CN117938680A

CN117938680A - Communication signal type identification method, device, computer equipment and storage medium

Info

Publication number: CN117938680A
Application number: CN202311718292.8A
Authority: CN
Inventors: 李雨晴; 刘鹏; 齐望东; 尤肖虎; 黄永明; 刘升恒
Original assignee: Network Communication and Security Zijinshan Laboratory
Current assignee: Network Communication and Security Zijinshan Laboratory
Priority date: 2023-12-13
Filing date: 2023-12-13
Publication date: 2024-04-26

Abstract

The present application relates to a method, an apparatus, a computer device, a storage medium and a computer program product for type identification of a communication signal. The method comprises the following steps: acquiring a plurality of channel state information corresponding to communication signals of a base station at a target sampling point by a terminal device, wherein each channel state information corresponds to one antenna of the base station; according to each channel state information, initial channel impulse response information corresponding to each channel state information is obtained; obtaining a channel map of a target sampling point according to initial channel impulse response information corresponding to each piece of channel state information; the method comprises the steps of inputting a channel spectrum of a target sampling point into a type recognition model, and obtaining a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of a communication signal of the target sampling point, and the type recognition model is generated after training by taking historical channel spectrums and type labels of different sampling points as sample sets. By adopting the method, the efficiency and the instantaneity of the identification result can be improved.

Description

Communication signal type identification method, device, computer equipment and storage medium

Technical Field

The present application relates to the field of communications technologies, and in particular, to a method, an apparatus, a computer device, a storage medium, and a computer program product for identifying a type of a communication signal.

Background

In real environments, such as urban environments, in buildings, and underground tunnels, etc., when encountering various obstacles such as walls, human bodies, and doors, the obstacles may block and reflect wireless signals, thereby generating Non-Line of Sight (NLOS). Signal propagation under NLOS propagation experiences additional time and distance compared to Line of Sight (LOS), creating signal angle of arrival bias, and also causing additional power LOSs. NLOS propagation problems, such as positioning, direction finding, ranging, inter-base station clock calibration, etc., exist in many application scenarios. When NLOS propagation exists between a terminal device, such as a User Equipment (UE), and a base station, if measurement information of signal NLOS propagation is directly used without any preprocessing of the signals, positioning or calibration accuracy will be significantly degraded. Thus, for wireless communication system applications that rely on LOS propagation, accurately identifying NLOS propagation will play a critical role in system performance.

In the traditional technology, non-line-of-sight propagation and line-of-sight propagation type identification based on artificial intelligence are often carried out by extracting statistical characteristic values of channel state information and inputting a machine learning model for identification, however, a spectrogram input type-based method has high calculation complexity, so that signal identification efficiency is low, real-time requirements of an actual system are difficult to meet, and the method is difficult to use in an actual deployment environment.

Disclosure of Invention

In view of the foregoing, it is desirable to provide a type recognition method, apparatus, computer device, computer-readable storage medium, and computer program product of a communication signal capable of improving signal recognition efficiency.

In a first aspect, the present application provides a method for identifying a type of a communication signal, including:

Acquiring a plurality of channel state information corresponding to communication signals of a base station at a target sampling point of a terminal device, wherein each channel state information corresponds to one antenna of the base station;

obtaining initial channel impulse response information corresponding to each channel state information according to each channel state information;

Obtaining a channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information;

Inputting the channel spectrum of the target sampling point into a type recognition model, and acquiring a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of the communication signal of the target sampling point, and the type recognition model is generated after training by taking historical channel spectrum and type labels of different sampling points as a sample set.

In one embodiment, the obtaining the channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information includes:

acquiring amplitude information of impulse response information of each initial channel;

Performing time delay normalization and amplitude normalization processing on the amplitude information of each initial channel impulse response information to obtain normalized initial channel impulse response information;

denoising the initial channel impulse response information after the normalization processing to obtain a plurality of target channel impulse response information;

And splicing the target channel impulse response information to obtain the channel map of the target sampling point.

In one embodiment, the performing delay normalization and amplitude normalization processing on the amplitude information of each initial channel impulse response information to obtain normalized initial channel impulse response information includes:

Determining the maximum amplitude value of each initial channel impulse response message and the sampling point corresponding to the maximum amplitude value of each initial channel impulse response message according to the amplitude information of each initial channel impulse response message;

circularly shifting the sampling point corresponding to the maximum amplitude value of each initial channel impulse response message to a first sampling point according to the sampling point corresponding to the maximum amplitude value of each initial channel impulse response message so as to normalize the time delay of each initial channel impulse response message;

Dividing the amplitude information of each initial channel impulse response information by the maximum amplitude value to normalize the amplitude of each initial channel impulse response information.

In one embodiment, the splicing the multiple target channel impulse response information to obtain the channel spectrum of the target sampling point includes:

acquiring a transmission channel corresponding to each target channel impulse response message; the transmission channel is used for transmitting communication signals between the terminal equipment and any antenna of the base station at the target sampling point;

And splicing the target channel impulse response information corresponding to the target sampling point by taking the transmission channel as a dimension to obtain a channel map corresponding to the target sampling point.

In one embodiment, inputting the channel spectrum of the target sampling point into the type recognition model, and obtaining the recognition result output by the type recognition model includes:

Extracting characteristic values from an input channel map of a target sampling point to obtain a characteristic matrix;

Leveling the feature matrix to obtain a one-dimensional feature vector of the channel map;

and determining the recognition result according to the preset weight value, the bias matrix and the one-dimensional feature vector.

In one embodiment, before the inputting the channel spectrum of the target sampling point into the type recognition model and obtaining the recognition result output by the type recognition model, the method further includes:

Acquiring historical communication signals of different sampling points and type labels corresponding to the different sampling points respectively, wherein the type labels are used for indicating that the propagation type of the historical communication signals is line-of-sight propagation or non-line-of-sight propagation;

performing channel estimation on the historical communication signals of the different sampling points to generate a historical channel state information set;

Generating historical channel maps of different sampling points according to the historical channel state information set;

And training the type recognition model by taking the historical channel atlas of the different sampling points and the type labels corresponding to the different sampling points as sample sets.

