of the received signal, wherein the target statistical characteristic is used for indicating statistical information of posterior probability of the signal to be detected; finally, the device adopts a neural network to process the target statistical characteristics at least once and then outputs target bit soft information corresponding to the signal to be detected. The signal detection device determines the statistical characteristics of the signal to be detected according to the filtering parameters of the received signal, and outputs target bit soft information corresponding to the signal to be detected after processing the target statistical characteristics at least once by using a neural network, namely, the neural network is used for processing the statistical characteristics of the signal to be detected and outputting the corresponding bit soft information.

In a possible implementation manner of the first aspect of the embodiment of the present application, the process of outputting the target bit soft information corresponding to the signal to be detected after at least one processing on the target statistical feature by using a neural network may specifically include: and in the neural network, mapping the target statistical characteristics and outputting bit soft information corresponding to the signal to be detected.

In this embodiment, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in a neural network in multiple manners, and output corresponding target bit soft information, where after the target statistical feature of the signal to be detected is obtained, the target bit soft information of the signal to be detected is obtained through mapping processing in the neural network. Therefore, a specific implementation mode for processing and outputting the target bit soft information in the neural network is provided, and the feasibility of the scheme is improved.

In a possible implementation manner of the first aspect of the embodiment of the present application, the process of outputting the target bit soft information corresponding to the signal to be detected after at least one processing on the target statistical feature by using a neural network may specifically include: in the neural network, carrying out fusion processing on the target statistical characteristics to obtain first statistical characteristics, wherein the first statistical characteristics are used for indicating statistical information of the target bit soft information; and mapping the first statistical characteristic and then outputting target bit soft information corresponding to the signal to be detected.

In this embodiment, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in multiple manners in the neural network, and output corresponding target bit soft information, where after the target statistical feature of the signal to be detected is obtained, a first statistical feature for indicating the statistical information of the target bit soft information is obtained through fusion processing in the neural network, and then the first statistical feature is mapped, and the target bit soft information of the signal to be detected is output. Therefore, the input of the mapping process is optimized through the fusion process, and a specific implementation mode for processing and outputting the target bit soft information in the neural network is provided, so that the feasibility of the scheme is improved.

In a possible implementation manner of the first aspect of the embodiments of the present application, outputting, after processing the target statistical feature at least once by using a neural network, target bit soft information corresponding to the signal to be detected includes: in the neural network, mapping the target statistical characteristics to obtain first bit soft information corresponding to the signal to be detected; and correcting the value of the first bit soft information corresponding to the signal to be detected according to the first value range, and outputting the target bit soft information corresponding to the signal to be detected.

In this embodiment, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in a neural network in multiple manners, and output corresponding target bit soft information, where after obtaining the target statistical feature of the signal to be detected, after obtaining the first bit soft information corresponding to the signal to be detected through mapping processing in the neural network, the first bit soft information corresponding to the signal to be detected is subjected to correction processing according to the first value range, and the target bit soft information of the signal to be detected is output. Therefore, the mapping result of the mapping process is optimized through the correction process, and a specific implementation mode for processing and outputting the target bit soft information in the neural network is provided, so that the feasibility of the scheme is improved.

In a possible implementation manner of the first aspect of the embodiments of the present application, outputting, after processing the target statistical feature at least once by using a neural network, target bit soft information corresponding to the signal to be detected includes: in the neural network, the target statistical features are fused to obtain second statistical features, and the second statistical features are used for indicating statistical information of second bit soft information corresponding to the signal to be detected; mapping the second statistical characteristics to obtain second bit soft information corresponding to the signal to be detected; and correcting the value of the second bit soft information corresponding to the signal to be detected according to the second value range, and outputting the target bit soft information corresponding to the signal to be detected.

In this embodiment, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in multiple manners in the neural network, and output corresponding target bit soft information, where after the target statistical feature of the signal to be detected is obtained, a second statistical feature for indicating the statistical information of the second bit soft information corresponding to the signal to be detected is obtained through fusion processing in the neural network, and after the second statistical feature is mapped to obtain the second bit soft information, correction processing is performed on the value of the second bit soft information according to a second value range, and the target bit soft information of the signal to be detected is output. Therefore, the input of the mapping process is optimized through the fusion process, the mapping result of the mapping process is optimized through the correction process, a specific implementation mode for processing and outputting the target bit soft information in the neural network is provided, and the feasibility of the scheme is improved.

In a possible implementation manner of the first aspect of the embodiments of the present application, the target activation function of the neural network includes a ReLu activation function and/or a softplus activation function and/or a softmax activation function.

In this embodiment, the target activation function used during at least one processing of the neural network may include a ReLu activation function and/or a softplus activation function and/or a softmax activation function, etc. Thus, a number of specific implementations of the activation function used in the neural network are provided, enhancing the feasibility of the scheme.

In a possible implementation manner of the first aspect of the embodiments of the present application, the weight coefficient of the neural network is obtained by training a preset signal after the neural network generates a data tag through a preset algorithm.

In this embodiment, in the processing procedure of the neural network, the weight coefficient used is obtained by training the preset signal after the data label is generated by the preset algorithm, and compared with the processing procedure of using the actually transmitted real data as the label in the neural network, the design of the loss function can be simplified, the processing result of the neural network is optimized, and the optimal performance of the preset algorithm is approximated.

In a possible implementation manner of the first aspect of the embodiments of the present application, the preset algorithm may include a maximum likelihood detection MLD algorithm, or a linear minimum mean square error LMMSE algorithm.

In this embodiment, the neural network may participate in the implementation by using an MLD algorithm or an LMMSE algorithm in the process of training to obtain the weight coefficient, so that various specific implementation modes of the algorithm used in the process of training to obtain the weight coefficient of the neural network are provided, and the feasibility of the scheme is improved.

In a possible implementation manner of the first aspect of the embodiments of the present application, determining the filtering parameter of the received signal according to the channel state information includes: initial filtering parameters of the received signal are determined in a target manner based on the channel state information, wherein the target manner comprises QR decomposition, SQR decomposition processing, cholesky decomposition, linear detection processing, and the like.

In this embodiment, in the process of determining the filtering parameters of the received signal, the signal detection device may participate in the calculation by using a QR decomposition, SQR decomposition, cholesky decomposition, or linear detection process, so as to provide various specific implementation manners of the signal detection device in the process of determining the filtering parameters of the received signal, thereby improving the feasibility of the scheme.

In a possible implementation manner of the first aspect of the embodiments of the present application, the process of determining the filtering parameter of the received signal according to the channel state information may specifically include: the channel state information is subjected to regular processing according to a first preset parameter to obtain processed channel state information; then, the received signal is subjected to normalization processing according to a second preset parameter to obtain a first received signal; then, determining the filtering parameters of the received signal in the target mode according to the processed channel state information; and/or the determining the target statistical feature of the signal to be detected according to the received signal and the filtering parameter of the received signal comprises: performing normalization processing on the received signals according to a third preset parameter to obtain second received signals; then, the filtering parameters of the received signals are subjected to normalization processing according to fourth preset parameters to obtain first filtering parameters; and then determining the target statistical characteristics of the signal to be detected according to the second received signal and the first filtering parameter.

In this embodiment, in determining the initial filtering parameter of the received signal according to the channel state information, the signal detection device may first perform a normalization process on the channel state information and the received signal in the determining process by using the first preset parameter and the second preset parameter, and then further determine the filtering parameter of the first received signal after obtaining the processed channel state information and the first received signal. In addition, in the process of determining the target statistical characteristics of the signal to be detected according to the received signal and the filtering parameters of the received signal, the signal detection device may perform model normalization processing on the received signal and the filtering parameters of the received signal, respectively, to obtain the second received signal and the first filtering parameters, and then further determine the target statistical characteristics of the signal to be detected. The normalization process may also be referred to as a model normalization process, and the specific implementation may include channel power de-normalization, value range adjustment, column sequencing, etc., or other model normalization process operations. Compared with the data before the normalization processing, the data after the normalization processing is more beneficial to the value area (such as continuous integer value) of the subsequent processing and obtains better ordering sequence.

In a possible implementation manner of the first aspect of the embodiments of the present application, the determining, according to the received signal and the filtering parameter of the received signal, the target statistical feature of the signal to be detected may specifically include: randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result, wherein the sampling result is used for indicating the discrete information of the posterior probability of the signal to be detected; and calculating according to the sampling result to obtain the target statistical characteristics of the signal to be detected.

In this embodiment, the signal detection device may randomly sample the received signal according to the filtering parameter of the received signal to obtain a sampling result of discrete information for indicating the posterior probability of the signal to be detected, and then calculate the target statistical feature of the signal to be detected according to the sampling result, that is, determine the target statistical feature of the signal to be detected through a random sampling process. Therefore, a specific implementation mode for determining the target statistical characteristics of the signal to be detected according to the received signal and the filtering parameters of the received signal is provided, and the feasibility of the scheme is improved.

In a possible implementation manner of the first aspect of the embodiments of the present application, the process of randomly sampling the received signal according to the filtering parameter of the received signal to obtain a sampling result may specifically include: and randomly sampling the received signal according to the filtering parameters of the received signal and a target sampling structure to obtain a sampling result, wherein the target sampling structure comprises a serial sampling structure and/or a parallel sampling structure.

In this embodiment, the signal detection device may randomly sample the received signal through a serial sampling structure and/or a parallel sampling structure to obtain a sampling result. Thus, a specific implementation manner of the random sampling process is provided, and the feasibility of the scheme is improved.

The second aspect of the embodiments of the present application provides a signal detection method, where the method is applied to a signal detection apparatus, where the signal detection apparatus may be a network device or a terminal device, or may be a network device or a component (such as a processor, a chip, or a chip system) of the terminal device. In the method, a signal detection device determines channel state information of a received signal after receiving the received signal, wherein the received signal comprises a signal to be detected; determining a filtering parameter of the received signal according to the channel state information; randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result, wherein the sampling result is used for indicating the discrete information of the posterior probability of the signal to be detected; and obtaining the target bit soft information corresponding to the signal to be detected according to the sampling result. The signal detection device obtains a sampling result of discrete information for indicating posterior probability of the signal to be detected according to a mode of randomly sampling the received signal according to filtering parameters of the received signal, and compared with an implementation process of participating in calculation through multiple QR decomposition, the complexity of the calculation process can be reduced through the mode of randomly sampling, the signal processing process of the signal detection device is optimized, and the communication efficiency is improved.

