Departure angle estimation method and apparatus based on unsupervised learning, and related media

Abstract

The embodiment of the application provides a kind of based on the angle of departure estimation method, device and related medium of unattended learning, belong to wireless communication technical field.The method comprises: obtaining the observation sequence data transmitted by base station;Based on communication protocol, pilot information is obtained, and pilot related information is obtained;When pilot related information is specific pilot information, the angle of departure estimation of observation sequence data and pilot information is carried out by the first angle of departure estimation model pre-trained, and target angle of departure information is obtained;When pilot related information is pilot information pilot statistics information, the angle of departure estimation of observation sequence data and pilot statistics information is carried out by the second angle of departure estimation model pre-trained, and target angle of departure information is obtained.The embodiment of the application can make terminal device accurately estimate the angle of departure information of base station terminal transmitting signal.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, and particularly relates to a method and device for angle of departure estimation based on unsupervised learning and related media. BACKGROUND

[0002] Angle estimation plays a vital role in many fields such as astronomy, navigation, radar sensing, etc. The downlink positioning method refers to transmitting signals to a terminal device by a base station, and the terminal device receives and processes the signals to determine its own position. The angle of departure (AOD) refers to the angle at which the signal departs from the antenna array.

[0003] In the related art, many Internet of Things devices need to implement device positioning through downlink angle estimation. However, most devices in the Internet of Things are only configured with a single antenna, and many related angle estimation methods are mainly designed for multi-antenna array systems. Such systems are composed of multiple antennas, and these angle estimation methods are not applicable to Internet of Things devices with only a single antenna. The antenna resources of Internet of Things devices are scarce, which makes it very difficult for single-antenna devices to estimate the angle of departure of the sending end, and it is difficult to obtain accurate results. Therefore, how to enable single-antenna terminal devices to accurately estimate the angle of departure information of the base station terminal transmitted signal has become a technical problem to be solved. SUMMARY

[0004] The main purpose of the embodiments of the present application is to provide a method and device for angle of departure estimation based on unsupervised learning and related media, which aims to enable terminal devices to accurately estimate the angle of departure information of the base station terminal transmitted signal.

[0005] To achieve the above purpose, a first aspect of the embodiments of the present application provides a method for angle of departure estimation based on unsupervised learning, which comprises:

[0006] Obtaining observation sequence data transmitted by a base station;

[0007] Obtaining pilot information based on a communication protocol to obtain pilot-related information;

[0008] In response to the pilot-related information being specific pilot information, performing angle of departure estimation on the observation sequence data and the pilot information by a pre-trained first angle of departure estimation model to obtain target angle of departure information;

[0009] In response to the pilot-related information being pilot statistical information of the pilot information, performing angle of departure estimation on the observation sequence data and the pilot statistical information by a pre-trained second angle of departure estimation model to obtain target angle of departure information.

[0010] In some embodiments, the observation sequence data includes a first number of observation values;

[0011] The angle of departure estimation on the observation sequence data and the pilot information by the pre-trained first angle of departure estimation model includes:

[0012] Based on the pilot information, a pilot signal matrix is constructed;

[0013] A first model input tensor is generated according to the pilot signal matrix and the first number of observation values;

[0014] The first model input tensor is input into the pre-trained first angle of departure estimation model for angle of departure estimation to obtain target angle of departure information;

[0015] The angle of departure estimation on the observation sequence data and the pilot statistical information by the pre-trained second angle of departure estimation model includes:

[0016] Based on the pilot statistical information, a pilot vector matrix is constructed;

[0017] Based on the pilot vector matrix, a Kronecker product processing is performed to obtain a pilot processing matrix;

[0018] A second model input tensor is generated according to the pilot processing matrix and the first number of observation values;

[0019] The second model input tensor is input into the pre-trained second angle of departure estimation model for angle of departure estimation to obtain target angle of departure information.

[0020] In some embodiments, before the angle of departure estimation on the observation sequence data and the pilot information by the pre-trained first angle of departure estimation model to obtain target angle of departure information, the first angle of departure estimation model is pre-trained, including:

[0021] A first training data set is obtained; the first training data set includes first observation training data and pilot training data;

[0022] The first observation training data and the pilot training data are input into an original first angle of departure estimation model to output first predicted angle of departure information and a first channel gain estimation value;

[0023] Based on a deterministic maximum likelihood estimation, the original first angle of departure estimation model is unsupervised learned according to the first predicted angle of departure information and the first channel gain estimation value to obtain the pre-trained first angle of departure estimation model.

[0024] In some embodiments, the unsupervised learning of the original first angle of departure estimation model according to the first predicted angle of departure information and the first channel gain estimation value to obtain the pre-trained first angle of departure estimation model comprises:

[0025] performing a steering vector calculation based on the first predicted angle of departure information to obtain a first predicted angle steering vector;

[0026] constructing a first model loss function based on the pilot training data, the first predicted angle steering vector, the first channel gain estimation value and the first observation training data:

[0027]

[0028] wherein X1 represents the pilot training data, Y1 represents the first observation training data, D1 represents a first training data set, W1 represents the first model parameter of the original first angle of departure estimation model, ξ1 represents the first channel gain estimation value, θ1 represents the first predicted angle of departure information, and a(θ1) represents the first predicted angle steering vector;

[0029] adjusting the first model parameter of the original first angle of departure estimation model based on the first model loss function to obtain the pre-trained first angle of departure estimation model.

[0030] In some embodiments, before performing the angle of departure estimation on the observation sequence data and the pilot statistical information by the pre-trained second angle of departure estimation model to obtain target angle of departure information, the method further comprises pre-training the second angle of departure estimation model, comprising:

[0031] obtaining a second training data set; the second training data set comprises second observation training data and pilot training statistical data;

[0032] inputting the second observation training data and the pilot training statistical data into an original second angle of departure estimation model to output second predicted angle of departure information, a second channel gain estimation value and noise variance information;

[0033] generating a reconstructed covariance matrix based on the second predicted angle of departure information, the second channel gain estimation value and the noise variance information, and generating a sample covariance matrix based on the second observation training data;

[0034] performing unsupervised learning of the original second angle of departure estimation model according to the reconstructed covariance matrix and the sample covariance matrix based on a stochastic maximum likelihood estimation to obtain the pre-trained second angle of departure estimation model.