In a second aspect, the present application also provides a type identification device for a communication signal, including:

the first acquisition module is used for acquiring a plurality of channel state information corresponding to communication signals of the base station at a target sampling point of the terminal equipment, wherein each channel state information corresponds to one antenna of the base station;

The conversion module is used for obtaining initial channel impulse response information corresponding to each piece of channel state information according to each piece of channel state information;

the channel map generating module is used for obtaining the channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information;

the recognition result generation module is used for inputting the channel spectrum of the target sampling point into a type recognition model and obtaining a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of the communication signal of the target sampling point, and the type recognition model is generated after training by taking the historical channel spectrum and the type label of different sampling points as a sample set.

In one embodiment, the type-recognition model comprises a lightweight residual network model, the lightweight residual network comprising a residual module, a flat layer, and a full-connection layer, the residual module and the full-connection layer being connected by the flat layer;

The residual error module is used for extracting the characteristic value of the channel map to obtain a characteristic matrix, and inputting the characteristic matrix into the flat layer;

The flattening layer is used for flattening the feature matrix to obtain a one-dimensional feature vector of the channel map, and inputting the one-dimensional feature vector into the full-connection layer;

The full connection layer is used for determining the identification result according to a preset weight value, a preset bias matrix and the one-dimensional feature vector.

In a third aspect, the present application also provides a computer device comprising a memory storing a computer program and a processor implementing the steps of the method described above when the processor executes the computer program.

In a fourth aspect, the present application also provides a computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of the method described above.

In a fifth aspect, the application also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the method described above.

The method, the device, the computer equipment, the storage medium and the computer program product for identifying the type of the communication signal comprise the steps of firstly obtaining a plurality of channel state information of the terminal equipment corresponding to the communication signal of the base station at a target sampling point, wherein each channel state information corresponds to one antenna of the base station; secondly, according to the state information of each channel, initial channel impulse response information corresponding to the state information of each channel is obtained; obtaining a channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information; and finally, inputting the channel spectrum of the target sampling point into a type recognition model, and acquiring a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of the communication signal of the target sampling point, and the type recognition model is generated after training by taking the historical channel spectrum and the type label of different sampling points as a sample set. According to the method, the channel state information is subjected to low-complexity calculation processing, so that the channel spectrum of the target sampling point is obtained, the calculation processing of the channel spectrum of the target sampling point by the type recognition model is reduced by inputting the channel spectrum of the target sampling point into the pre-trained type recognition model, the recognition result output by the type recognition model is obtained, and the efficiency and the instantaneity of the recognition result are improved.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the related art, the drawings that are required to be used in the embodiments or the related technical descriptions will be briefly described, and it is apparent that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to the drawings without inventive effort for those skilled in the art.

FIG. 1 is an application environment diagram of a method of type identification of communication signals in one embodiment;

FIG. 2 is a flow chart of a method of type identification of communication signals in one embodiment;

FIG. 3 is a schematic flow chart of obtaining a channel map of a target sampling point according to initial channel impulse response information corresponding to each channel state information in an embodiment;

fig. 4 is a schematic waveform diagram of initial channel impulse response information tn_cir after normalization and denoising truncation in one embodiment under LOS and NLOS propagation;

FIG. 5 is a schematic diagram of a channel map in one embodiment;

FIG. 6 is a flow chart of a process of performing delay normalization and amplitude normalization on the amplitude information of each initial channel impulse response information in one embodiment;

FIG. 7 is a schematic flow chart of a channel map obtained by splicing a plurality of target channel impulse response information in one embodiment;

FIG. 8 is a schematic flow chart of inputting a channel spectrum of a target sampling point into a type recognition model and obtaining a recognition result output by the type recognition model in one embodiment;

FIG. 9 is a schematic diagram of a lightweight residual network in one embodiment;

FIG. 10 is a schematic diagram of a residual module structure in one embodiment;

FIG. 11 is a schematic flow chart before inputting the channel spectrum of the target sampling point into the type recognition model and obtaining the recognition result output by the type recognition model in another embodiment;

FIG. 12 is a block diagram of a communication signal type recognition device in one embodiment;

fig. 13 is an internal structural view of a computer device in one embodiment.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

The method for identifying the type of the communication signal provided by the embodiment of the application can be applied to an application environment shown in fig. 1. Wherein the terminal device 102 communicates with the base station 104 via a network. The terminal equipment 102 acquires a plurality of channel state information corresponding to communication signals of the base station at a target sampling point, wherein each channel state information corresponds to one antenna of the base station; obtaining initial channel impulse response information corresponding to each channel state information according to each channel state information; obtaining a channel map of a target sampling point according to initial channel impulse response information corresponding to each piece of channel state information; the method comprises the steps of inputting a channel spectrum of a target sampling point into a type recognition model, and obtaining a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of a communication signal of the target sampling point, and the type recognition model is generated after training by taking historical channel spectrums and type labels of different sampling points as sample sets. The terminal device 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, and portable wearable devices, where the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like. The method for identifying the type of the communication signal according to the embodiment of the application can be applied to a device for identifying the type of the communication signal, wherein the device for identifying the type of the communication signal can be the terminal equipment, a base station or other equipment.

In an exemplary embodiment, as shown in fig. 2, a method for identifying a type of a communication signal is provided, and an example in which the method is applied to the terminal device 102 in fig. 1 is described, including the following steps S202 to S208. Wherein:

step S202, a plurality of channel state information corresponding to a communication signal of a base station at a target sampling point of a terminal device is obtained, where each channel state information corresponds to an antenna of the base station.