In a possible implementation manner of the second aspect of the embodiments of the present application, determining the filtering parameter of the received signal according to the channel state information includes: initial filtering parameters of the received signal are determined in a target manner based on the channel state information, wherein the target manner comprises QR decomposition, SQR decomposition processing, cholesky decomposition, linear detection processing, and the like.

In a possible implementation manner of the second aspect of the embodiments of the present application, the process of determining the filtering parameter of the received signal according to the channel state information may specifically include: the channel state information is subjected to regular processing according to a first preset parameter to obtain processed channel state information; then, the received signal is subjected to normalization processing according to a second preset parameter to obtain a first received signal; then, determining the filtering parameters of the received signal in the target mode according to the processed channel state information; and/or the determining the target statistical feature of the signal to be detected according to the received signal and the filtering parameter of the received signal comprises: performing normalization processing on the received signal according to a third preset parameter to obtain a second received signal; then, the filtering parameters of the received signals are subjected to normalization processing according to fourth preset parameters to obtain first filtering parameters; and then determining the target statistical characteristics of the signal to be detected according to the second received signal and the first filtering parameter.

In this embodiment, in the process of determining the initial filtering parameter of the received signal according to the channel state information, the signal detection device may use the first preset parameter and the second preset parameter to perform normalization processing on the channel state information and the received signal in the determining process, so as to obtain the processed channel state information and the first received signal, and then further determine the filtering parameter of the first received signal. In addition, in the process of determining the target statistical characteristics of the signal to be detected according to the received signal and the filtering parameters of the received signal, the signal detection device may perform model normalization processing on the received signal and the filtering parameters of the received signal, respectively, to obtain the second received signal and the first filtering parameters, and then further determine the target statistical characteristics of the signal to be detected. The normalization process may also be referred to as a model normalization process, and the specific implementation may include channel power de-normalization, value range adjustment, column sequencing, etc., or other model normalization process operations. Compared with the data before the normalization processing, the data after the normalization processing is more beneficial to the value area (such as continuous integer value) of the subsequent processing and obtains better ordering sequence.

In a possible implementation manner of the second aspect of the embodiments of the present application, the process of randomly sampling the received signal according to the filtering parameter of the received signal to obtain a sampling result may specifically include: and randomly sampling the received signal according to the filtering parameters of the received signal and a target sampling structure to obtain a sampling result, wherein the target sampling structure comprises a serial sampling structure and/or a parallel sampling structure.

A third aspect of the embodiments of the present application provides a signal detection apparatus, where the signal detection apparatus may be a network device or a terminal device, and may also be a network device or a component (such as a processor, a chip, or a chip system) of the terminal device. The device comprises a determining unit and a processing unit: the determining unit is used for determining channel state information of a received signal, wherein the received signal comprises a signal to be detected; the determining unit is further configured to determine a filtering parameter of the received signal according to the channel state information; the determining unit is further configured to determine a target statistical feature of the signal to be detected according to the received signal and a filtering parameter of the received signal, where the target statistical feature is used to indicate statistical information of a posterior probability of the signal to be detected; the processing unit is used for processing the target statistical characteristics at least once by adopting a neural network and then outputting target bit soft information corresponding to the signal to be detected. The processing unit in the signal detection device determines the statistical characteristics of the signal to be detected according to the filtering parameters of the received signal, and outputs target bit soft information corresponding to the signal to be detected after processing the target statistical characteristics at least once by using the neural network, namely, the processing unit processes the statistical characteristics of the signal to be detected by using the neural network and outputs the corresponding bit soft information.

In a possible implementation manner of the third aspect of the embodiments of the present application, the processing unit is specifically configured to:

and in the neural network, mapping the target statistical characteristics and outputting bit soft information corresponding to the signal to be detected.

in the neural network, carrying out fusion processing on the target statistical characteristics to obtain first statistical characteristics, wherein the first statistical characteristics are used for indicating statistical information of the target bit soft information;

and mapping the first statistical characteristic and then outputting target bit soft information corresponding to the signal to be detected.

in the neural network, mapping the target statistical characteristics to obtain first bit soft information corresponding to the signal to be detected;

and correcting the value of the first bit soft information corresponding to the signal to be detected according to the first value range, and outputting the target bit soft information corresponding to the signal to be detected.

In the neural network, the target statistical features are fused to obtain second statistical features, and the second statistical features are used for indicating statistical information of second bit soft information corresponding to the signal to be detected;

mapping the second statistical characteristics to obtain second bit soft information corresponding to the signal to be detected;

and correcting the value of the second bit soft information corresponding to the signal to be detected according to the second value range, and outputting the target bit soft information corresponding to the signal to be detected.

In a possible implementation manner of the third aspect of the embodiments of the present application, the target activation function of the neural network includes a ReLu activation function and/or a softplus activation function and/or a softmax activation function.

In a possible implementation manner of the third aspect of the embodiments of the present application, the weight coefficient of the neural network is obtained by training a preset signal after the neural network generates a data tag through a preset algorithm.

In a possible implementation manner of the third aspect of the embodiment of the present application, the preset algorithm includes a maximum likelihood detection MLD algorithm, or a linear minimum mean square error LMMSE algorithm.

In a possible implementation manner of the third aspect of the embodiments of the present application, the determining unit is specifically configured to:

The filtering parameters of the received signal are determined in a target manner based on the channel state information, the target manner including Cholesky decomposition process, QR decomposition process, SQR decomposition process, or linear detection process.

In a possible implementation manner of the third aspect of the embodiments of the present application, the determining unit is specifically configured to: the channel state information is subjected to regular processing according to a first preset parameter to obtain processed channel state information; performing normalization processing on the received signal according to a second preset parameter to obtain a first received signal; determining a filtering parameter of the first received signal in the target mode according to the processed channel state information;

and/or the number of the groups of groups,

performing normalization processing on the received signal according to a third preset parameter to obtain a second received signal;

the filtering parameters of the received signals are subjected to normalization processing according to fourth preset parameters to obtain first filtering parameters;

and determining the target statistical characteristics of the signal to be detected according to the second received signal and the first filtering parameter.

randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result, wherein the sampling result is used for indicating the discrete information of the posterior probability of the signal to be detected;

And calculating according to the sampling result to obtain the target statistical characteristics of the signal to be detected.

and randomly sampling the received signal according to the filtering parameters of the received signal and a target sampling structure to obtain a sampling result, wherein the target sampling structure comprises a serial sampling structure and/or a parallel sampling structure.

In the third aspect of the embodiments of the present application, the constituent modules of the signal detection apparatus may also be configured to execute the steps executed in each possible implementation manner of the first aspect, and all details may refer to the first aspect, which is not repeated herein.

A fourth aspect of the present embodiment provides a signal detection apparatus, including a determining unit and a processing unit:

the determining unit is used for determining channel state information of a received signal, wherein the received signal comprises a signal to be detected;

the determining unit is further configured to determine a filtering parameter of the received signal according to the channel state information;

the determining unit is further configured to randomly sample the received signal according to a filtering parameter of the received signal, to obtain a sampling result, where the sampling result is used to indicate discrete information of posterior probability of the signal to be detected;

And the processing unit is used for obtaining the target bit soft information corresponding to the signal to be detected according to the sampling result.

In a possible implementation manner of the fourth aspect of the embodiments of the present application, the determining unit is specifically configured to:

the channel state information is subjected to regular processing according to a first preset parameter to obtain processed channel state information;

performing normalization processing on the received signal according to a second preset parameter to obtain a first received signal;

determining a filtering parameter of the first received signal in the target mode according to the processed channel state information;

and/or the number of the groups of groups,

In the fourth aspect of the embodiments of the present application, the constituent modules of the signal detection apparatus may also be used to perform the steps performed in each possible implementation manner of the second aspect, and reference may be specifically made to the second aspect, which is not described herein.

A fifth aspect of the embodiments of the present application provides a communication device, where the communication device includes a processor and a communication interface, where the communication interface is coupled to the processor, and where the processor is configured to execute a computer program or instructions, to perform a method according to any one of the foregoing first aspect or any one of the foregoing possible implementation manners of the first aspect, or to perform a method according to any one of the foregoing second aspect or any one of the possible implementation manners of the second aspect.

A sixth aspect of embodiments of the present application provides a computer-readable storage medium storing one or more computer-executable instructions which, when executed by a processor, perform a method as described above in the first aspect or any one of the possible implementations of the first aspect; alternatively, when the computer-executable instructions are executed by a processor, the processor performs the method as described above in relation to the second aspect or any one of the possible implementations of the second aspect.

A seventh aspect of the embodiments of the present application provides a computer program product (or computer program) storing one or more computers, which when executed by the processor performs the method of the first aspect or any one of the possible implementations of the first aspect; alternatively, the computer program product, when executed by the processor, performs the method of the second aspect or any one of the possible implementations of the second aspect.

An eighth aspect of the embodiments of the present application provides a chip system, which includes a processor, configured to support a network device to implement the functions involved in the first aspect or any one of the possible implementations of the first aspect; or for supporting the network device to implement the functionality referred to in the second aspect or any one of the possible implementations of the second aspect. In one possible design, the system on a chip may further include memory to hold the program instructions and data necessary for the network device. The chip system can be composed of chips, and can also comprise chips and other discrete devices.

A ninth aspect of the embodiments of the present application provides a communication system, which includes at least the communication device of the third aspect, or at least the communication device of the fourth aspect, or at least the communication device of the fifth aspect.

From the above technical solutions, the embodiments of the present application have the following advantages: the signal detection device can determine the statistical characteristics of the signal to be detected according to the filtering parameters of the received signal, and output the target bit soft information corresponding to the signal to be detected after at least one time of processing the target statistical characteristics by using the neural network, namely, the neural network is used for processing the statistical characteristics of the signal to be detected and outputting the corresponding bit soft information. In addition, the signal detection device can also obtain a sampling result of discrete information for indicating posterior probability of the signal to be detected by randomly sampling the received signal according to the filtering parameter of the received signal, and compared with the implementation process of participating in calculation through multiple QR decomposition, the complexity of the calculation process can be reduced through the random sampling mode, the signal processing process of the signal detection device is optimized, and the communication efficiency is improved.