[0035] In some embodiments, the generating a reconstructed covariance matrix based on the second predicted angle of departure information, the second channel gain estimation value and the noise variance information, and generating a sample covariance matrix based on the second observed training data, comprises:

[0036] performing a steering vector calculation based on the second predicted angle of departure information to obtain a second predicted angle steering vector;

[0037] generating a reconstructed covariance matrix based on the second predicted angle steering vector, the second channel gain estimation value and the noise variance information:

[0038]

[0039] wherein C y represents the reconstructed covariance matrix, ξ2 represents the second channel gain estimation value, w (q) represents the noise variance information, θ2 represents the second predicted angle of departure information, a(θ2) represents the second predicted angle steering vector, X (q) (X (q) ) H represents a covariance of pilot information in the pilot training statistics, represents an expectation value.

[0040] generating a sample covariance matrix based on the second observed training data:

[0041]

[0042] wherein, represents the sample covariance matrix, Q represents a quantity of the second observed training data, y (q) represents a qth data of the second observed training data, (y (q) H represents a conjugate transpose matrix of y (q) .

[0043] In some embodiments, the unsupervised learning of the original second angle of departure estimation model based on the reconstructed covariance matrix and the sample covariance matrix to obtain the pre-trained second angle of departure estimation model comprises:

[0044] constructing a second model loss function based on the reconstructed covariance matrix and the sample covariance matrix:

[0045]

[0046] wherein C y represents the reconstructed covariance matrix, ​represents the sample covariance matrix, X2represents the pilot training statistics, Y2represents the second observation training data, D2represents the second training data set, W2represents the second model parameter of the original second angle of departure estimation model, tr() refers to the trace of a matrix, that is, the sum of the diagonal elements of the main diagonal of the matrix, det(C y ) represents the determinant value of the reconstructed covariance matrix;

[0047] Adjust the second model parameter of the original second angle of departure estimation model based on the second model loss function to obtain a pre-trained second angle of departure estimation model.

[0048] To achieve the above object, a second aspect of the embodiment of the present application provides an angle of departure estimation device based on unsupervised learning, the device comprising:

[0049] An observation data acquisition module is configured to acquire observation sequence data transmitted by a base station;

[0050] A pilot information acquisition module is configured to acquire pilot information based on a communication protocol to obtain pilot-related information;

[0051] A first angle estimation module is configured to, in response to the pilot-related information being specific pilot information, perform angle of departure estimation on the observation sequence data and the pilot information by using a pre-trained first angle of departure estimation model to obtain target angle of departure information;

[0052] A second angle estimation module is configured to, in response to the pilot-related information being pilot statistical information of the pilot information, perform angle of departure estimation on the observation sequence data and the pilot statistical information by using a pre-trained second angle of departure estimation model to obtain target angle of departure information.

[0053] To achieve the above object, a third aspect of the embodiment of the present application provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the angle of departure estimation method based on unsupervised learning of the first aspect when executing the computer program.

[0054] To achieve the above object, a fourth aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the angle of departure estimation method based on unsupervised learning of the first aspect.

[0055] This application proposes a departure angle estimation method, apparatus, and related medium based on unsupervised learning. It acquires observation sequence data transmitted from a base station and obtains pilot information based on a communication protocol to obtain pilot-related information. When the pilot-related information is specific pilot information, a pre-trained first departure angle estimation model is used to estimate the departure angle of the observation sequence data and pilot information to obtain the target departure angle information. When the pilot-related information is pilot statistics, a pre-trained second departure angle estimation model is used to estimate the departure angle of the observation sequence data and pilot statistics to obtain the target departure angle information. Therefore, to improve the accuracy of departure angle estimation, this application uses not only the received observation sequence data as input to the neural network model but also the pilot information. Based on the acquired pilot-related information, if the pilot-related information is specific pilot information, the first departure angle estimation model is used for departure angle estimation. If the pilot-related information is pilot statistics, the second departure angle estimation model is used for departure angle estimation, enabling the user equipment to perform departure angle estimation according to different states of the pilot information, thus obtaining accurate target departure angle information. Attached Figure Description

[0056] Figure 1 This is a flowchart of the departure angle estimation method based on unsupervised learning provided in the embodiments of this application;

[0057] Figure 2A yes Figure 1 The flowchart of step S103 in the process;

[0058] Figure 2B yes Figure 1 The flowchart of step S104 in the process;

[0059] Figure 3 yes Figure 1 The flowchart preceding step S103;

[0060] Figure 4 yes Figure 3 The flowchart of step S303 in the process;

[0061] Figure 5 yes Figure 3 The flowchart preceding step S104;

[0062] Figure 6 yes Figure 5 The flowchart of step S503 in the process;

[0063] Figure 7 yes Figure 5 The flowchart of step S504 in the process;

[0064] Figure 8is a relationship diagram of average absolute error indicators of different methods provided by embodiments of the present application varying with sampling time;

[0065] Figure 9 is a relationship diagram of average absolute error indicators of different methods provided by embodiments of the present application varying with base station transmission power;

[0066] Figure 10 is a structural schematic diagram of an angle of departure estimation device based on unsupervised learning provided by embodiments of the present application;

[0067] Figure 11 is a hardware structural schematic diagram of an electronic device provided by embodiments of the present application. DETAILED DESCRIPTION

[0068] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0069] It should be noted that although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a manner different from the module division in the device or the order in the flowchart. The terms "first", "second", and the like in the specification and claims and the above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0070] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

[0071] Artificial intelligence (AI): is a new technical science of researching, developing, simulating, extending and expanding human intelligence, and is a branch of computer science. Artificial intelligence attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence. The research in this field includes robots, language recognition, image recognition, natural language processing and expert systems. Artificial intelligence can simulate the information process of human consciousness and thinking. Artificial intelligence is also the theory, method, technology and application system of using digital computers or digital computer controlled machines to simulate, extend and expand human intelligence, to perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0072] Unsupervised Learning is a branch of machine learning that deals with unlabeled datasets. Unlike supervised learning, in unsupervised learning, the algorithm is not told which data belongs to the same class or the value of the target variable to be predicted, but needs to find the structure, pattern or relationship in the data itself.