The base station, i.e., the public mobile communication base station, is one type of radio station, and refers to a radio transceiver station that performs information transfer with a terminal device through a mobile communication switching center in a certain radio coverage area. The main function of the base station is to provide wireless coverage, i.e. to enable wireless signal transmission between a wired communication network and a wireless terminal. The base station transmits and receives messages via an antenna. One base station may correspond to one or more antennas.

An antenna is a transducer that converts a guided wave propagating on a transmission line into an electromagnetic wave propagating in an unbounded medium (usually free space) or vice versa. A component for transmitting or receiving electromagnetic waves in a radio device. Engineering systems such as radio communication, broadcasting, television, radar, navigation, electronic countermeasure, remote sensing, radio astronomy and the like all rely on antennas to work when information is transmitted by electromagnetic waves. In addition, in terms of energy transfer with electromagnetic waves, an antenna is also required for energy radiation other than signals. The common antennas are reversible, i.e. the same pair of antennas can be used as both a transmitting antenna and a receiving antenna. The same antenna is the same as the basic characteristic parameters of transmission or reception.

The channel state information is the channel properties of the communication link and describes the attenuation factors of the signal on each transmission path, such as signal Scattering (Scattering), environmental attenuation (multipath fading or shadowingfading), distance attenuation (power decay of distance), etc. The channel state information can adapt the communication system to the current channel condition, and guarantees are provided for high-reliability and high-speed communication in the multi-antenna system.

Optionally, the terminal device comprises a user device. Acquiring communication signals of terminal equipment at a target sampling point and a base station, wherein the communication signals can be communication signals sent to the base station by user equipment at the target sampling point, and an antenna of the base station receives the communication signals sent by the user equipment at the target sampling point; or the communication signal sent by the base station to the user equipment by the antenna of the base station, and the user equipment receives the communication signal sent by the base station at the target sampling point. The communication signal may be either line-of-sight or non-line-of-sight. The target sampling point where the user terminal is located may be mobile or stationary. And carrying out channel estimation on communication signals between the target sampling point and a plurality of antennas of the base station by the user equipment to obtain a plurality of pieces of channel state information, wherein each piece of channel state information corresponds to one antenna of the base station. The terminal equipment acquires a plurality of channel state information corresponding to the communication signals of the base station at the target sampling point.

Optionally, the terminal device acquires, as the data set, a plurality of channel state information corresponding to the communication signal of the base station at the target sampling point, where each channel state information corresponds to one antenna of the base station. The data set consists of the location of the base station, the unique identification of the base station antenna and a channel estimation matrix of the communication signal between the user devices at the sampling points. Specifically, assuming that the number of antennas in a base station is N _ant, performing channel estimation on antenna receiving signals to obtain Channel State Information (CSI), and constructing a data set Wherein/>Representing the channel estimation matrix of the mth antenna of the base station and the user equipment at the p-th position.

Step S204, according to each channel state information, obtaining the initial channel impulse response information corresponding to each channel state information.

The channel impulse response information refers to signal energy information that the communication signal arrives at the receiving side at different times (different propagation paths lead to different propagation time). I.e. the communication signal is affected during transmission in the channel, reflecting the signal processing characteristics of the channel. It should be noted that, both the channel state information and the channel impulse response information are a representation form of the channel signal information, and the information provided by both are the same.

Optionally, the terminal device performs the inverse fast fourier transform (INVERSE FAST Fourier Transform, IFFT) of formula (1) on each channel state information in the data set, so as to convert the channel state information to obtain initial channel impulse response information corresponding to each channel state information.

CIR(n)＝IFFT(CSI(k)) (1)

Wherein, CIR (n) represents initial channel impulse response information, n is sampling point sequence number; CSI (k) represents channel state information, k being a subcarrier sequence number; the IFFT represents an inverse fast fourier transform.

Step S206, obtaining the channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information.

The channel spectrum is used for indicating channel impulse response information of a sample signal of the terminal equipment between the target sampling point and the base station, and the channel spectrum can be a one-dimensional channel spectrum or a multi-dimensional channel spectrum.

Optionally, the terminal device has a plurality of signal transmission channels between the target sampling point and the base station, and correspondingly, based on the target sampling point where the terminal device is located, the terminal device orders and combines channel impulse response information of sample signals in the plurality of channel transmission channels according to a certain rule, thereby constructing a channel map.

Step S208, inputting the channel spectrum of the target sampling point into a type recognition model, and obtaining a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of the communication signal of the target sampling point, and the type recognition model is generated after training by taking the historical channel spectrum and the type label of different sampling points as a sample set.

The type recognition model is generated after training a sample set based on historical channel maps and type labels of different sampling points. The identification result is used for determining the propagation type of the communication signal of the target sampling point, and the propagation type of the communication signal can be line-of-sight propagation or non-line-of-sight propagation. The type tag indicates a true value tag of a communication signal between a target sampling point where the terminal equipment is located and the base station, and the terminal equipment marks LOS/NLOS true values between different P sampling points and the base station as a tag set Y= { Ind _p }, p=1, 2,3 … P, wherein Ind _p＝(y₁,y₂) for distinguishing LOS from NLOS. Without LOSs of generality, such as Indp = (1, 0) when the terminal device propagates for LOS between a certain sampling point and the base station, indp = (0, 1) when it propagates for NLOS.

Optionally, the terminal device may construct a type recognition model based on a convolutional neural network (convolutionalneural network, CNN), a cyclic neural network (recurrent neural network, RNN), a time-series classification algorithm (connectionist temporal classification, CTC) network, and a machine learning model such as a lightweight neural network model. The terminal equipment constructs a sample set based on historical channel maps and type labels of different sampling points, inputs the sample set into a type recognition model, and trains the type recognition model.

Further, the terminal device inputs the channel spectrum of the target sampling point into a trained type recognition model to obtain the propagation type of the communication signal of the target sampling point, wherein the propagation type can be LOS or NLOS. Wherein the target sampling point may be a base station or a terminal device.