Drawings

Fig. 1 is a schematic diagram of a communication system according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a signal processing process implemented by a signal detection apparatus using the QR algorithm;

fig. 3 is a schematic diagram of a signal processing method according to an embodiment of the present application;

FIG. 4-1 is a schematic diagram of a signal processing method according to an embodiment of the present disclosure;

fig. 4-2 is another schematic diagram of a signal processing method according to an embodiment of the present application;

fig. 4-3 are another schematic diagram of a signal processing method according to an embodiment of the present application;

fig. 5 is another schematic diagram of a signal processing method according to an embodiment of the present application;

fig. 6 is a schematic diagram of a signal processing apparatus according to an embodiment of the present application;

fig. 7 is another schematic diagram of a signal processing apparatus according to an embodiment of the present application.

Detailed Description

The following description of the technical solutions in the embodiments of the present application will be made with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.

First, some terms in the embodiments of the present application are explained for easy understanding by those skilled in the art.

1. Terminal equipment: may be a wireless terminal device capable of receiving network device scheduling and indication information, which may be a device providing voice and/or data connectivity to a user, or a handheld device having wireless connection capabilities, or other processing device connected to a wireless modem.

The terminal device may communicate with one or more core networks or the internet via a radio access network (radio access network, RAN), and may be a mobile terminal device, such as a mobile phone (or "cellular" phone), a computer and a data card, for example, a portable, pocket, hand-held, computer-built-in or vehicle-mounted mobile device that exchanges voice and/or data with the radio access network. Such as personal communication services (personal communication service, PCS) phones, cordless phones, session Initiation Protocol (SIP) phones, wireless local loop (wireless local loop, WLL) stations, personal digital assistants (personal digital assistant, PDAs), tablet computers (Pad), computers with wireless transceiver capabilities, and the like. The wireless terminal device may also be referred to as a system, subscriber unit (subscriber unit), subscriber station (subscriber station), mobile Station (MS), remote station (AP), access Point (AP), remote terminal device (remote), access terminal device (access terminal), user terminal device (user terminal), user agent (user agent), user station (subscriber station, SS), user equipment (customer premises equipment, CPE), terminal (terminal), user Equipment (UE), mobile Terminal (MT), etc. The terminal device may also be a wearable device as well as a next generation communication system, e.g. a terminal device in a 5G communication system or a terminal device in a future evolved public land mobile network (public land mobile network, PLMN), etc.

2. Network equipment: may be a device in a wireless network, for example, a network device may be a radio access network (radio access network, RAN) node (or device), also referred to as a base station, that accesses a terminal device to the wireless network. Currently, some examples of RAN equipment are: a new generation base station (generation Node B, gNodeB), a transmission reception point (transmission reception point, TRP), an evolved Node B (eNB), a radio network controller (radio network controller, RNC), a Node B (Node B, NB), a base station controller (base station controller, BSC), a base transceiver station (base transceiver station, BTS), a home base station (e.g., home evolved Node B, or home Node B, HNB), a baseband unit (BBU), or a wireless fidelity (wireless fidelity, wi-Fi) Access Point (AP), etc. in a 5G communication system. In addition, in one network structure, the network device may include a Centralized Unit (CU) node, or a Distributed Unit (DU) node, or a RAN device including a CU node and a DU node.

The network device can send configuration information (for example, carried in a scheduling message and/or an indication message) to the terminal device, and the terminal device further performs network configuration according to the configuration information, so that network configuration between the network device and the terminal device is aligned; or, the network configuration between the network device and the terminal device is aligned through the network configuration preset in the network device and the network configuration preset in the terminal device. Specifically, "alignment" refers to the coincidence of the two understandings of the carrier frequency of the interactive messaging, the determination of the type of interactive message, the meaning of field information carried in the interactive message, or other configuration of the interactive message, when there is an interactive message between the network device and the terminal device.

Furthermore, the network device may be other means of providing wireless communication functionality for the terminal device, as other possibilities. The embodiment of the application does not limit the specific technology and the specific device form adopted by the network device. For convenience of description, embodiments of the present application are not limited.

The network devices may also include core network devices including, for example, access and mobility management functions (access and mobility management function, AMF), user plane functions (user plane function, UPF), or session management functions (session management function, SMF), etc.

In the embodiment of the present application, the means for implementing the function of the network device may be the network device, or may be a means capable of supporting the network device to implement the function, for example, a chip system, and the apparatus may be installed in the network device. In the technical solution provided in the embodiments of the present application, the device for implementing the function of the network device is exemplified by the network device, and the technical solution provided in the embodiments of the present application is described.

3. Neural network: the neural network may be composed of neural units, which may refer to an arithmetic unit having xs and intercept 1 as inputs, and the output of the arithmetic unit may be:

Where s=1, 2, … … n, n is a natural number greater than 1, ws is the weight of xs, and b is the bias of the neural unit. f is an activation function (activation functions) of the neural unit for introducing a nonlinear characteristic into the neural network to convert an input signal in the neural unit to an output signal. The output signal of the activation function may be used as an input to the next convolutional layer. The activation function may be a sigmoid function. A neural network is a network formed by joining together a number of the above-described single neural units, i.e., the output of one neural unit may be the input of another. The input of each neural unit may be connected to a local receptive field of a previous layer to extract features of the local receptive field, which may be an area composed of several neural units. One neural network may include a "layer" of a plurality of neural elements, such as an input layer, a hidden layer, and an output layer. The input layer is responsible for receiving input data and distributing the input data to the hidden layers, the hidden layers are responsible for required calculation and output results to the output layer, and the output layer outputs the output results.

4. The terms "system" and "network" in embodiments of the present application may be used interchangeably. "at least one" means one or more, and "a plurality" means two or more. "and/or", describes an association relationship of an association object, and indicates that there may be three relationships, for example, a and/or B, and may indicate: cases where A alone, both A and B together, and B alone, where A and B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, "at least one of A, B, and C" includes A, B, C, AB, AC, BC, or ABC. And, unless otherwise specified, references to "first," "second," etc. in the embodiments herein are for distinguishing between multiple objects and not for defining the order, timing, priority, or importance of the multiple objects.

Fig. 1 is a schematic diagram of a communication system in the present application. In the communication system shown in fig. 1, at least one network device and at least one terminal device are included, such as the network device 101, and the terminal device 102, the terminal device 103, the terminal device 104, the terminal device 105, the terminal device 106, and the terminal device 107 shown in fig. 1, for example. In the example shown in fig. 1, terminal device 102 is a vehicle, terminal device 103 is an intelligent air conditioner, terminal device 104 is an intelligent fuel dispenser, terminal device 105 is a mobile phone, terminal device 106 is an intelligent teacup, and terminal device 107 is a printer.

In the communication process of the communication system shown in fig. 1, the signal transmitting end may be a network device and/or a terminal device, the signal receiving end may also be a network device and/or a terminal device, and the signal transmission process between the signal transmitting end and the signal receiving end may apply a multiple-input multiple-output (MIMO) technology. MIMO technology is a hotspot of current research in the field of wireless communication, and various novel mobile communication systems adopt MIMO technology to improve the spectrum efficiency of the system. The MIMO technology can increase the spatial dimension of data multiplexing, so that multiple data can be spatially multiplexed to the same time-frequency resource, and the same data can be transmitted by multiple antennas and/or received by multiple receiving antennas, thereby obtaining the spatial diversity gain.

In the transmission process of applying the MIMO technology, the signal transmitting end maps bit information to constellation symbols, modulates the constellation symbols into signals, and transmits the signals through the transmitting antennas. Correspondingly, the signal receiving end receives the signal through the receiving antenna, demodulates the signal into the sending probability of the constellation symbol, determines bit soft information according to the probability result, and finally decodes according to the bit soft information, for example, through searching the path result, and determines the bit information in the searching path result. In the signal transmitting end, the baseband processing generally comprises two important processes of channel coding and modulation, and correspondingly, in the signal receiving end, the baseband processing comprises demodulation and channel decoding processes. The processing procedures of the signal transmitting side and the signal receiving side will be described below, respectively.

In the signal transmitting end, the modulation process maps the coded bit information (with a value of 0 or 1) into constellation symbols, taking 16QAM modulation in quadrature amplitude modulation (quadrature amplitude modulation, QAM) modulation as an example, that is, maps every 4 bits (corresponding to 2^4 =16 states) of information into constellation symbols of a two-dimensional (commonly called I/Q path) plane, and there is a fixed mapping relationship between each bit combination state and each modulation symbol, such as gray code (gray code) mapping. The transmitting end then carrier modulates and transmits the constellation symbols, wherein the carrier modulation may be implemented by an orthogonal frequency division multiplexing technique (orthogonal frequency division multiplexing, OFDM).

In the signal receiving end, the signal receiving end needs to calculate bit soft information, namely, the probability that each information bit is 0 or 1 is obtained according to the fixed mapping relation of the bit state and the modulation symbol, and the process can be represented by the formula (1):

in formula (1), P _s (i) Representing the demodulation probability corresponding to the ith symbol, P _b (j) Representing the probability corresponding to the j-th bit.