[0073] Downlink (DL) refers to the process of sending information from a base station (BS) or other network node to a terminal device such as a mobile phone, computer or other user equipment (UE).

[0074] Channel State Information (CSI) is a term used in the field of wireless communication, which refers to the channel properties of a communication link. It describes the attenuation factors of the signal on each transmission path, i.e. the value of each element in the channel gain matrix H, such as signal scattering, environmental attenuation (fading, multipath fading or shadowing fading), distance attenuation (power decay of distance) and other information. CSI can make the communication system adapt to the current channel conditions and provide high reliability and high speed communication in multi-antenna systems.

[0075] The embodiments of the present application provide a kind of based on the angle of departure estimation method, device and related medium of unsupervised learning, to make terminal equipment accurately estimate the angle of departure information of base station terminal transmission signal.

[0076] The angle of departure estimation method, device and related medium based on unsupervised learning provided by the embodiments of the present application are specifically described by the following embodiments. First, the angle of departure estimation method based on unsupervised learning in the embodiments of the present application is described.

[0077] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. Artificial intelligence (AI) is the use of digital computers or digital computer-controlled machines to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0078] Artificial intelligence basic technologies generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics, etc. Artificial intelligence software technology mainly includes computer vision technology, robot technology, biometric technology, speech processing technology, natural language processing technology and machine learning / deep learning, etc.

[0079] The method for estimating the angle of departure based on unsupervised learning provided by the embodiments of the present application relates to the technical field of wireless communication. The method for estimating the angle of departure based on unsupervised learning provided by the embodiments of the present application can be applied to a terminal, can also be applied to a server end, and can also be software running in the terminal or the server end. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc. The server end can be configured as a stand-alone physical server, can also be configured as a server cluster or a distributed system composed of multiple physical servers, and can also be configured as a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and big data and artificial intelligence platform. The software can be an application for implementing the method for estimating the angle of departure based on unsupervised learning, but is not limited to the above forms.

[0080] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment in which tasks are performed by remote processing devices connected by a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0081] Figure 1 is an optional flowchart of the method for estimating the angle of departure based on unsupervised learning provided by the embodiments of the present application, Figure 1 The method in the above can include but is not limited to including steps S101 to S104.

[0082] Step S101, acquiring observation sequence data transmitted by a base station;

[0083] Step S102, acquiring pilot information based on a communication protocol to obtain pilot-related information;

[0084] Step S103, when the pilot-related information is specific pilot information, estimating the angle of departure of the observation sequence data and the pilot information by using a pre-trained first angle of departure estimation model to obtain target angle of departure information;

[0085] Step S104, in response to the pilot correlation information being pilot statistical information of the pilot information, performing angle of departure estimation on the observation sequence data and the pilot statistical information by using the pre-trained second angle of departure estimation model to obtain target angle of departure information.

[0086] The steps S101 to S106 shown in the embodiments of the present application obtain the observation sequence data transmitted by the base station, and obtain the pilot information based on the communication protocol to obtain the pilot correlation information. In response to the pilot correlation information being specific pilot information, the first angle of departure estimation model is used to perform angle of departure estimation on the observation sequence data and the pilot information to obtain target angle of departure information. In response to the pilot correlation information being pilot statistical information of the pilot information, the second angle of departure estimation model is used to perform angle of departure estimation on the observation sequence data and the pilot statistical information to obtain target angle of departure information. Thus, in order to improve the accuracy of angle of departure estimation, the present application not only uses the received observation sequence data as the input of the neural network model, but also uses the pilot information as the input. According to the obtained pilot correlation information, if the pilot correlation information is specific pilot information, the first angle of departure estimation model is used to perform angle of departure estimation. If the pilot correlation information is pilot statistical information, the second angle of departure estimation model is used to perform angle of departure estimation, so that the user equipment can perform angle of departure estimation according to different states of the pilot information to obtain accurate target angle of departure information.

[0087] In step S101 of some embodiments, the observation sequence data is received and obtained from the base station. In a wireless communication system, the observation sequence data generally refers to the actual measurement values of the signals periodically transmitted by the base station, which are affected by various factors such as path loss, multipath effect, signal attenuation, etc. during propagation to the user equipment. These observation sequence data contain rich information, which can be used to estimate the propagation characteristics of the signal, including the angle of arrival or the angle of departure of the signal. The user equipment refers to most of the devices in the Internet of Things.

[0088] In step S102 of some embodiments, the pilot information is obtained through the communication protocol. However, in some communication protocols, the specific pilot information may not be disclosed, but the statistical information about the pilot information, i.e. the pilot statistical information, is disclosed. The pilot statistical information here refers to the first-order statistical information and the second-order statistical information of the pilot information, i.e. the expected value, the variance and the covariance, etc. Therefore, when the pilot information is obtained through the communication protocol, the obtained pilot correlation information may be specific pilot information or pilot statistical information.

[0089] For a user equipment, pilot information can be obtained through a communication protocol and is a known signal. A base station can periodically send these pilot information, and a user equipment estimates channel state information (CSI) by comparing the received pilot information (i.e., observation sequence data) with the known pilot information. In the case of a single antenna device, due to the lack of spatial diversity in a multi-antenna system, the acquisition and processing of pilot information become more important and are the main source of angle of departure estimation.