Further, the terminal equipment acquires the real-time CSI based on the base station or the terminal equipment as a sampling point, constructs a channel spectrum, and inputs the channel spectrum into a type identification model deployed on the base station or the terminal equipment so as to perform LOS/NLOS identification on the real-time measured CSI. Class identification model outputAssumption/>Representing LOS probability,/>Representing NLOS probability, discriminating LOS/NLOS according to the method of formula (2):

When the output result is When the LOS probability is larger than the NLOS probability, the signal propagation type is considered to be LOS; when the output result is/>When the LOS probability is smaller than the NLOS probability, the signal propagation type is considered as NLOS.

In the above method for identifying the type of the communication signal, firstly, a plurality of channel state information corresponding to the communication signal of the base station at a target sampling point of the terminal equipment is obtained, and each channel state information corresponds to one antenna of the base station; secondly, according to the state information of each channel, initial channel impulse response information corresponding to the state information of each channel is obtained; obtaining a channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information; and finally, inputting the channel spectrum of the target sampling point into a type recognition model, and acquiring a recognition result output by the type recognition model, wherein the recognition result is used for determining the propagation type of the communication signal of the target sampling point, and the type recognition model is generated after training by taking the historical channel spectrum and the type label of different sampling points as a sample set. According to the method, the channel state information is subjected to low-complexity calculation processing, so that the channel spectrum of the target sampling point is obtained, the calculation processing of the channel spectrum of the target sampling point by the type recognition model is reduced by inputting the channel spectrum of the target sampling point into the pre-trained type recognition model, the recognition result output by the type recognition model is obtained, and the efficiency and the instantaneity of the recognition result are improved.

In an exemplary embodiment, as shown in fig. 3, obtaining a channel map of the target sampling point according to the initial channel impulse response information corresponding to each channel state information includes steps S302 to S306.

Wherein:

Step S302, the amplitude information of each initial channel impulse response information is acquired.

Wherein the amplitude information is information indicating stability of the initial channel impulse response information.

Optionally, the initial channel impulse response information corresponds to a set of sampling points, and the amplitudes of all sampling points in the initial channel impulse response information are taken as the amplitudes of the channel impulse response information. That is, each initial channel impulse response information corresponds to a set of amplitudes, and the number of amplitudes of each initial channel impulse response information is identical to the corresponding number of sampling points. The terminal equipment acquires amplitude information of each initial channel impulse response information. Specifically, the calculation is performed according to formula (3):

CIR^amp(n)＝abs(CIR(n)) (3)

Where abs (·) represents the amplitude, and CIR ^amp (n) represents the amplitude information of the initial channel impulse response information.

Step S304, carrying out time delay normalization and amplitude normalization processing on the amplitude information of each initial channel impulse response information to obtain initial channel impulse response information after normalization processing.

Optionally, in order to eliminate time delay random jitter and eliminate instability of channel impulse response information, the terminal device performs time delay normalization and amplitude normalization processing on the amplitude information of each initial channel impulse response information, so as to obtain initial channel impulse response information after normalization processing.

After normalization processing is carried out on each initial channel impulse response information, the influence of unstable factors is eliminated. In the processing of the initial channel impulse response information, the processing order of performing the time delay normalization and the amplitude normalization processing on the amplitude information of each initial channel impulse response information is not limited.

Step S306, denoising the initial channel impulse response information after the normalization processing to obtain a plurality of target channel impulse response information.

After the initial signal impulse response information after normalization processing is obtained, in order to eliminate the noise part of the channel impulse response information, the number of sampling points of the initial signal impulse response signals is determined according to preset sampling points, the initial channel impulse response information is truncated, and the target channel impulse response information is obtained. It should be noted that, the numerical value of the sampling point needs to be determined according to the actual use environment, and the basic principle is to remove noise sampling points to the greatest extent on the basis of retaining signals, and reduce the input dimension of the subsequent type recognition model while removing noise redundancy.

Optionally, the terminal device performs truncation processing on the initial channel impulse response information after the normalization processing, specifically, the truncation method is as shown in formula (4):

TN_CIR＝CIR^apm_norm(1：N_cut) (4)

N _cut represents the reserved CIR ^apm_norm sampling points, which are a preset fixed value, and the specific value is determined according to the actual use environment; CIR ^amp_norm represents initial channel impulse response information after normalization processing; tn_cir represents truncated normalized channel impulse response, i.e., target channel impulse response information.

According to the sample index (sample index), it is determined that the initial channel impulse response information from the target sampling point after normalization processing and denoising and truncation processing is read, that is, the waveforms of the TN_CIR transmitted by the LOS and the NLOS are shown in fig. 4, it can be seen that the TN_CIR of the LOS and the NLOS have significant differences, the signal energy under the LOS transmission is more concentrated, the signal energy under the NLOS transmission has significant main peaks, and the signal energy under the NLOS transmission is dispersed, and a multimodal state is presented, and the differences can provide key information for the subsequent type identification model.

And step S308, splicing the impulse response information of the plurality of target channels to obtain a channel map of the target sampling points.

Optionally, the terminal device acquires a plurality of target channel impulse response information between a target sampling point where the terminal device is located and a plurality of antennas of the base station, and splices the plurality of target channel impulse response information according to a rule, so as to obtain a channel map of the target sampling point where the terminal device is located.

Further, the terminal device obtains each target channel impulse response information between the target sampling point where the terminal device is located and the antennas 1 and 2.

In this embodiment, the target channel impulse response information is obtained by performing delay normalization, amplitude normalization and denoising on the amplitude information of each initial channel impulse response information, and the target channel impulse response information is spliced into the channel map of the target sampling point. The normalization and denoising processes of low-complexity calculation eliminate time delay and instability among the initial channel impulse response information, the truncation process ensures the effectiveness of the initial channel response information, redundant data of the initial channel response information are eliminated, the target channel impulse response information is spliced into a channel map and is used as a type recognition model input sample, the data volume of the type recognition model for data processing is reduced, and the instantaneity and the efficiency of the type recognition model are improved.