Thereafter, the decoder in the signal receiving end generates a bit-soft information based on the logarithmic ratio (i.eAnd/or the number of the groups of groups,) Decoding is achieved. Obviously, in the formula (1), the probability sum of all constellation symbols corresponding to the j-th bit value of 0 and the j-th bit value of 1 can be obtained: p (P) _b (b _j ＝0)+P _b (b _j ＝1)＝1。

In addition, due to the influence of noise, interference and the like in the practical communication system and the wide application of the MIMO technology, before constellation demodulation, that is, before the process of the calculation formula (1), the MIMO detection technology is needed to obtain the constellation probability estimation result, where the constellation probability estimation result may be denoted as P _s (i) A. The invention relates to a method for producing a fibre-reinforced plastic composite Specifically, under the assumption of gaussian noise and ideal channel information, maximum likelihood detection (maximal likelihood detection, MLD) is an optimal detection algorithm, but the complexity thereof increases exponentially with the increase of modulation modes and MIMO multiplexing streams, and the practical system cannot be realized. Practically implementable algorithms can be divided into The nonlinear detection algorithm is also called a suboptimal MLD algorithm, which has performance approaching that of the optimal MLD algorithm. The linear detection algorithm has low implementation complexity, but the performance is larger than that of the optimal MLD algorithm; the nonlinear detection algorithm performance approaches the optimal MLD algorithm, but is highly complex to implement and disadvantageous for friendly implementation with general purpose processors, such as graphics processors (graphic processing unit, GPU), digital signal processors (digital signal processor, DSP), etc.

Wherein, for the signal receiving end, the received signal model can be represented by formula (2):

y＝Hs+n (2)

in formula (2), y represents a received signal vector, H represents a channel matrix of the MIMO system, s represents a transmitted constellation symbol vector, and n represents independent additive zero-mean complex gaussian noise.

In an actual communication system, n is an unknown random noise, y is a received signal, and the received signal y at least includes pilot symbols (s is known) and data symbols (s is unknown). For the signal receiving end, H (called channel estimation) can be obtained by pilot symbol (known as y and s) estimation, and then this channel estimation result is used for data symbol together for s estimation.

Based on the received signal model shown in formula (2), the sub-optimal MLD soft detection process using the nonlinear detection algorithm can be approximated by formula (3):

/>

in equation (3), y-Hs ² The value of the metric is indicated and,represents all possible sets of constellation symbols s, { s ₁ ,s ₂ ,…,s _K And the symbol s represents k constellation combination paths corresponding to the minimum value of the metric value.

Further, taking into account the highThe influence of the SIGMA noise can be calculated approximately by using max-log algorithm to solve the formula (1), namely, bit-by-bit from the set { s } ₁ ,s ₂ ,…,s _K The seed picks the smallest metric as output, noted asNamely by the formula (4):

thereafter, the results of taking 0 and 1 for each bit are subtracted to obtain the final bit log-likelihood ratio (LLR), i.e., there areOr->

In order to solve the formula (3), a classical multipath parallel search algorithm may be adopted, and an implementation block diagram of the overall scheme is shown in fig. 2, specifically including:

step 1, performing column replacement on a channel matrix H, so as to obtain channel matrix results of different column ordering combinations, and marking as: { H ₁ ,H ₂ ,…,H _M }, wherein H _m ＝HP _m ,m＝1,2,…,M，P _m Is a column permutation matrix.

Step 2, sequentially performing orthogonal triangular (QR) decomposition on all the ordered channel matrixes obtained in the step 1, namely H _m ＝Q _m R _m M=1, 2, …, M. Here Q _m As unitary matrices, i.e.R _m Is an upper triangular matrix.

Step 3, using matrixDockingThe received signal y and the channel matrix H are filtered to obtain an equivalent received signal model, namely:

step 4, R is utilized _m The triangular structure of the matrix can realize the search of constellation points and the calculation of corresponding measurement values in a serial equation solving mode. In particular, for the sake of presentation brevity, notes) The method can obtain:

wherein, for the last N layer, the following can be obtainedThen according to the modulation mode, from the constellation set +.>Select distance->The nearest P constellations are used as candidate constellations, then the constellation value is carried into the equation formed by the previous layers to be eliminated, and the equation solution is hard judged to be the constellation set +.>And brought into the previous layers, i.e. to achieve multipath parallel serial interference cancellation (successive interference cancellation, SIC). And finally, calculating the metric value corresponding to each group of search results, wherein the metric value calculation expression corresponding to the ith group is as follows:

η _i ＝||y-Hs _i || ² ,i＝1,2,…,M×P，

wherein s is _i For the ith group of constellation search results, η _i And measuring values corresponding to the set of search results.

And (3) executing the steps 2 to 4 on each column replacement result in the step (1), obtaining M multiplied by P constellation path search results and metric values thereof, and finally, approximately solving the problem of the formula (1) according to the solving method of the formula (4), thereby obtaining the bit log likelihood ratio LLR value.

From the above, the signal receiving end has at least the following problems in the process of implementing bit soft information estimation: when LLR calculation is carried out, namely, in the solving process of the formula (4), a max-log algorithm is adopted to select from the search constellation path results bit by bit, and the LLR calculation is irregular searching and comparing operation, so that the efficient realization of a general processor is not facilitated; in addition, in the nonlinear detection process, namely, in the solving process of the formula (3), QR decomposition needs to be carried out for a plurality of times, so that the algorithm implementation complexity is high, and the communication efficiency is influenced.

In order to solve the problem that the processing procedure for implementing bit soft information estimation is low in efficiency, the embodiments of the present application provide various schemes, which can be implemented from different angles for solving the problem, and will be described in detail below.

Fig. 3 is a schematic diagram of a signal detection method according to an embodiment of the present application, and as shown in fig. 3, the signal detection method includes the following steps.

S101, determining channel state information of a received signal;

in this embodiment, the signal detection apparatus determines channel state information of a received signal after receiving the received signal, where the received signal includes a signal to be detected.

Specifically, the signal detection device may be a network device or a terminal device, or may be a network device or a component (such as a processor, a chip, or a chip system) of the terminal device.

In step S101, the signal detection apparatus may acquire a received signal including a signal to be detected by means of wired communication or wireless communication, and determine channel state information of the received signal. The received signal may be composed of transmission signals of a plurality of different transmission ends, and for the signal detection device, the received signal includes signals to be detected of a plurality of transmission ends, and the signals to be detected of each transmission end are detected through the process from step S101 to step S104 in this embodiment, so as to obtain soft bit information corresponding to each signal to be detected.

Illustratively, in step S101, the signal detection apparatus may determine channel state information of the received signal through a model shown in equation (2), that is, the channel matrix H used to represent the MIMO system in equation (2) is determined as the channel state information of the received signal.

S102, determining a filtering parameter of the received signal according to the channel state information;

in this embodiment, the signal detection device may further determine the filtering parameters of the received signal according to the channel state information determined in step S101.

In a possible implementation manner, in the implementation process of step S102, the signal detection device may determine, according to the channel state information, a filtering parameter of the received signal in a target manner, where the target manner may include a Cholesky decomposition process, a QR decomposition process, a sequential QR decomposition (sorted QR decomposition, SQR) process, or a linear detection process, or other processing manners are implemented, which are not limited herein. The following will be described by way of specific examples:

the process of determining the filtering parameters of the received signal in the mode a and the step S102 may be obtained through Cholesky decomposition, and specifically includes the following steps:

step A1, calculating a channel correlation matrix according to the channel state information, namely:

R _hh ＝H ^H H+I；

wherein R is _hh Is a channel correlation matrix, H ^H Is the transposed conjugate of H, I is used to indicate noise covariance information.

Specifically, when the variance of the complex gaussian noise is 1, I, which is used to indicate noise covariance information, is an identity matrix. More generally, when the noise is colored noise caused by interference or the likeIn the case of sound, the noise covariance information may be in other forms, but all may be equivalently transformed into an identity matrix by a whitening operation. Illustratively, the whitening operation refers to, for a general noise covariance matrix R _uu Can be subjected to Cholesky decomposition to give R _uu ＝LL ^H Then inverting the decomposition result to obtain a whitening matrix L ^-1 The noise covariance matrix of the equivalent model can be obtained by multiplying the whitening matrix by the received signal and the channel matrix H as the identity matrix I.

And A2, performing Cholesky decomposition on the channel correlation matrix to obtain:

/>

wherein L is _h In the form of a lower triangular matrix,is L _h Is a transposed conjugate matrix of (a).

Step A4, solving the triangular matrix L _h The inverse of (2), noted as:and use +.>Filtering the received signal to obtain:

or->

Wherein y is used for indicating the received signal before filtering or the equivalent signal after model normalization,for indicating the filtered signal. The filtering parameters include L _h 、/>H ^H And->H ^H Or in other forms transformed therefrom, without limitation.

In mode B, the process of determining the filtering parameters of the received signal in step S102 may be obtained by QR decomposition, and specifically includes the following steps:

step B1, performing QR decomposition or SQR decomposition according to the channel state information to obtain:

wherein I is used to indicate noise covariance information; q, R is used for indicating the unitary matrix and the upper triangular matrix obtained after decomposition; the SQR decomposition yields decomposition results in a form that is exactly identical to the QR decomposition, except that it is further provided with a matrix P for column permutation.

Step B2, according to the QR decomposition result, the following steps are obtained:

where y is used to indicate the received signal before filtering or the equivalent signal after model normalization,for indicating the filtered signal. The filtering parameters include Q ^H And R, and the partial elements transformed thereby or extracted therefrom, e.g. extracted from a block matrix, i.e. fromExtract Q ₁ 、Q ₁ 、Q ₂ 、Q ₃ 、Q ₄ One or more of them as filtering parametersThe numbers, and other forms resulting from the transformation, are not limited herein.

In the mode C, the process of determining the filtering parameters of the received signal in the step S102 may be obtained through a linear detection method, for example, a linear minimum mean square error (linear minimum mean square error, LMMSE) algorithm, a maximum ratio combining detection algorithm, a zero forcing detection algorithm, or the like. A linear minimum mean square error (linear minimum mean square error, LMMSE) algorithm is described herein as an example of a linear detection method. The LMMSE algorithm can be used to obtain:

step C1: and filtering the received signal or the equivalent signal after the model is regulated by utilizing a filtering parameter corresponding to the LMMSE, namely:

wherein I is used to indicate noise covariance information, (H) ^H H+I) ^-1 H ^H The filtering parameters, or other forms resulting from the transformation, are not limited herein.

Optionally, in the implementation of step S102, the implementation may be further optimized according to an introduced normalization process, where the normalization process may also be referred to as a model normalization process, and the specific implementation may include channel power de-normalization, value range adjustment, column permutation, and the like, or other model normalization processing operations. Compared with the data before the normalization processing, the data after the normalization processing is more beneficial to the value area (such as continuous integer value) of the subsequent processing and obtains better ordering sequence.