[0090] After obtaining the pilot-related information, it is necessary to input it into the first angle of departure estimation model or the second angle of departure estimation model for angle of departure estimation according to whether the specific pilot information or the pilot statistical information is obtained.

[0091] In step S103 of some embodiments, if the obtained pilot-related information is specific pilot information, the observation sequence data and the pilot information are input into the pre-trained first angle of departure estimation model for angle of departure estimation to obtain target angle of departure information. The user equipment will use the pre-trained first angle of departure estimation model to process the observation sequence data and the pilot information. This model is designed based on the principle of deterministic maximum likelihood (DML) estimation, which uses specific pilot information to estimate the angle of departure (AoD) of the transmitted signal.

[0092] Referring to Figure 2A In some embodiments, the observation sequence data includes a first number of observation values, and step S103 can include but is not limited to steps S201 to S203:

[0093] Step S201, constructing a pilot signal matrix based on the pilot information;

[0094] Step S202, generating a first model input tensor according to the pilot signal matrix and the first number of observation values;

[0095] Step S203, inputting the first model input tensor into the pre-trained first angle of departure estimation model for angle of departure estimation to obtain target angle of departure information;

[0096] Referring to Figure 2B In some embodiments, step S104 can include but is not limited to steps S204 to S206:

[0097] Step S204, constructing a pilot vector matrix based on the pilot statistical information;

[0098] Step S205, performing Kronecker product processing based on the pilot vector matrix to obtain a pilot processing matrix;

[0099] Step S206, generating a second model input tensor according to the pilot processing matrix and the first number of observation values;

[0100] Step S207, inputting the second model input tensor into the pre-trained second angle of departure estimation model to perform angle of departure estimation, and obtaining target angle of departure information.

[0101] In step S201 of some embodiments, the observation sequence data is pilot information received by the user equipment from the base station during a transmission time, and the transmission time T of the observation sequence data can be divided into a first number Q of equal length blocks, each block being L unit time long. T = Q * L can be obtained. Denote the pilot information transmitted in block q as

[0102] Assuming that the user equipment remains stationary during the Q blocks and receives the observation sequence data, the observation value in block q can be represented as where Y (q) The value in formula (1) can be calculated by formula (1), and formula (1) is as follows:

[0103] Y (q) = ξX (q) a(θ) + w (q) (1)

[0104] where w (q) represents additive white Gaussian noise, and the additive white Gaussian noise (AWGN) describes the random noise superimposed on the transmitted signal in the process of signal transmission in a wireless communication system. This noise model is very important in both theoretical and practical communication system design, and it affects the quality of the signal and the performance of the system. The amplitude distribution of the additive white Gaussian noise follows the Gaussian distribution. ξ represents the channel gain parameter, which is determined by the distance between the base station and the user equipment. Since the user equipment remains stationary during the Q blocks, the channel gain parameter can be considered as a constant value. θ is the angle of departure, and a(θ) represents the steering vector calculated based on the angle of departure.

[0105] Therefore, the first number of observation values of the observation sequence data can be represented as Corresponding to the observation sequence data, the first number of pilot information is constructed into a pilot information matrix, and

[0106] In steps S202 to S203 of some embodiments, a first model input tensor is generated from the pilot signal matrix and the first number of observation values, wherein the first number of observation values is The first model input tensor S is obtained as follows:

[0107]

[0108] wherein Y L,Q represents the (L, Q)th item in the first number of observations B, X M,L represents the (M, L)th item in the first number of pilot signals A. Then the first model input tensor S is split into real and imaginary parts and input into the pre-trained first angle of departure estimation model to obtain the target angle of departure information.

[0109] In steps S204 to S205 of some embodiments, the pilot information is generated by wherein c (q) is the value of the pilot information of the qth block transmission, is the beamforming vector, and the pilot statistical information can be understood as the first-order statistical information and the second-order statistical information of c (q) , and the beamforming vector is also known, to obtain the pilot vector matrix Then, according to the Kronecker product of the pilot vector matrix and the all-1 vector of the first number of dimensions, the pilot processing matrix C is obtained.

[0110] In steps S206 to S207 of some embodiments, the pilot processing matrix and the first number of observations are used to generate a second model input tensor, wherein the first number of observations is to obtain the second model input tensor S, the specific steps are the same as steps S202, except that the pilot processing matrix C and the first number of observations B constitute the second model input tensor S. Then, the second model input tensor is split into real and imaginary parts and input into the pre-trained second angle of departure estimation model to obtain the target angle of departure information.

[0111] Please refer to Figure 3 In some embodiments, before step S103, the first angle of departure estimation model is pre-trained, which can include but is not limited to steps S301 to S303:

[0112] Step S301: obtaining a first training data set; the first training data set includes first observation training data and pilot training data;

[0113] Step S302: inputting the first observation training data and the pilot training data into the original first angle of departure estimation model to output first predicted angle of departure information and first channel gain estimation value;

[0114] Step S303, unsupervised learning is performed on the original first angle of departure estimation model based on the first predicted angle of departure information and the first channel gain estimation value according to the deterministic maximum likelihood estimation, to obtain a pre-trained first angle of departure estimation model.

[0115] In step S301 of some embodiments, a first training data set is obtained, which contains first observation training data for simulating real-world signal propagation and specific pilot information, i.e., pilot training data, for training. The first angle of departure estimation model is pre-trained using the first observation training data and the pilot training data, so that the pre-trained first angle of departure estimation model can accurately estimate the angle of departure information for specific pilot information.

[0116] In step S302 of some embodiments, the first observation training data and the pilot training data are input into the original first angle of departure estimation model for angle estimation, to obtain a prediction result including first predicted angle of departure information and a first channel gain estimation value.