In an exemplary embodiment, as shown in fig. 6, the delay normalization and the amplitude normalization are performed on the amplitude information of each initial channel impulse response information, so as to obtain the normalized initial channel impulse response information, which includes steps S602 to S606. Wherein:

step S602, according to the amplitude information of each initial channel impulse response information, determining the maximum amplitude value of each initial channel impulse response information and the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information.

Optionally, the terminal device determines a maximum amplitude value (max (CIR ^amp)) of each initial channel impulse response information and a sampling point (n _max) corresponding to the maximum amplitude value of each initial channel impulse response information according to the amplitude information of each initial channel impulse response information.

Further, the terminal device determines a maximum amplitude value according to the amplitude information of each initial channel impulse response information, and marks max (CIR ^amp), wherein max (·) represents taking the maximum value.

Further, in order to eliminate the time delay random jitter, the terminal device selects the maximum amplitude from the amplitudes corresponding to the plurality of sampling points of the initial channel impulse response information, and determines the sampling point corresponding to the maximum amplitude as the sampling point corresponding to the maximum amplitude value of the initial channel impulse response information. The process of determining the sampling point n _max corresponding to the maximum amplitude value is characterized by the formula (5):

n_max＝argmax(CIR^amp(n)) (5)

Wherein, CIR ^amp represents the amplitude information of the initial channel impulse response information, argmax (·) represents the sampling point position corresponding to the maximum amplitude value of CIR ^amp.

Step S604, according to the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information, cyclically shifting the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information to the first sampling point so as to normalize the time delay of each initial channel impulse response information.

Optionally, the terminal device circularly shifts the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information to the nth ₁ sampling point according to the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information, so as to eliminate the time delay between the initial channel impulse response information. The operation of the cyclic shift can be characterized by equation (6):

CIR^amp＝cirshift(CIR^amp,-n_max+n₁) (6)

Wherein cirshift (x (n), n ₀) represents that the sequence x (n) is circularly shifted to the right by n ₀ points, n _max represents a sampling point corresponding to the maximum amplitude value, and the parameter n ₁ can be selected according to the practical application environment, so as to ensure the integrity of the main path of the CIR ^amp after the circular shift operation, and is generally greater than 1.

Step S606 divides the amplitude information of each initial channel impulse response information by the maximum amplitude value to normalize the amplitude of each initial channel impulse response information.

Optionally, in order to eliminate instability of the channel impulse response information, after the terminal device obtains the amplitude of each initial channel impulse response information and the maximum amplitude value of each initial channel impulse response information, the terminal device divides the amplitude of each initial channel impulse response information by the maximum amplitude value to obtain an amplitude normalization processing result of each initial channel impulse response. The process of amplitude normalization is characterized according to formula (7):

CIR^amp_norm＝(CIR^amp/max(CIR^amp) (7)

Where max (CIR ^amp) represents the maximum amplitude value of each initial channel impulse response information, CIR ^amp represents the amplitude value of each initial channel impulse response information, and CIR ^amp_norm represents the amplitude normalization result of the initial channel impulse response information.

In an exemplary embodiment, as shown in fig. 7, a plurality of target channel impulse response information is spliced to obtain a channel map of target sampling points, which includes steps S702 to S704. Wherein:

Step S702, a transmission channel corresponding to each target channel impulse response information is obtained; the transmission channel is used for transmitting communication signals between the terminal equipment and any antenna of the base station at the target sampling point.

The transmission channel is used for transmitting communication signals between the terminal equipment and any antenna of the base station at the target sampling point.

Optionally, the terminal device acquires a transmission channel corresponding to each target channel impulse response information, for example, when the terminal device is located at the target sampling point p, the target channel impulse response information corresponds to the 1 st to the m antennas of the base station. Correspondingly, the terminal equipment acquires a target channel impulse response message corresponding to a transmission channel between a target sampling point p where the terminal equipment is located and a1 st antenna; and the target sampling point p where the terminal equipment is positioned and the 2 nd antenna are used as target channel impulse response information corresponding to a transmission channel, and the target sampling point p where the terminal equipment is positioned and the m th antenna are used as target channel impulse response information corresponding to a transmission channel.

And step S704, splicing the target channel impulse response information corresponding to the target sampling points by taking the transmission channel as a dimension to obtain a channel map corresponding to the target sampling points.

Optionally, the terminal device uses the transmission channel as a dimension to splice the target channel impulse response information corresponding to the target sampling points according to the arrangement sequence of the transmission channel, so as to form a one-dimensional spectrogram with N _ant channels, wherein the dimension is (N _ant×1×N_cut), and N _ant represents the number of base station antennas; n _cut represents the number of reserved CIR ^apm_norm sampling points, namely the number of channels in the type identification model, namely the transmission channels of the terminal equipment between the target sampling points and any antenna of the base station are in one-to-one correspondence with the channels in the type identification model, so that a channel map corresponding to the target sampling points is obtained, and the channel map corresponding to the target sampling points is shown in fig. 5.

In this embodiment, by splicing multiple target channel impulse response information corresponding to the target sampling points by taking the transmission channel as a dimension, a one-dimensional channel map corresponding to the target sampling points can be generated, so that time domain information of an initial CIR can be reserved, key information is prevented from being lost, space domain information of multiple antennas is fully utilized, and abundant LOS/NLOS implicit features are provided for subsequent network identification. By inputting the one-dimensional channel spectrum into the type recognition model, the calculation amount of the type recognition model on the one-dimensional channel spectrum is reduced, and therefore the instantaneity and the efficiency of the type recognition model are improved.

In an exemplary embodiment, as shown in fig. 8, a channel map of a target sampling point is input into a type recognition model, and a recognition result output by the type recognition model is obtained, including steps S802 to S806.

Wherein:

Step S802, extracting characteristic values from the input channel spectrum of the target sampling points to obtain a characteristic matrix.