Specifically, in the implementation process of step S102, the signal detection apparatus may further determine the filtering parameters of the received signal after performing the model normalization process on the channel state information. The signal detection device performs normalization processing on the channel state information according to a first preset parameter to obtain processed channel state information, and performs normalization processing on the received signal according to a second preset parameter to obtain a first received signal; then, step A1, step A2 and step A3 in the above-mentioned mode a are executed again, or step B1 and step B2 in the above-mentioned mode B are executed again, or step C1 in the above-mentioned mode C is executed again, that is, the filtering parameters of the first received signal are determined according to the processed channel state information. The implementation of the first preset parameter and the second preset parameter may be a preset numerical value range, a preset vector, a preset matrix or other parameter implementation, which is not limited herein.

Exemplary, the implementation process of the model normalization in step S102 specifically includes the following steps:

firstly, finishing model normalization operation according to the modulation mode of the signal to be detected, and aiming at transforming the signal to be detected to a value area (such as continuous integer value) which is more favorable for subsequent processing and obtaining a better ordering sequence. Taking a wireless communication system as a QAM modulation transmission scheme commonly used in a long term evolution (long term evolution, LTE) system or a New Radio (NR) system as an example, channel state information may be converted into:

accordingly, the received signal may be further transformed into:

wherein e represents a column vector of all 1 elements, and the dimension of the column vector is the same as the column number of H; the gamma vector is a power normalization factor determined by the modulation mode of the signal to be detected, and the values of the gamma vector are shown in the following table 1 according to different modulation modes; diag { γ } represents a diagonal matrix composed of vectors γ (i.e., the first preset parameter is implemented by 2diag { γ } in the above formula, and the second preset parameter is implemented by Hdiag { γ } e in the above formula).

Optionally, the matrix can be further processedProceeding withAnd (3) proper column replacement operation, and realizing the reordering of each stream. The permutation order or column permutation matrix P can be applied by P +. >The SQRD operation may be performed to obtain a preset configuration in other manners. Then the reordered equivalent channel is +.>

In addition, the processing of the model normalization operation is an alternative implementation, and the model normalization operation may be used in the processing of this and subsequent embodimentsOr->As channel state information->As the received signal, H may be used as channel state information and y may be used as the received signal to participate in the calculation without undergoing the model normalization operation, which is not limited herein.

TABLE 1

S103, determining target statistical characteristics of the signal to be detected according to the received signal and the filtering parameters of the received signal;

in this embodiment, the signal detection device determines a target statistical feature of the signal to be detected according to the received signal and the filtering parameter of the received signal obtained in step S102, where the target statistical feature is used to indicate statistical information of a posterior probability of the signal to be detected.

In one possible implementation, the optimization may also be further based on introducing a normalization process, which may also be referred to as a model normalization process, which may include channel power de-normalization, value range adjustment, column ordering, etc., or other model normalization process operations. Compared with the data before the normalization processing, the data after the normalization processing is more beneficial to the value area (such as continuous integer value) of the subsequent processing and obtains better ordering sequence.

Specifically, in the implementation process of step S103, the signal detection apparatus may further determine the target statistical feature of the signal to be detected after performing the model normalization process on the received signal and the filtering parameter of the received signal to obtain the second received signal and the first filtering parameter. The third preset parameter and the fourth preset parameter may be preset numerical value ranges, preset vectors, preset matrixes or other parameter implementations, and are not limited herein.

Exemplary, the filtering parameters of the received signal obtained by means of mode A in step S102, i.e. Cholesky decomposition, areFor example, in the implementation of the model normalization in step S103, the received signal may be converted into:

the filtering parameters of the received signal may be transformed into:

wherein, the liquid crystal display device comprises a liquid crystal display device,ρ is a matrix +.>Is a master pair of (2)Diagonal matrix composed of corner line elements; gamma is a diagonal matrix formed by power normalization factors generated by constellation modulation of the signal to be detected, and the corresponding power normalization factors under different modulation modes are as shown in the table 1.

In addition, the processing of the model normalization operation is an alternative implementation, and the model normalization operation may be used in the processing of this and subsequent embodiments As filtering parameters for the received signal, < >>The received signal can be used as a received signal to participate in calculation without model normalization operation>(or other forms of the foregoing forms a, B, and C) participate in calculation as a filtering parameter of the received signal and y as the received signal, and are not limited herein.

In a possible implementation manner, in step S103, the target statistical feature of the signal to be detected may be determined through a linear processing procedure, where the linear processing procedure may be implemented through an LMMSE algorithm, a maximum ratio combining detection algorithm, a zero forcing detection algorithm, and the like, and is not limited herein.

Taking the implementation procedure of the LMMSE algorithm in step S102 (i.e., mode C) as an example, the filtering parameter of the received signal can be determined to be (H by mode C ^H H+I) ^-1 H ^H The method comprises the steps of carrying out a first treatment on the surface of the Filtering resultIs the corresponding first moment statistical feature, (H) ^H H+I) ^- ¹ H ^H The main diagonal elements of the matrix shown in H are the corresponding second moment statistics, i.e. it can be determined in step S103 by LMMSE algorithm that the target statistics of the signal to be detected include the first moment statistics +.>And second moment statistics (H) ^H H+I) ^- ¹ H ^H H。

In a possible implementation manner, in step S103, the process of determining the target statistical feature of the signal to be detected according to the received signal and the filtering parameter of the received signal may specifically include: randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result, wherein the sampling result is used for indicating the discrete information of the posterior probability of the signal to be detected; and calculating according to the sampling result to obtain the target statistical characteristics of the signal to be detected. The random sampling is used for selecting a plurality of constellation combination paths with the smallest searching metric value with a maximum probability, namely determining target statistical characteristics of statistical information for indicating posterior probability of a signal to be detected.

Specifically, the process of randomly sampling the received signal according to the filtering parameter of the received signal to obtain the sampling result may specifically include: and randomly sampling the received signal according to the filtering parameters of the received signal and a target sampling structure to obtain a sampling result, wherein the target sampling structure comprises a serial sampling structure and/or a parallel sampling structure. That is, the signal detection device may randomly sample the received signal through a serial sampling structure and/or a parallel sampling structure to obtain a sampling result.

The random sampling process in step S103 may be implemented by various algorithms, and an exemplary sampling algorithm is provided in this embodiment to obtain the received signal in step B2And a filter parameter R, specifically including:

step D1, estimating the signal to be detected layer by layer from bottom to top by using the upper triangle structure of the matrix R and eliminating the interference between serial streams (the estimated result of the ith layer is recorded asAnd according to the modulation mode of each layer/stream, randomly quantizing the estimation result to a candidate constellation point to be marked as +.>

For example, mode one:where Q (-) represents the most recent constellation quantization operation, epsilon _i Representing a preset random sampling number; mode two: / >Where Q (-) represents the most recent constellation quantization operation, Δ _i Representing legal constellation random offset, and randomly offsetting the current constellation to other constellation points, wherein the current constellation random offset can be given through pre-configuration; or by other means, not limited herein.

Step D2, repeating step D1 (i.e. configuring different random sampling numbers or offsets) independently for multiple times, so as to obtain multiple sets of sampling results, or referred to as sampling path results, which are recorded as:

wherein N is the total sampling times,representing the sampling path result, where N _t Is the total number of streams of the signal to be detected.

Step D3, calculating a metric value corresponding to each sampling path result, wherein eta= [ eta ] ₁ ,…,η _N ]。

One possible calculation method is:

wherein, the liquid crystal display device comprises a liquid crystal display device, I.I. | ² Taking the square of the mouldAnd (3) operating. In addition, the measurement value can be obtained by equivalent calculation of the filtering parameter, which is not described herein.

Optionally, the outputs of the steps D1 to D3 are derived from the received signal obtained in the step B2And the filtering parameters are replaced by the received signal obtained in step A3 of mode A, respectively +.>And filtering parameters->Then, the steps D1 to D4 may be executed to obtain the sampling result and the corresponding measurement value, or the sampling result and the corresponding measurement value may be realized in the mode C or in other modes, and the implementation process and the steps D1 to D3 are not repeated here.

Optionally, the random sampling algorithm uses serial sampling structures in step D1, step D2 and step D3, and may be adjusted by replacing parallel sampling structures to obtain the received signal in step A3And filtering parameters->For example, the following is specific:

e1, calculating a random sampling estimation result:

where ε represents a random disturbance vector that is set in advance.

E2, carrying out nearest constellation point quantization on the sampling result:

where Q (-) represents the most recent constellation quantization operation.

And E3, independently repeating the steps E1 to E2 for a plurality of times (configuring different random disturbance vectors epsilon), and obtaining a plurality of groups of sampling results, namely:

Step E4, calculating a metric value corresponding to each sampling path, wherein eta= [ eta ] ₁ ,…,η _N ]。

One possible calculation method is:

wherein, the liquid crystal display device comprises a liquid crystal display device, I.I. | ² And (5) performing modular squaring operation. In addition, the measurement value can be obtained by equivalent calculation of the filtering parameter, which is not described herein.

In addition, the signal detection device may further calculate, after obtaining the sampling result, a target statistical feature of the signal to be detected according to the sampling result, where the sampling result is obtained based on the above example of the random sampling algorithm And η, the calculation process comprises:

step F1, calculating path normalization probability to obtain:

p＝softmax(η)

where p is used to indicate a path normalized probability row vector, each element of which represents the normalized probability value corresponding to the sampled path, softmax () represents the activation function,for normalizing the input vector to a probability value, the calculation expression of the nth output value isη is used to indicate the metric value to which the sampling path corresponds.

And F2, respectively calculating first moment characteristics by a real part and an imaginary part to obtain:

wherein mu _r/i Column vectors for indicating the first order feature moment composition of the real/imaginary part correspondence of each stream, real/imag () for indicating the real/imaginary part taking operation, p ^T Indicating the transpose of p.

And F3, respectively calculating second moment characteristics by a real part and an imaginary part to obtain:

wherein, the liquid crystal display device comprises a liquid crystal display device,for indicating the corresponding additional moment characteristics of the real part/imaginary part of each stream, e for indicating the full 1-row vector with the dimension equal to the sampling path number N, p ^T Indicating the transpose of p.