[0117] In step S303 of some embodiments, the original first angle of departure estimation model is unsupervised learned based on the deterministic maximum likelihood estimation theory according to the prediction result obtained in step S302. Unsupervised learning is a training method that does not rely on labeled data, which allows the model to improve its performance through self-learning and self-adjustment. In this stage, the model may use various optimization algorithms to adjust its internal parameters, so as to provide more accurate predictions when encountering similar input data in the future. In this way, the model can gradually adapt to the characteristics of the training data and eventually obtain a pre-trained first angle of departure estimation model.

[0118] Through steps S301 to S303, the first angle of departure estimation model is pre-trained to enable it to provide accurate angle of departure information based on specific pilot information.

[0119] Please refer to Figure 4 In some embodiments, step S303 can include but is not limited to steps S401 to S403:

[0120] Step S401, based on the first predicted angle of departure information, a first predicted angle direction vector is calculated;

[0121] Step S402, a first model loss function is constructed based on the pilot training data, the first predicted angle direction vector, the first channel gain estimation value, and the first observation training data;

[0122] Step S403, based on the first model loss function, the first model parameters of the original first angle of departure estimation model are adjusted to obtain a pre-trained first angle of departure estimation model.

[0123] In step S401 of some embodiments, before the steering vector calculation, the pilot training data is parsed for data or according to the communication protocol to obtain the signal wavelength, the number of base station antennas and the base station antenna spacing. The base station will generally actively broadcast the number of base station antennas and other configuration information about the base station.

[0124] Then, based on the signal wavelength, the number of base station antennas, the base station antenna spacing and the first predicted angle of departure information, the steering vector calculation is performed to obtain the following formula (2):

[0125]

[0126] Wherein a(θ1) is the predicted angle steering vector, θ1 is the first predicted angle of departure information, λ represents the signal wavelength, M represents the number of base station antennas equipped by the base station, d represents the base station antenna spacing between the antennas configured by the base station, and j represents the imaginary part of the complex number.

[0127] In step S402 of some embodiments, the first model loss function is constructed according to the pilot training data, the first predicted angle steering vector, the first channel gain estimate and the first observation training data.

[0128] If the user equipment knows all the pilot information X (q) of the specific pilot information, then the actual measurement value Y (q) obeys the complex Gaussian distribution in formula (1). Thus, based on formula (1), the maximum likelihood estimation of the log-likelihood p(y (1) ,…,y (Q) ; θ1, ξ1) is equivalent to solving the least squares problem to obtain the first model loss function, as shown in the following formula (3):

[0129]

[0130] Wherein X1 represents the pilot training data, Y1 represents the signal of the first observation training data, D1 represents the first training data set, W1 represents the first model parameter of the original first angle of departure estimation model, ξ1 represents the first channel gain estimate, θ1 represents the first predicted angle of departure information, and a(θ1) represents the first predicted angle steering vector.

[0131] In step S403 of some embodiments, the first model loss function is obtained by step S402, and the original first angle of departure estimation model can learn how to better estimate the angle of departure and the channel gain according to the first observation training data and the pilot training data. With the continuous adjustment of the parameters, the performance of the model will be improved, and finally a pre-trained first angle of departure estimation model is obtained.

[0132] Through unsupervised learning, the model can be trained without real label data, which is very valuable in practical applications, because it can be difficult and time-consuming to obtain such label data.

[0133] Through steps S401 to S403, the original first angle of departure estimation model is pre-trained by using the first model loss function, so as to provide accurate angle of departure estimation based on specific pilot information.

[0134] In step S104 of some embodiments, if the obtained pilot-related information is the statistical information of the pilot information, the user equipment will use the pre-trained second angle of departure estimation model. Such a model is designed based on the principle of Stochastic Maximum Likelihood (SML) estimation, which can estimate the angle of departure in the case of only the first-order and second-order statistical information of the pilot information. Unlike deterministic models, stochastic models use statistical information of pilot information, such as the covariance matrix of pilot information, to estimate the angle of departure of the signal.

[0135] Please refer to Figure 5 In some embodiments, step S104 further includes pre-training the second angle of departure estimation model, which can include but is not limited to steps S501 to S504:

[0136] Step S501, obtaining a second training data set; the second training data set includes second observation training data and pilot training statistical data;

[0137] Step S502, inputting the second observation training data and the pilot training statistical data into the original second angle of departure estimation model to output second predicted angle of departure information, a second channel gain estimation value, and noise variance information;

[0138] Step S503, generating a reconstructed covariance matrix based on the second predicted angle of departure information, the second channel gain estimation value, and the noise variance information, and generating a sample covariance matrix based on the second observation training data;

[0139] Step S504, based on the stochastic maximum likelihood estimation, performing unsupervised learning on the original second angle of departure estimation model according to the reconstructed covariance matrix and the sample covariance matrix to obtain the pre-trained second angle of departure estimation model.

[0140] In step S501 of some embodiments, a second training dataset is obtained, which contains second observation training data and pilot training statistics for simulating real-world signal propagation scenarios. The second observation training data represents the signals received from the base station, while the pilot training statistics provide first-order and second-order statistics of the pilot information. Thus, the pre-trained second angle-of-departure estimation model can accurately estimate the angle-of-departure information in the case of only knowing the statistics of the pilot information.

[0141] In step S502 of some embodiments, the second observation training data and the pilot training statistics are input into the original second angle-of-departure estimation model for angle estimation, obtaining a prediction result including second predicted angle-of-departure information, a second channel gain estimate, and noise variance information.

[0142] In step S503 of some embodiments, based on the prediction result obtained in step S502, a reconstructed covariance matrix is generated. This matrix reflects the estimation of the signal and noise characteristics by the second angle-of-departure estimation model. At the same time, a sample covariance matrix is generated based on the second observation training data, which represents the statistical information of the actual received signal, i.e., the statistical information of the observation data. These two covariance matrices will be used to evaluate the performance of the model.