The type recognition model includes a lightweight Residual network model, which includes a Residual module (Residual Block), a flat layer (flat) and a full Connection layer (Fc), and the Residual module and the full Connection layer are connected through the flat layer, as shown in fig. 9. The residual error module is used for extracting the characteristic value of the channel map to obtain a characteristic matrix, and inputting the characteristic matrix into the flat layer; the flattening layer is used for flattening the feature matrix to obtain a one-dimensional feature vector of the channel map, and inputting the one-dimensional feature vector into the full-connection layer; the full connection layer is used for determining the identification results of two propagation types, namely LOS and NLOS, according to a preset weight value, a preset bias matrix and a one-dimensional feature vector; the fully connected layer contains two neurons, the output of the first being the probability of being predicted as LOS and the output of the second being the probability of being predicted as NLOS.

The residual module is shown in fig. 10, and comprises 2 one-dimensional convolution layers, wherein a first convolution layer is followed by a batch normalization layer and a first rectification linear Unit (RECTIFIED LINEAR Unit, reLU); the first rectifying linear unit is followed by an adder and a second rectifying linear unit, wherein the adder is used for correspondingly adding the input data of the 0 th layer and the output data of the 4 th layer of the lightweight residual network model, and the specific layer number is shown in table 1. The first and second rectifying linear units are each used as an activation function for each one-dimensional convolution layer.

Table 1 lightweight residual network layer parameter settings

Optionally, the terminal device inputs the channel spectrum of the target sampling point into the type identification model, and the terminal device performs one-dimensional convolution on the input channel spectrum of the target sampling point through the residual module so as to extract the eigenvalue for distinguishing LOS/NLOS propagation, thereby obtaining the eigenvalue matrix.

Step S804, flattening the feature matrix to obtain a one-dimensional feature vector of the channel map.

Optionally, the terminal device performs leveling operation on the feature matrix through the flat layer, transforms the feature matrix into a one-dimensional feature vector, obtains the one-dimensional feature vector of the channel map, and transmits the one-dimensional feature vector to the full-connection layer of the model.

Step S806, determining the identification result according to the preset weight and the bias matrix and the one-dimensional feature vector.

Optionally, the full connection layer is connected with the flat layer, and the terminal device determines the identification result according to the preset weight and bias matrix between the flat layer and the full connection layer and the one-dimensional feature vector output by the flat layer.

The connection between the planar layer and the fully connected layer is achieved by equation (8):

Y＝"*X+B (8)

Wherein W and B represent weights and bias matrices between the flat layer and the fully connected layer, respectively, X is a one-dimensional feature vector output by the flat layer, Is the full connectivity layer output, i.e. the output of a lightweight network model.

For LOS/NLOS classification problems, cross entropy is used as a LOSs function, as shown in equations (9) and (10).

Where N is the number of samples, yi is the type tag,Is a lightweight residual network output.

Optionally, the terminal device outputs a lightweight residual networkThe cross entropy of the corresponding type label yi is used as a loss function, and whether the loss function value meets the target precision is judged; when the loss function value does not meet the target precision, the weight and the bias matrix in the lightweight residual network are updated for multiple times, and the lightweight residual network outputs/>, after updating the weight and the bias matrix for multiple timesUntil the loss function value meets the target precision.

The weight between the flat layer and the fully connected layer is updated by equation (11):

The bias matrix between the flat layer and the fully connected layer is updated by equation (12):

Where lr is the learning rate, W ^old and W ^new are the old and updated weights, respectively, and B ^old and B ^new are the old and updated biases, respectively.

In each training phase, the network weights and offsets are iteratively updated multiple times until the network loss function value meets the target accuracy.

And under the condition that the target precision is met, determining the signal propagation type according to the light residual error network output. Such as lightweight residual network outputAssumption/>Representing LOS probability,/>Representing NLOS probability, LOS/NLOS is determined according to formula (2):

When the output result is When the LOS probability is larger than the NLOS probability, the signal propagation type is considered to be LOS; when the output result is/>When the LOS probability is smaller than the NLOS probability, the signal propagation type is considered as NLOS. /(I)

In the embodiment, a channel map of a target sampling point is input into a type identification model, and a one-dimensional channel map is processed by using a low-complexity type identification model after an identification result is obtained; and the calculation amount of the type identification model is small, the memory overhead of the system can be reduced, and the efficiency of result output is improved.

In an exemplary embodiment, as shown in fig. 11, before inputting the channel map of the target sampling point into the type recognition model and acquiring the recognition result output by the type recognition model, steps S1102 to S1108 are further included. Wherein:

in step S1102, a history communication signal of different sampling points and type labels corresponding to the different sampling points are obtained, where the type labels are used to indicate that the propagation type of the history communication signal is line-of-sight propagation or non-line-of-sight propagation.

The LOS/NLOS truth values between the terminal device and the base station at different P sampling points are marked as a type tag set y= { Ind _p }, p=1, 2,3 … P, where Ind _p＝(y₁,y₂), which is used for distinguishing LOS from NLOS, such as Indp = (1, 0) when the terminal device propagates LOS between a certain sampling point and the base station, and Indp = (0, 1) when the terminal device propagates NLOS.

Optionally, the terminal device acquires, as the data set, a plurality of historical communication signals corresponding to the communication signals of the base station at different sampling points by the terminal device. The data set consists of the location of the base station, the unique identification of the base station antenna and a channel estimation matrix of the historical communication signal between the user devices at the sampling points.

Further, acquiring historical communication signals of different sampling points where the terminal equipment is located and type labels corresponding to the different sampling points respectively, wherein the type label of the historical communication signals can be LOS or NLOS, and if the type label is LOS, the type label can be recorded as Indp = (1, 0); if the class tag is NLOS, it can be denoted Indp = (0, 1).

Step S1104, performing channel estimation on the historical communication signals of different sampling points, and generating a historical channel state information set.