As can be obtained from the above calculation process, it can be determined in step S103 that the target statistical features of the signal to be detected include the above first moment feature μ _r/i And second moment features

S104, after at least one time of processing is carried out on the target statistical characteristics by adopting a neural network, outputting target bit soft information corresponding to the signal to be detected.

In this embodiment, after the signal detection device obtains the target statistical feature in step S103, the target statistical feature is used as an input of the neural network, and after the neural network processes the target statistical feature at least once, the signal detection device outputs and obtains the target bit soft information corresponding to the signal to be detected, thereby completing the detection process of the signal to be detected in the received signal.

The training process of the neural network referred to in the present application will be exemplarily described below.

Referring to fig. 4-1, an embodiment of the present invention provides a system architecture 100. The data acquisition device 160 is configured to acquire the received signal and/or the target statistical features of the signal to be detected in the received signal as input data and store the input data in the database 130, and the training device 120 generates the target model/rule 101 based on the input data maintained in the database 130. How the training device 120 derives the target model/rule 101 based on the input data will be described in more detail below.

The operation of each layer in the deep neural network can be expressed mathematicallyTo describe: the work of each layer in a physical layer deep neural network can be understood as completing the transformation of input space into output space (i.e., row space to column space of the matrix) by five operations on the input space (set of input vectors), including: 1. dimension increasing/decreasing; 2. zoom in/out; 3. rotating; 4. translating; 5. "bending". Wherein the operations of 1, 2, 3 are defined by +. >The operation of 4 is completed by +b, and the operation of 5 is implemented by a (). The term "space" is used herein to describe two words because the object being classified is not a single thing, but rather a class of things, space referring to the collection of all individuals of such things. Where W is a weight vector, each value in the vector representing a weight value of a neuron in the layer neural network. The vector W determines the spatial transformation of the input space into the output space described above, i.e. the weights W of each layer control how the space is transformed. The purpose of training deep neural networks, namely, the mostThe weight matrix of all layers of the trained neural network (the weight matrix formed by the vectors W of many layers) is finally obtained. Thus, the training process of the neural network is essentially a way to learn and control the spatial transformation, and more specifically to learn the weight matrix.

Since the output of the deep neural network is expected to be as close as possible to the value actually expected, the weight vector of each layer of the neural network can be updated by comparing the predicted value of the current network with the actually expected target value and according to the difference between the predicted value and the actually expected target value (of course, there is usually an initialization process before the first update, that is, the preconfigured parameters of each layer in the deep neural network). For example, if the predicted value of the network is high, the weight vector is adjusted to make it predicted lower, and the adjustment is continued until the neural network can predict the truly desired target value. Thus, it is necessary to define in advance "how to compare the difference between the predicted value and the target value", which is a loss function (loss function) or an objective function (objective function), which are important equations for measuring the difference between the predicted value and the target value. Taking the loss function as an example, the higher the output value (loss) of the loss function is, the larger the difference is, and then the training of the deep neural network becomes a process of reducing the loss as much as possible.

The target models/rules obtained by training device 120 may be applied in different systems or devices. In fig. 4-1, the executing device 110 is configured with an I/O interface 112 for data interaction with external devices, and a "user" may input data to the I/O interface 112 through the client device 140.

The execution device 110 may call data, code, etc. in the data storage system 150, or may store data, instructions, etc. in the data storage system 150. The signal detection device in the embodiment of the present application may include a process of implementing the neural network by the execution device 110, or a process of implementing the neural network by externally connecting the execution device 110, which is not limited herein.

The calculation module 111 processes the input data by using the target model/rule 101, for example, in step S104, the calculation module 111 processes the input target statistical feature at least once to obtain the target bit soft information corresponding to the signal to be detected

Finally, the I/O interface 112 returns the processing results to the client device 140 for presentation to the user.

Further, the training device 120 may generate corresponding target models/rules 101 for different targets based on different data to provide better results to the user.

It should be noted that fig. 4-1 is only a schematic diagram of a system architecture provided by an embodiment of the present invention, and the positional relationship among devices, apparatuses, modules, etc. shown in the figure is not limited in any way, for example, in fig. 4-1, the data storage system 150 is an external memory with respect to the execution device 110, and in other cases, the data storage system 150 may be disposed in the execution device 110.

In a possible implementation manner, in step S104, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in a neural network in multiple manners, and output corresponding target bit soft information, and specifically, may obtain the target bit soft information of the signal to be detected through mapping processing in the neural network after obtaining the target statistical feature of the signal to be detected.

The mapping process is exemplified by fig. 4-2, in which, illustratively, the target statistical features (first moment features μ _r/i And second moment features) As input data, after the mapping process of the hidden layer, it is output as bit soft information. The hidden layer may have a plurality of computing nodes, and for different modulation modes, the number of the computing nodes may have different values (for example, 4, 8, 16 and …). The output node number is used for indicating the modulation bit number and is related to the modulation mode. Taking 256QAM modulation as an example, and the bit mapping employs gray mapping, i.e. 8 bits for each constellation symbol (due to 2 ⁸ =256), the real part and the imaginary part of each constellation symbol are mapped separately,i.e. 4 bits each, so the output node is 4; taking 16QAM modulation as an example, and the bit mapping employs gray mapping, i.e. 4 bits per constellation symbol (due to 2 ⁴ =16), the real part and the imaginary part of each constellation symbol are mapped separately, i.e. each occupies 2 bits, under gray mapping, so the output node is 2.

In a possible implementation manner, in step S104, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in multiple manners in the neural network, and output corresponding target bit soft information, where after obtaining the target statistical feature of the signal to be detected, after obtaining a first statistical feature of statistical information indicating the target bit soft information through fusion processing in the neural network, mapping processing is performed on the first statistical feature, and the target bit soft information of the signal to be detected is output.

Taking fig. 4-2 as an example, the process of obtaining the first statistical feature of the statistical information for indicating the target bit soft information through the fusion processing in the neural network specifically includes: multiple types of input target statistical features (first moment features and second moment features) are combined into one type, so that the feature dimension after combination is larger than the feature dimension before combination is input. Illustratively, the input first-order moment may be absolute-valued to |μ| and subjected to a linear transformation, and the transformation result is then multiplied by a second-order moment nonlinear transformation value (here For example, other sigma-dependent designs may alternatively be designed ² A function form which becomes larger and becomes smaller monotonically or a form which is transformed by a network structure), and then a calculation process of a nonlinear activation function is carried out to obtain a calculation result as an output of a hidden layer. Wherein the activation function may be a ReLu activation function and/or a softplus activation function and/or a softmax activation function, implemented as a softplus activation function, for example, in the form of f (x) =ln (1+e ^x )。

In a possible implementation manner, in step S104, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in multiple manners in the neural network, and output corresponding target bit soft information, where after obtaining the target statistical feature of the signal to be detected, after obtaining the first bit soft information corresponding to the signal to be detected through mapping processing in the neural network, correction processing is performed on the first bit soft information corresponding to the signal to be detected according to the first value range, and the target bit soft information of the signal to be detected is output.

Taking fig. 4-2 as an example, the process of correcting the first bit soft information corresponding to the signal to be detected according to the first value range may specifically include: determining a first different value range according to different input target statistical characteristics, for example, the first value range can be (a, 0), or (0, b), or (c, d), wherein a is smaller than 0, b is larger than 0, and d is larger than c; further, the corresponding output symbol adjustment (i.e. the corresponding number can be taken) is performed according to the first value range, or the correction is performed to limit the maximum and minimum value of the value range of the bit soft information. Illustratively, the first moment μmay be input _r/i For the (positive or negative) sign of the output low order bit result (i.e., eta in FIG. 4-2 ₀ ) Correction is made, i.e. if μ is greater than 0, η ₀ The result of (2) takes the inverse output.

In a possible implementation manner, in step S104, the signal detection device may perform at least one processing procedure on the target statistical feature of the signal to be detected in multiple manners in the neural network, and output corresponding target bit soft information, where after the target statistical feature of the signal to be detected is obtained, a second statistical feature for indicating the statistical information of the second bit soft information corresponding to the signal to be detected is obtained through fusion processing in the neural network, and after the second statistical feature is mapped to obtain the second bit soft information, correction processing is performed on the value of the second bit soft information according to a second value range, and the target bit soft information of the signal to be detected is output.

In this embodiment, the mapping process, the fusion process and the correction process may refer to the foregoing descriptions about the contents shown in fig. 4-2, and will not be repeated here.

In one possible implementation manner, in the processing process of the neural network, the weight coefficient is obtained by training the preset signal after the data label is generated by the preset algorithm, and compared with the processing process of using the actually transmitted real data as the label in the neural network, the design of the loss function can be simplified, the processing result of the neural network is optimized, and the optimal performance of the preset algorithm is approximated. Specifically, the preset algorithm may include a maximum likelihood detection MLD algorithm, or a linear minimum mean square error LMMSE algorithm, or other algorithm implementation, which is not limited herein.

Preferably, the weight coefficient of the neural network may be obtained by training a preset signal after the data label is generated for the neural network through a maximum likelihood detection MLD algorithm. Specifically, in the process of signal detection processing, under the assumption of Gaussian noise and the assumption of ideal channel information, MLD is an optimal detection algorithm, so that a weight coefficient obtained by training a preset signal after a data tag is generated through the MLD algorithm is applied to a neural network, and the subsequent processing process of the neural network can be optimized.

In this embodiment, after receiving a received signal, the signal detection apparatus determines channel state information of the received signal, where the received signal includes a signal to be detected; determining a filtering parameter of the received signal according to the channel state information; determining a target statistical characteristic of the signal to be detected according to the received signal and the filtering parameter of the received signal, wherein the target statistical characteristic is used for indicating statistical information of posterior probability of the signal to be detected; and processing the target statistical characteristics at least once by adopting a neural network, and outputting target bit soft information corresponding to the signal to be detected. The signal detection device determines the statistical characteristics of the signal to be detected according to the filtering parameters of the received signal, and outputs target bit soft information corresponding to the signal to be detected after processing the target statistical characteristics at least once by using a neural network, namely, the neural network is used for processing the statistical characteristics of the signal to be detected and outputting the corresponding bit soft information.