[0143] Please refer to Figure 6 In some embodiments, step S503 can include, but is not limited to, steps S601 to S603:

[0144] Step S601, based on the second predicted angle-of-departure information, a steering vector calculation is performed to obtain a second predicted angular steering vector;

[0145] Step S602, based on the second predicted angular steering vector, the second channel gain estimate, and the noise variance information, a reconstructed covariance matrix is generated;

[0146] Step S603, based on the second observation training data, a sample covariance matrix is generated.

[0147] In step S601 of some embodiments, the operation is synchronous with step S401, except that the calculation is based on the second predicted angle-of-departure information.

[0148] In step S602 of some embodiments, since the pilot training statistics obtained by the user equipment can only obtain statistical information about the pilot information, such as mean, variance, and covariance.

[0149] Based on the second predicted angular steering vector, the second channel gain estimate, and the noise variance information, a reconstructed covariance matrix is generated, as shown in the following formula (4):

[0150]

[0151] wherein C y denotes the reconstructed covariance matrix, ξ2denotes the second channel gain estimate, w (q) denotes the noise variance information, θ2denotes the second predicted angle of departure information, a(θ2) denotes the second predicted angle steering vector, X (q) (X (q) ) H denotes the covariance of the pilot information in the pilot training statistics, denotes the mean.

[0152] In step S603 of some embodiments, a sample covariance matrix is generated based on the second observation training data, as follows:

[0153]

[0154] wherein, denotes the sample covariance matrix, Q denotes the number of second observation training data, y (q) denotes the qth data of the second observation training data, (y (q) ) H denotes the conjugate transpose matrix of y (q) .

[0155] Through steps S601 to S603, the necessary input and feedback are provided for the training of the second angle of departure estimation model. By constructing a loss function based on the reconstructed covariance matrix and the sample covariance matrix, the prediction performance of the model can be evaluated, and the model parameters can be adjusted accordingly. This method of constructing a loss function based on a covariance matrix allows the model to perform unsupervised learning without real label data, thereby optimizing its prediction ability.

[0156] In step S504 of some embodiments, an unsupervised learning of the original second angle of departure estimation model is performed based on the reconstructed covariance matrix and the sample covariance matrix to construct a loss function. This process involves adjusting the model parameters using the loss function, in this way, the model can learn how to better estimate the angle of departure based on the statistical information of the pilot information. With the continuous adjustment of the parameters, the performance of the model will be improved, and finally a pre-trained second angle of departure estimation model is obtained.

[0157] Please refer to Figure 7 , in some embodiments, step S504 can include but is not limited to including steps S701 to S702:

[0158] Step S701, constructing a second model loss function based on the reconstructed covariance matrix and the sample covariance matrix;

[0159] Step S702, adjusting the second model parameters of the original second angle of departure estimation model based on the second model loss function to obtain a pre-trained second angle of departure estimation model.

[0160] In step S701 of some embodiments, the mean of the pilot of the lth time slot of the qth block of all pilot training statistics can be known by the statistical information of the pilot training statistics, that is, l∈[L] and q∈[Q], and the correlation between the pilot information of the lth and l' time slots Therefore, the second observation training data obeys a complex Gaussian distribution. Thus, the maximum likelihood estimation problem of the log-likelihood p(y (1) ,…,y (Q) ; θ2, ξ2) can be obtained based on formula (4) and formula (5), and the second model loss function is obtained by minimizing the problem, as shown in the following formula (6):

[0161]

[0162] Where C y represents the reconstructed covariance matrix, represents the sample covariance matrix, X2 represents the pilot training statistics, Y2 represents the second observation training data, D2 represents the second training data set, W2 represents the second model parameters of the original second angle of departure estimation model, and tr() represents the trace of the matrix, that is, the sum of the diagonal elements of the main diagonal of the matrix. det(C y ) represents the determinant value of the reconstructed covariance matrix;

[0163] In step S702 of some embodiments, the second model loss function is obtained by step S402, and the original second angle of departure estimation model can learn how to better estimate the angle of departure and channel gain according to the first observation training data and the pilot training data. With the continuous adjustment of the parameters, the performance of the model will be improved, and finally a pre-trained first angle of departure estimation model is obtained.

[0164] Through steps S701 to S702, the original second angle of departure estimation model is pre-trained by using the second model loss function, so that accurate angle of departure estimation can be provided according to the pilot statistical information.

[0165] Through steps S501 to S504, by constructing the reconstructed covariance matrix and the sample covariance matrix, the original second angle of departure estimation model is unsupervised learned based on the random maximum likelihood estimation theory, and a pre-trained second angle of departure estimation model is obtained, so that accurate target angle of departure information can be obtained according to the observation sequence data and the pilot statistical information.

[0166] In some experimental embodiments, the mean absolute error (MAE) metric is used to evaluate the error performance of the estimated departure angle. See also... Figure 8 The same experimental environment was set up, and training and validation sets were constructed using simulated or actual measurement data. Different sampling times (L) and different base station transmission powers (P) were tested, and the mean absolute error (MAE) of the estimated departure angle was obtained using various methods. In the figure, ESPRIT represents the method of estimating signal parameters via rotational invariance techniques, MUSIC represents the Multiple Signal Classification (MUSIC) algorithm, AE-based represents the signal processing method based on an autoencoder, DFT-based represents the signal processing method based on the Discrete Fourier Transform, Proposed (DML) represents the first departure angle estimation model designed based on the deterministic maximum likelihood estimation principle in this application, and Proposed (SML) represents the second departure angle estimation model designed based on the stochastic maximum likelihood estimation principle in this application.

[0167] pass Figure 8 As can be seen, the first departure angle estimation model proposed in this application can accurately estimate the departure angle with a very short sampling time, while other methods or models require more sampling time, i.e., more observation sequence data, to achieve similar accuracy. The second departure angle estimation model proposed in this application, designed based on the principle of stochastic maximum likelihood estimation, can also achieve high accuracy after a certain sampling time. The ability to achieve accurate departure angle estimation with a shorter sampling time means that only a small amount of data is needed to make an accurate departure angle estimate. For IoT devices with limited antenna resources, this can significantly reduce resource pressure and accurately estimate the departure angle information of the base station terminal's transmitted signal.