Optionally, the terminal device performs channel estimation on historical communication signals between different sampling points and the base station antenna, and generates a historical channel state information CSI set.

Step S1106, according to the historical channel state information set, historical channel maps of different sampling points are generated.

Optionally, the terminal device performs inverse fast fourier transform (INVERSE FAST Fourier Transform, IFFT) on each historical channel state information in the historical channel state information set according to the historical channel state information set, so as to convert the historical channel state information to obtain historical initial channel impulse response information corresponding to each historical channel state information; acquiring amplitude information of each historical initial channel impulse response information; performing time delay normalization and amplitude normalization processing on the amplitude information of each historical initial channel impulse response information to obtain normalized historical initial channel impulse response information; denoising the plurality of normalized historical initial channel impulse response information to obtain a plurality of historical target channel impulse response information; and splicing the plurality of historical target channel impulse response information, so as to generate historical channel maps of different sampling points.

Step S1108, training a type recognition model by taking the historical channel atlas of different sampling points and type labels corresponding to different sampling points as sample sets.

Optionally, the terminal device acquires historical channel maps of different sampling points and type labels corresponding to the different sampling points respectively as training sample sets. Obtaining an initial lightweight network model output by inputting a historical channel map into an initial type identification model, namely an initial lightweight network modelJudging whether the loss function value meets the target precision; wherein the loss function is based on the output/>, of the historical channel spectrum light-weight network model of different sampling pointsAnd the type labels respectively corresponding to the different sampling points.

When the loss function value does not meet the target precision, the weight and the bias matrix in the lightweight residual network are updated for multiple times, and the initial lightweight residual network outputs the updated weight and bias matrix for multiple timesAnd until the loss function value meets the target precision, finishing training, and thus obtaining the trained lightweight network model.

In this embodiment, by training the type recognition model with the historical channel maps of different sampling points and the type labels corresponding to different sampling points as sample sets, a trained type recognition model can be obtained, and the accuracy of recognition is improved by using the type recognition model to perform signal type recognition.

The effectiveness of the application is illustrated by two practical experiments. The experiment is based on a fifth generation mobile communication technology (5th Generation Mobile Communication Technology,5G) positioning system of a sub6G frequency band, a 5G Sounding REFERENCE SIGNAL, SRS is adopted as a positioning signal, SRS bandwidth is configured to be 100MHz in the experiment, 1632 subcarriers are occupied, and subcarrier intervals are 60kHz. In the experimental process, 1 5G remote radio unit with 4 array element uniform linear array antenna is used as receiving equipment. In this experiment, the parameter N ₁ in the formula CIR ^amp＝cirshift(CIR^amp,-n_max+n_i) was set to 16, and the parameter N _cut in the formula tn_cir=cir ^apm-norm(1：N_cut) was set to 150. As shown in table 2, the model parameters of the lightweight residual network were 1.314K and the total amount of floating point operations 18K.

TABLE 2

Model parameter quantity	Total number of floating point operations (FLOPs)
		1.314K	18K

In an exemplary embodiment, to verify the validity of the application in a stationary condition of the user equipment UE. In a typical obstacle environment (including pillars, walls, vehicles and the like) of an indoor parking lot, LOS of 310 different sampling points and NLOS acquisition points of 310 different sampling points are set, and 10 frames of data are acquired by each sampling point. The training set was constructed using 3100 LOS and 3100 NLOS data as described above. After one week, LOS of 100 different sampling points and NLOS of 100 different sampling points are randomly set again, each sampling point collects 100 frames of data in total, and the 10000 frames of LOS and 10000 frames of NLOS data are used for constructing a test set. After 15 iterative learning, the recognition accuracy obtained by the light residual error network is shown in table 3, the LOS recognition accuracy reaches 0.93, and the NLOS recognition accuracy reaches 0.90.

TABLE 3 Table 3

LOS identification accuracy	NLOS recognition accuracy
		0.93	0.90

In another exemplary embodiment, the validity of the application is verified under the motion conditions of the user equipment UE. Under a typical obstacle environment (including pillars, walls and the like) of an indoor parking lot, UE does non-uniform motion at a speed of 1-3 m/s, the data acquisition rate is 40 ms/frame, and 17979 frames of LOS data and 18033 frames of NLOS data are acquired to form a training set. LOS data of 11908 frames and NLOS data of 22687 frames were additionally collected under the same conditions as the current day, to construct test set 1. After two months, the distribution of the parking vehicles in the parking lot is obviously changed, and the positioning system is electrified and electrified for multiple times, the UE is enabled to move at a speed of 1-3 m/s again, 11885 frames of LOS data and 21304 frames of NLOS data are collected at a collection rate of 40 ms/frame, and the data form a test set 2. The recognition accuracy of the lightweight residual error network on the test set 1 and the test set 2 after training based on the training set is shown in the table 4, the LOS recognition accuracy of the test set 1 reaches 0.924, and the NLOS recognition accuracy reaches 0.910; the LOS recognition accuracy of the test set 2 reaches 0.911, and the NLOS recognition accuracy reaches 0.917.

TABLE 4 Table 4

	LOS identification accuracy	NLOS recognition accuracy
			Test set 1	0.924	0.910
Test set 2	0.911	0.917

The above recognition accuracy is taken as the recall rate in the classification problem, namely

LOS recognition accuracy = number of correctly recognized LOS/total number of LOS;

NLOS recognition accuracy = number of correctly recognized NLOS/total number of NLOS;

comprehensive experiments prove that in a typical indoor parking lot environment, LOS/NLOS recognition accuracy of more than 90% can be obtained under the condition that a terminal is stationary or in motion; under the conditions that the distribution of the obstacles in the parking lot is changed and the equipment is powered on and powered off, the identification precision is almost unchanged, which shows that the application has stronger generalization on the environment and hardware changes; the model parameter of the light residual error network is only 18K, which is beneficial to saving the memory overhead, the floating point operand is only 1.314K, which is far smaller than the common depth network, and the light residual error network is suitable for a system with higher requirement on real-time performance.