Fig. 4-3 are schematic diagrams of a signal detection method according to an embodiment of the present application, and as shown in fig. 4-3, the signal detection method includes the following steps.

S1, carrying out linear processing on a received signal according to channel information and noise covariance information;

in this embodiment, the signal detection device obtains a received signal, and may obtain channel information H (i.e., channel state information) through the processing procedure of equation (2), and perform linear processing on the received signal y according to the noise covariance information I, so as to obtain a linear processing result.

In the linear processing in step S1, the implementation of the mode a, the mode B, the mode C and the related steps in the foregoing step S101 may be referred to, and the details are not repeated here.

S2, randomly sampling according to the linear processing result and the equivalent covariance matrix;

in this embodiment, the signal detection device performs random sampling according to the result after the linear processing and the equivalent covariance matrix, so as to obtain a random sampling result.

The random sampling process in step S2 may refer to the sampling process in step S103, that is, the implementation processes of steps D1 to D3, and steps E1 to E4 and related steps, which are not described herein.

S3, calculating moment characteristic values corresponding to each sampling dimension according to the sampling result;

In this embodiment, the signal detection device calculates a moment feature value corresponding to each sampling dimension according to the sampling result.

The process of calculating the moment feature value in step S3 may refer to the implementation processes of steps F1 to F3 and related steps in step S103, which are not described herein.

S4, mapping the moment characteristics into final bit LLR values and outputting the final bit LLR values;

in this embodiment, each moment feature pair obtained in step S3 is mapped to a corresponding bit LLR value output in the neural network in turn according to the modulation mode of the received signal, and the mapping method is just to pass through a given neural network.

The mapping process in the neural network (e.g., as shown in fig. 4-2) in step S4 may refer to the implementation process in step S104, which is not described herein.

In the embodiment, in the processing part, the equivalence only needs to carry out single Cholesky decomposition, and the scheme based on multiple QR decomposition obviously has lower complexity, in addition, in the LLR mapping part, the moment characteristics are directly mapped into LLR values by using a neural network, and the comparison operation in the traditional max-log algorithm is replaced by more general and efficient multiply-add operation; further, the performance of the calculation process approaches to the optimal MLD performance, and the linear processing method with lower complexity has obvious performance gain in the actual scene.

Fig. 5 is a schematic diagram of a signal detection method according to an embodiment of the present application, as shown in fig. 4-3, the signal detection method includes the following steps.

S201, determining channel state information of a received signal;

The implementation process in step S201 may be implemented by referring to the related steps in step S101, which is not described herein.

S202, determining a filtering parameter of the received signal according to the channel state information;

in this embodiment, the signal detection device may further determine the filtering parameters of the received signal according to the channel state information determined in step S201.

The implementation process in step S202 may be implemented by referring to the related steps in step S102, which is not described herein.

S203, randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result;

in this embodiment, the signal detection device randomly samples the received signal according to the filtering parameters of the received signal obtained in step S202, to obtain a sampling result. Wherein the sampling result is used for indicating discrete information of posterior probability of the signal to be detected.

In one possible implementation manner, the process of randomly sampling the received signal according to the filtering parameter of the received signal to obtain the sampling result may specifically include: and randomly sampling the received signal according to the filtering parameters of the received signal and a target sampling structure to obtain a sampling result, wherein the target sampling structure comprises a serial sampling structure and/or a parallel sampling structure. That is, the signal detection device may randomly sample the received signal through a serial sampling structure and/or a parallel sampling structure to obtain a sampling result.

The sampling result obtained in the random sampling process in step S203 may be implemented by various algorithms, and reference may be made to the foregoing sampling process in step S103, that is, the implementation process of calculating the sampling result (or sampling path result) in steps D1 to D3, and steps E1 to E4 and related steps thereof, which are not described herein.

S204, obtaining the target bit soft information corresponding to the signal to be detected according to the sampling result.

In this embodiment, the signal detection device further calculates, according to the sampling result obtained in step S203, the target soft bit information corresponding to the signal to be detected.

In one possible implementation manner, the process of obtaining the target soft bit information corresponding to the signal to be detected according to the sampling result may be implemented by the foregoing manner of formula (4). Illustratively, taking the sample result calculated by steps E1 to E4 as an example, the sample result may be expressed as:

Further, considering the Gaussian additive noise effect, in order to solve the formula (1), the bit soft information can be obtained by approximate calculation by adopting a max-log algorithm, namely, the bit-by-bit slave setThe smallest metric value is chosen as output, denoted +.>Namely by the formula (4):

then, the results of taking 0 and 1 for each bit are subtracted to obtain the final bit Log Likelihood Ratio (LLR), namely the target bit soft information corresponding to the signal to be detected is obtainedAnd/or +.>

In a possible implementation manner, in step S204, the implementation process of calculating the target bit soft information corresponding to the signal to be detected may refer to the implementation of the target bit soft information through the neural network in step S104, which is not described herein.

In this embodiment, after receiving a received signal, the signal detection apparatus determines channel state information of the received signal, where the received signal includes a signal to be detected; determining a filtering parameter of the received signal according to the channel state information; randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result, wherein the sampling result is used for indicating the discrete information of the posterior probability of the signal to be detected; and obtaining the target bit soft information corresponding to the signal to be detected according to the sampling result. The signal detection device obtains a sampling result of discrete information for indicating posterior probability of the signal to be detected according to a mode of randomly sampling the received signal according to filtering parameters of the received signal, and compared with an implementation process of participating in calculation through multiple QR decomposition, the complexity of the calculation process can be reduced through the mode of randomly sampling, the signal processing process of the signal detection device is optimized, and the communication efficiency is improved.

The embodiments of the present application are described above in terms of methods, and the signal detection device in the embodiments of the present application is described below in terms of implementation of a specific device.

Referring to fig. 6, an embodiment of the present application provides a schematic diagram of a signal detection apparatus 600, where the signal detection apparatus 600 at least includes a determining unit 601 and a processing unit 602.

In one implementation of the signal detection apparatus 600,

the determining unit 601 is configured to determine channel state information of a received signal, where the received signal includes a signal to be detected;

the determining unit 601 is further configured to determine a filtering parameter of the received signal according to the channel state information;

the determining unit 601 is further configured to determine a target statistical feature of the signal to be detected according to the received signal and a filtering parameter of the received signal, where the target statistical feature is used to indicate statistical information of a posterior probability of the signal to be detected;

the processing unit 602 is further configured to output target bit soft information corresponding to the signal to be detected after processing the target statistical feature at least once by using a neural network.

In one possible implementation, the processing unit 602 is specifically configured to:

In one possible implementation, the target activation function of the neural network includes a ReLu activation function and/or a softplus activation function and/or a softmax activation function.

In one possible implementation manner, the weight coefficient of the neural network is obtained by training a preset signal after the neural network generates a data tag through a preset algorithm.

In one possible implementation, the preset algorithm includes a maximum likelihood detection MLD algorithm, or a linear minimum mean square error LMMSE algorithm.

In a possible implementation manner, the determining unit 601 is specifically configured to:

the filtering parameters of the received signal are determined in a target manner based on the channel state information, the target manner including Cholesky decomposition processing, QR decomposition processing, or linear detection processing.

and/or the number of the groups of groups,

It should be noted that, for details of the information execution process of the unit of the signal detection apparatus 600, reference may be made to the description in the foregoing method embodiments of the present application, and details are not repeated here.

In one implementation of the signal detection apparatus 600,

the determining unit 601 is further configured to randomly sample the received signal according to a filtering parameter of the received signal, to obtain a sampling result, where the sampling result is used to indicate discrete information of a posterior probability of the signal to be detected;

the processing unit 602 is further configured to obtain target bit soft information corresponding to the signal to be detected according to the sampling result.

and/or the number of the groups of groups,

Referring to fig. 7, a schematic structural diagram of a communication device according to the foregoing embodiment provided in an embodiment of the present application, where the communication device may specifically be a network device in the foregoing embodiment, and the structure of the communication device may refer to the structure shown in fig. 7.

The communication device includes at least one processor 711, at least one memory 712, at least one transceiver 713, at least one network interface 714, and one or more antennas 715. The processor 711, memory 712, transceiver 713, and network interface 714 are coupled, for example, via a bus, which in the present embodiment may include various interfaces, transmission lines, buses, etc., which are not limited in this embodiment. An antenna 715 is coupled to the transceiver 713. The network interface 714 is used to enable the communication apparatus to connect with other communication devices via a communication link, e.g., the network interface 714 may comprise a network interface, e.g., an S1 interface, between the communication apparatus and a core network device, and the network interface may comprise a network interface, e.g., an X2 or Xn interface, between the communication apparatus and other network devices, e.g., other access network devices or core network devices.

The processor 711 is mainly used for processing communication protocols and communication data, and controlling the entire communication apparatus, executing software programs, processing data of the software programs, for example, for supporting the communication apparatus to perform the actions described in the embodiments. The communication device may include a baseband processor, which is mainly used for processing the communication protocol and the communication data, and a central processor, which is mainly used for controlling the entire network device, executing the software program, and processing the data of the software program. The processor 711 in fig. 7 may integrate the functions of a baseband processor and a central processor, and those skilled in the art will appreciate that the baseband processor and the central processor may also be separate processors, interconnected by bus technology, etc. Those skilled in the art will appreciate that the network device may include multiple baseband processors to accommodate different network formats, and that the network device may include multiple central processors to enhance its processing capabilities, and that the various components of the network device may be connected by various buses. The baseband processor may also be expressed as a baseband processing circuit or a baseband processing chip. The central processing unit may also be expressed as a central processing circuit or a central processing chip. The function of processing the communication protocol and the communication data may be built in the processor, or may be stored in a memory in the form of a software program, which is executed by the processor to realize the baseband processing function.

The memory is mainly used for storing software programs and data. The memory 712 may be separate and coupled to the processor 711. Alternatively, the memory 712 may be integrated with the processor 711, for example, within a single chip. The memory 712 is capable of storing program codes for implementing the technical solutions of the embodiments of the present application, and the processor 711 controls the execution of the program codes, and various types of computer program codes that are executed may also be regarded as drivers of the processor 711.