[0168] The base station's transmit power is related to the signal-to-noise ratio (SNR) of the received signal. The higher the base station's transmit power, the higher the SNR and the higher the signal quality of the signal received by the user equipment. Figure 9It can be seen that the first angle of departure estimation model proposed in the application can obtain good mean absolute error even in the case of very low base station transmit power, which shows that even in the case of poor signal quality, the first angle of departure estimation model can still obtain an angle of departure estimation result with high accuracy based on the received observation sequence data. This means that for Internet of Things devices with scarce antenna resources, even if a poor quality signal is received, accurate angle of departure estimation can still be performed through the first angle of departure estimation model. The second angle of departure estimation model designed based on the random maximum likelihood estimation principle proposed in the application can also obtain high accuracy when the base station transmit power is increased to a certain extent.

[0169] In some embodiments, since the first angle of departure estimation model and the second angle of departure estimation model are both designed by maximum likelihood estimation, although the second angle of departure estimation model outputs one more noise variance information than the first angle of departure estimation model, the first angle of departure estimation model can also be designed to output the same noise variance information, except that it does not use it as a key parameter for training. In this way, the second angle of departure estimation model and the first angle of departure estimation model are designed in the same architecture, thus simplifying the design of the neural network model.

[0170] The embodiment of the application obtains observation sequence data transmitted by a base station, and obtains pilot-related information based on a communication protocol. In response to the pilot-related information being specific pilot information, the first angle of departure estimation model is used to estimate the angle of departure of the observation sequence data and the pilot information to obtain target angle of departure information. In response to the pilot-related information being pilot statistical information of the pilot information, the second angle of departure estimation model is used to estimate the angle of departure of the observation sequence data and the pilot statistical information to obtain target angle of departure information. Thus, in order to improve the accuracy of angle of departure estimation, the application not only uses the received observation sequence data as the input of the neural network model, but also uses the pilot information as the input. According to the obtained pilot-related information, if the pilot-related information is specific pilot information, the first angle of departure estimation model is used for angle of departure estimation. If the pilot-related information is pilot statistical information, the second angle of departure estimation model is used for angle of departure estimation, so that the user equipment can perform angle of departure estimation according to different states of the pilot information to obtain accurate target angle of departure information.

[0171] Please refer to Figure 10 The embodiment of the application also provides an angle of departure estimation device based on unsupervised learning, which can implement the above-mentioned angle of departure estimation method based on unsupervised learning. The device comprises:

[0172] An observation data acquisition module is configured to obtain observation sequence data transmitted by a base station;

[0173] The pilot information acquisition module is configured to acquire pilot information based on a communication protocol to obtain pilot-related information.

[0174] The first angle estimation module is configured to, in response to the pilot-related information being specific pilot information, perform angle of departure estimation on the observation sequence data and the pilot information by using a pre-trained first angle of departure estimation model to obtain target angle of departure information.

[0175] The second angle estimation module is configured to, in response to the pilot-related information being pilot statistical information of the pilot information, perform angle of departure estimation on the observation sequence data and the pilot statistical information by using a pre-trained second angle of departure estimation model to obtain target angle of departure information.

[0176] The specific implementation of the angle of departure estimation device based on unsupervised learning is basically the same as the specific implementation of the angle of departure estimation method based on unsupervised learning, and will not be repeated here.

[0177] The embodiments of the present application also provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the angle of departure estimation method based on unsupervised learning when executing the computer program. The electronic device can be any intelligent terminal, such as a tablet computer or a vehicle-mounted computer.

[0178] Please refer to Figure 11 , Figure 11 The hardware structure of the electronic device of another embodiment is illustrated, which includes:

[0179] The processor 1101 can be implemented in the form of a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, and is used to execute related programs to implement the technical solutions provided by the embodiments of the present application.

[0180] The memory 1102 can be implemented in the form of a ROM (Read Only Memory), a static storage device, a dynamic storage device, or a RAM (Random Access Memory). The memory 1102 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are stored in the memory 1102 and called and executed by the processor 1101 to implement the angle of departure estimation method based on unsupervised learning.

[0181] The input / output interface 1103 is configured to realize information input and output.

[0182] The communication interface 1104 is configured to realize communication interaction between the device and other devices, and the communication can be realized in a wired manner (for example, a USB, a network cable, and the like) or in a wireless manner (for example, a mobile network, WIFI, Bluetooth, and the like).

[0183] The bus 1105 is configured to transmit information between various components (for example, the processor 1101, the memory 1102, the input / output interface 1103, and the communication interface 1104) of the device.

[0184] The processor 1101, the memory 1102, the input / output interface 1103, and the communication interface 1104 are connected to each other through the bus 1105 to realize communication connection between the device.

[0185] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the above-mentioned angle of departure estimation method based on unsupervised learning.

[0186] The memory is a non-transitory computer readable storage medium, and can be used to store a non-transitory software program and a non-transitory computer executable program. In addition, the memory can include a high-speed random access memory, and can further include a non-transitory memory, for example, at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor, and the remote memory can be connected to the processor through a network. Examples of the network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0187] The method, device and related medium for angle of departure estimation based on unsupervised learning provided by the embodiments of the present application obtain observation sequence data transmitted by a base station, and obtain pilot information based on a communication protocol to obtain pilot-related information. In response to the pilot-related information being specific pilot information, a first angle of departure estimation model is used to perform angle of departure estimation on the observation sequence data and the pilot information to obtain target angle of departure information. In response to the pilot-related information being pilot statistical information of the pilot information, a second angle of departure estimation model is used to perform angle of departure estimation on the observation sequence data and the pilot statistical information to obtain target angle of departure information. Thus, in order to improve the accuracy of angle of departure estimation, the embodiments of the present application not only use the received observation sequence data as input of a neural network model, but also use pilot information as input. According to the obtained pilot-related information, if the pilot-related information is specific pilot information, the first angle of departure estimation model is used to perform angle of departure estimation. If the pilot-related information is pilot statistical information, the second angle of departure estimation model is used to perform angle of departure estimation, so that the user equipment can perform angle of departure estimation according to different states of the pilot information to obtain accurate target angle of departure information.