It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides a communication signal type recognition device for realizing the above related communication signal type recognition method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in the embodiments of the type identifying device for one or more communication signals provided below may refer to the limitation of the type identifying method for a communication signal hereinabove, and will not be repeated herein.

In an exemplary embodiment, as shown in fig. 12, there is provided a type recognition apparatus of a communication signal, including: a first acquisition module 1201, a conversion module 1202, a channel map generation module 1203, and a recognition result generation module 1204, wherein:

A first obtaining module 1201 is configured to obtain a plurality of channel state information corresponding to a communication signal of a base station at a target sampling point, where each channel state information corresponds to one antenna of the base station.

The conversion module 1202 is configured to obtain initial channel impulse response information corresponding to each piece of channel state information according to each piece of channel state information.

The channel map generating module 1203 is configured to obtain a channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information.

The recognition result generation module 1204 is configured to input the channel spectrum of the target sampling point into a type recognition model, and obtain a recognition result output by the type recognition model, where the recognition result is used to determine a propagation type of the communication signal of the target sampling point, and the type recognition model is generated after training with historical channel spectrums and type labels of different sampling points as a sample set.

In one exemplary embodiment, the type-recognition model includes a lightweight residual network model, the lightweight residual network including a residual module, a flat module, and a full connection module, the residual module and the full connection module being connected by the flat module.

And the residual error module is used for extracting characteristic values from the input channel spectrum of the target sampling points to obtain a characteristic matrix.

And the flattening module is used for flattening the feature matrix to obtain a one-dimensional feature vector of the channel map.

The flattening module is used for determining the recognition result according to the preset weight value, the bias matrix and the one-dimensional feature vector.

In an exemplary embodiment, the channel map generation module 1203 includes:

and the amplitude information acquisition unit is used for acquiring the amplitude information of each initial channel impulse response information.

And the processing unit is used for carrying out time delay normalization and amplitude normalization processing on the amplitude information of each initial channel impulse response message to obtain normalized initial channel impulse response information.

And the denoising unit is used for denoising the plurality of normalized initial channel impulse response information to obtain a plurality of target channel impulse response information.

And the splicing unit is used for splicing the multiple target channel impulse response information to obtain a channel map of the target sampling point.

In one exemplary embodiment, a processing unit includes:

And the amplitude sampling point determining subunit is used for determining the maximum amplitude value of each initial channel impulse response information and the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information according to the amplitude information of each initial channel impulse response information.

And the time delay normalization subunit is used for circularly shifting the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information to the first sampling point according to the sampling point corresponding to the maximum amplitude value of each initial channel impulse response information so as to normalize the time delay of each initial channel impulse response information.

And the amplitude normalization subunit is used for dividing the amplitude value of each initial channel impulse response information by the maximum amplitude value so as to normalize the amplitude value of each initial channel impulse response information.

In one exemplary embodiment, a splice unit includes:

A transmission channel acquisition subunit, configured to acquire a transmission channel corresponding to each target channel impulse response information; the transmission channel is used for transmitting communication signals between the terminal equipment and any antenna of the base station at the target sampling point.

And the splicing subunit is used for splicing the target channel impulse response information corresponding to the target sampling point by taking the transmission channel as a dimension to obtain a channel map corresponding to the target sampling point.

In an exemplary embodiment, the recognition result generation module 1204 includes:

And the characteristic matrix determining unit is used for extracting characteristic values from the input channel spectrum of the target sampling points to obtain a characteristic matrix.

And the one-dimensional feature vector determining unit is used for carrying out leveling treatment on the feature matrix to obtain a one-dimensional feature vector of the channel map.

The identification result determining unit is used for determining the identification result according to the preset weight and the bias matrix and the one-dimensional feature vector.

The respective modules in the above-described type recognition device for communication signals may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one exemplary embodiment, a computer device is provided, which may be a server, and the internal structure thereof may be as shown in fig. 13. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is for storing type identification data of the communication signals. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of type identification of a communication signal.

It will be appreciated by those skilled in the art that the structure shown in FIG. 13 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

In an embodiment, there is also provided a computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, carries out the steps of the method embodiments described above.

In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magneto-resistive random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (PHASE CHANGE Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in various forms such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.

Claims

1. A method for identifying a type of a communication signal, the method comprising:

2. The method of claim 1, wherein the obtaining the channel map of the target sampling point according to the initial channel impulse response information corresponding to each piece of channel state information includes:

3. The method of claim 2, wherein the performing delay normalization and amplitude normalization on the amplitude information of each initial channel impulse response information to obtain normalized initial channel impulse response information includes:

4. The method of claim 2, wherein the concatenating the plurality of target channel impulse response information to obtain the channel map of the target sampling point comprises:

5. The method according to claim 1, wherein inputting the channel spectrum of the target sampling point into a type recognition model and acquiring the recognition result output by the type recognition model comprises:

extracting characteristic values from the input channel spectrum of the target sampling points to obtain a characteristic matrix;

And determining the identification result according to the preset weight and the bias matrix and the one-dimensional feature vector.

6. The method according to any one of claims 1 to 5, further comprising, before said inputting the channel map of the target sampling point into a type recognition model and acquiring the recognition result output by the type recognition model:

7. A type recognition device for a communication signal, the device comprising:

8. The apparatus of claim 7, wherein the type-recognition model comprises a lightweight residual network model, the lightweight residual network comprising a residual module, a flat module, and a fully connected module, the residual module and the fully connected module connected by the flat module;

The residual error module is used for extracting characteristic values from the input channel spectrum of the target sampling point to obtain a characteristic matrix;

The flattening module is used for flattening the feature matrix to obtain a one-dimensional feature vector of the channel map;

The flattening module is used for determining the identification result according to a preset weight and a preset bias matrix and the one-dimensional feature vector.

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.

10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.

11. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.