Fig. 7 shows only one memory and one processor. In an actual network device, there may be multiple processors and multiple memories. The memory may also be referred to as a storage medium or storage device, etc. The memory may be a memory element on the same chip as the processor, i.e., an on-chip memory element, or a separate memory element, as embodiments of the present application are not limited in this regard.

The transceiver 713 may be used to support the reception or transmission of radio frequency signals between the communication device and the terminal, and the transceiver 713 may be connected to an antenna 715. The transceiver 713 includes a transmitter Tx and a receiver Rx. Specifically, the one or more antennas 715 may receive radio frequency signals, and the receiver Rx of the transceiver 713 is configured to receive the radio frequency signals from the antennas, convert the radio frequency signals into digital baseband signals or digital intermediate frequency signals, and provide the digital baseband signals or digital intermediate frequency signals to the processor 711, so that the processor 711 performs further processing, such as demodulation processing and decoding processing, on the digital baseband signals or digital intermediate frequency signals. The transmitter Tx in the transceiver 713 is also operative to receive and convert modulated digital baseband signals or digital intermediate frequency signals from the processor 711 to radio frequency signals and transmit the radio frequency signals via the one or more antennas 715. In particular, the receiver Rx may selectively perform one or more steps of down-mixing and analog-to-digital conversion on the radio frequency signal to obtain a digital baseband signal or a digital intermediate frequency signal, where the order of the down-mixing and analog-to-digital conversion is adjustable. The transmitter Tx may selectively perform one or more stages of up-mixing processing and digital-to-analog conversion processing on the modulated digital baseband signal or the digital intermediate frequency signal to obtain a radio frequency signal, and the sequence of the up-mixing processing and the digital-to-analog conversion processing may be adjustable. The digital baseband signal and the digital intermediate frequency signal may be collectively referred to as a digital signal.

The transceiver may also be referred to as a transceiver unit, transceiver device, etc. Alternatively, the device for implementing the receiving function in the transceiver unit may be regarded as a receiving unit, and the device for implementing the transmitting function in the transceiver unit may be regarded as a transmitting unit, that is, the transceiver unit includes a receiving unit and a transmitting unit, where the receiving unit may also be referred to as a receiver, an input port, a receiving circuit, etc., and the transmitting unit may be referred to as a transmitter, or a transmitting circuit, etc.

It should be noted that, the communication apparatus shown in fig. 7 may be specifically used to implement steps implemented by the network device in the embodiment of the method corresponding to fig. 3 to 5, which is not described herein in detail.

Embodiments of the present application also provide a computer-readable storage medium storing one or more computer-executable instructions, where when the computer-executable instructions are executed by a processor, the processor performs a method as described in a possible implementation of the communication apparatus in the foregoing embodiment, where the communication apparatus may specifically be a communication apparatus in the foregoing embodiment.

Embodiments of the present application also provide a computer program product (or called a computer program) storing one or more computers, which when executed by the processor performs a method of possible implementation of the communication apparatus, where the communication apparatus may specifically be a communication apparatus in the foregoing embodiments.

The embodiment of the application also provides a chip system, which comprises a processor and is used for supporting the communication device to realize the functions involved in the possible realization mode of the communication device. In one possible design, the system-on-chip may further include a memory to hold the necessary program instructions and data for the communication device. The chip system may be formed by a chip, or may include a chip and other discrete devices, where the communication device may specifically be a signal detection device in the foregoing embodiment.

The embodiment of the application also provides a network system architecture, which includes the communication device, and the communication device may specifically be the signal detection device in any one of the foregoing embodiments.

In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

Claims

1. A method of signal detection, the method comprising:

determining channel state information of a received signal, wherein the received signal comprises a signal to be detected;

determining a filtering parameter of the received signal according to the channel state information;

determining a target statistical characteristic of the signal to be detected according to the received signal and the filtering parameter of the received signal, wherein the target statistical characteristic is used for indicating statistical information of posterior probability of the signal to be detected;

and processing the target statistical characteristics at least once by adopting a neural network, and outputting target bit soft information corresponding to the signal to be detected.

2. The method of claim 1, wherein the outputting the target bit soft information corresponding to the signal to be detected after the at least one processing of the target statistical feature by using the neural network comprises:

3. The method of claim 1, wherein the outputting the target bit soft information corresponding to the signal to be detected after the at least one processing of the target statistical feature by using the neural network comprises:

and mapping the first statistical characteristic and outputting target bit soft information corresponding to the signal to be detected.

4. The method of claim 1, wherein the outputting the target bit soft information corresponding to the signal to be detected after the at least one processing of the target statistical feature by using the neural network comprises:

5. The method of claim 1, wherein the outputting the target bit soft information corresponding to the signal to be detected after the at least one processing of the target statistical feature by using the neural network comprises:

in the neural network, the target statistical features are fused to obtain second statistical features, wherein the second statistical features are used for indicating statistical information of second bit soft information corresponding to the signal to be detected;

and correcting the value of the second bit soft information corresponding to the signal to be detected according to a second value range, and outputting the target bit soft information corresponding to the signal to be detected.

6. The method according to any one of claim 1 to 5, wherein,

the target activation function of the neural network comprises a ReLu activation function and/or a softplus activation function and/or a softmax activation function.

7. The method according to any one of claim 1 to 5, wherein,

the weight coefficient of the neural network is obtained by training a preset signal after the neural network generates a data tag through a preset algorithm.

8. The method of claim 7, wherein the predetermined algorithm comprises a maximum likelihood detection MLD algorithm or a linear minimum mean square error LMMSE algorithm.

9. The method according to any one of claims 1 to 5, wherein said determining filtering parameters of the received signal from the channel state information comprises:

and determining filtering parameters of the received signal in a target mode according to the channel state information, wherein the target mode comprises Cholesky decomposition processing, QR decomposition processing, SQR decomposition processing or linear detection processing.

10. The method according to any one of claims 1 to 5, wherein said determining filtering parameters of the received signal from the channel state information comprises:

performing normalization processing on the channel state information according to a first preset parameter to obtain processed channel state information;

performing normalization processing on the received signals according to a second preset parameter to obtain first received signals;

and/or the number of the groups of groups,

the determining the target statistical feature of the signal to be detected according to the received signal and the filtering parameter of the received signal comprises:

performing normalization processing on the received signals according to a third preset parameter to obtain second received signals;

performing normalization processing on the filtering parameters of the received signals according to fourth preset parameters to obtain first filtering parameters;

11. The method according to any one of claims 1 to 5, wherein said determining the target statistical characteristics of the signal to be detected from the received signal and the filtering parameters of the received signal comprises:

12. The method of claim 11, wherein randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result comprises:

and randomly sampling the received signal according to the filtering parameters of the received signal and a target sampling structure to obtain the sampling result, wherein the target sampling structure comprises a serial sampling structure and/or a parallel sampling structure.

13. A method of signal detection, the method comprising:

And obtaining target bit soft information corresponding to the signal to be detected according to the sampling result.

14. The method of claim 13, wherein said determining filter parameters of said received signal based on said channel state information comprises:

15. The method according to claim 13 or 14, wherein said determining filter parameters of the received signal from the channel state information comprises:

and/or the number of the groups of groups,

16. The method according to claim 13 or 14, wherein randomly sampling the received signal according to the filtering parameters of the received signal to obtain a sampling result comprises:

17. A signal detection apparatus, characterized in that the apparatus comprises a determination unit and a processing unit:

the determining unit is configured to determine channel state information of a received signal, where the received signal includes a signal to be detected;

the determining unit is further configured to determine a target statistical feature of the signal to be detected according to the received signal and a filtering parameter of the received signal, where the target statistical feature is used to indicate statistical information of a posterior probability of the signal to be detected;

And the processing unit is used for processing the target statistical characteristics at least once by adopting a neural network and then outputting target bit soft information corresponding to the signal to be detected.

18. The apparatus according to claim 17, wherein the processing unit is specifically configured to:

19. The apparatus according to claim 17, wherein the processing unit is specifically configured to:

20. The apparatus according to claim 17, wherein the processing unit is specifically configured to:

21. The apparatus according to claim 17, wherein the processing unit is specifically configured to:

22. The device according to any one of claims 17 to 21, wherein,

23. The device according to any one of claims 17 to 21, wherein,

24. The apparatus of claim 23, wherein the predetermined algorithm comprises a maximum likelihood detection, MLD, algorithm or a linear minimum mean square error, LMMSE, algorithm.

25. The apparatus according to any one of claims 17 to 21, wherein the determining unit is specifically configured to:

26. The apparatus according to any one of claims 17 to 21, wherein the determining unit is specifically configured to:

and/or the number of the groups of groups,

27. The apparatus according to any one of claims 17 to 21, wherein the determining unit is specifically configured to:

28. The apparatus according to claim 27, wherein the determining unit is specifically configured to:

29. A signal detection apparatus, characterized in that the apparatus comprises a determination unit and a processing unit:

the determining unit is further configured to randomly sample the received signal according to a filtering parameter of the received signal, so as to obtain a sampling result, where the sampling result is used to indicate discrete information of posterior probability of the signal to be detected;

30. The apparatus according to claim 29, wherein the determining unit is specifically configured to:

31. The apparatus according to claim 29 or 30, wherein the determining unit is specifically configured to:

and/or the number of the groups of groups,

32. The apparatus according to claim 29 or 30, wherein the determining unit is specifically configured to:

33. A communication device comprising at least one processor and interface circuitry, wherein,

the interface circuit is used for providing programs or instructions for the at least one processor;

the at least one processor is configured to execute the program or instructions to cause the communication device to implement the method of any one of claims 1 to 12 or to cause the communication device to implement the method of any one of claims 13 to 16.

34. A computer readable storage medium having instructions stored thereon which, when executed by a computer, implement the method of any of claims 1 to 12 or the method of any of claims 13 to 16.

35. A communication system comprising the signal processing apparatus of any one of claims 17 to 28, or the communication system comprising the signal processing apparatus of any one of claims 29 to 32, or the communication system comprising the communication apparatus of claim 33.