[0188] The embodiments described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0189] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than the figures, or combine certain steps or different steps.

[0190] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separate, that is, can be located in one place or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0191] Those skilled in the art can understand that all or some of the steps in the above disclosed method, the functions of the modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0192] The terms "first", "second", "third", "fourth", and the like in the description of this application and in the claims hereof, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of the terms so termed herein is solely for the convenience of the reader and does not limit the scope of the application. It is also to be understood that the description and examples in this application are intended to cover all possible combinations where any of the several elements can represent one or more elements.

[0193] It should be understood that, in the application, "at least one" means one or more, "multiple" means two or more. "And / or" is used to describe the relationship between associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0194] In several embodiments provided in the application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the above-mentioned units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. The coupling or direct coupling or communication connection between the displayed or discussed each other can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0195] The units described above as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment of the present application.

[0196] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0197] When the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in part, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions used to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods in the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various other media that can store programs.

[0198] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A method of angle of departure estimation based on unsupervised learning, characterized in that, The method comprises: obtaining observation sequence data transmitted by a base station; obtaining pilot information based on a communication protocol to obtain pilot-related information; in response to the pilot-related information being specific pilot information, performing angle of departure estimation on the observation sequence data and the pilot information through a pre-trained first angle of departure estimation model to obtain target angle of departure information; in response to the pilot-related information being pilot statistical information of the pilot information, performing angle of departure estimation on the observation sequence data and the pilot statistical information through a pre-trained second angle of departure estimation model to obtain target angle of departure information.

2. The method of claim 1, wherein, The observation sequence data comprises a first number of observation values; The angle of departure estimation on the observation sequence data and the pilot information through the pre-trained first angle of departure estimation model to obtain target angle of departure information comprises: constructing a pilot signal matrix based on the pilot information; generating a first model input tensor according to the pilot signal matrix and the first number of observation values; inputting the first model input tensor into the pre-trained first angle of departure estimation model for angle of departure estimation to obtain target angle of departure information; The angle of departure estimation on the observation sequence data and the pilot statistical information through the pre-trained second angle of departure estimation model to obtain target angle of departure information comprises: constructing a pilot vector matrix based on the pilot statistical information; performing Kronecker product processing based on the pilot vector matrix to obtain a pilot processing matrix; generating a second model input tensor according to the pilot processing matrix and the first number of observation values; inputting the second model input tensor into the pre-trained second angle of departure estimation model for angle of departure estimation to obtain target angle of departure information.

3. The method of claim 1, wherein, Before the angle of departure estimation on the observation sequence data and the pilot information through the pre-trained first angle of departure estimation model to obtain target angle of departure information, the method further comprises pre-training the first angle of departure estimation model, comprising: obtaining a first training data set; the first training data set comprises first observation training data and pilot training data; inputting the first observation training data and the pilot training data into an original first angle of departure estimation model to output first predicted angle of departure information and a first channel gain estimation value; performing unsupervised learning on the original first angle of departure estimation model based on deterministic maximum likelihood estimation according to the first predicted angle of departure information and the first channel gain estimation value to obtain the pre-trained first angle of departure estimation model.

4. The method of claim 3, wherein, The unsupervised learning on the original first angle of departure estimation model based on deterministic maximum likelihood estimation according to the first predicted angle of departure information and the first channel gain estimation value to obtain the pre-trained first angle of departure estimation model comprises: performing steering vector calculation based on the first predicted angle of departure information to obtain a first predicted angle steering vector; constructing a first model loss function based on the pilot training data, the first predicted angle steering vector, the first channel gain estimation value, and the first observation training data: Wherein, X1 represents the pilot training data, Y1 represents the first observation training data, D1 represents the first training data set, W1 represents the first model parameter of the original first angle of departure estimation model, ξ1 represents the first channel gain estimation value, θ1 represents the first predicted angle of departure information, a(θ1) represents the first predicted angle steering vector; Adjust the first model parameter of the original first angle of departure estimation model based on the first model loss function, to obtain the pre-trained first angle of departure estimation model.

5. The method of claim 1, wherein, Before the target angle of departure information is obtained by performing angle of departure estimation on the observation sequence data and the pilot statistical information through the pre-trained second angle of departure estimation model, the second angle of departure estimation model is further pre-trained, comprising: Obtain a second training data set; the second training data set comprises second observation training data and pilot training statistical data; Input the second observation training data and the pilot training statistical data into an original second angle of departure estimation model to output second predicted angle of departure information, a second channel gain estimation value and noise variance information; Generate a reconstructed covariance matrix based on the second predicted angle of departure information, the second channel gain estimation value and the noise variance information, and generate a sample covariance matrix based on the second observation training data; Based on the random maximum likelihood estimation, the original second angle of departure estimation model is unsupervised learning according to the reconstructed covariance matrix and the sample covariance matrix, to obtain the pre-trained second angle of departure estimation model.

6. The method of claim 5, wherein, The method further comprises: Based on the second predicted angle of departure information, a second predicted angle steering vector is calculated; Based on the second predicted angle steering vector, the second channel gain estimation value and the noise variance information, a reconstructed covariance matrix is generated: where C y represents the reconstructed covariance matrix, ξ2represents the second channel gain estimate, w (q) represents the noise variance information, θ2represents the second predicted angle of departure information, a(θ2) represents the second predicted angle steering vector, X (q) (X (q) ) H represents the covariance of pilot information in the pilot training statistics, represents the expectation value; Based on the second observation training data, a sample covariance matrix is generated: wherein, represents the sample covariance matrix, Q represents the number of the second observed training data, y (q) represents the qth data of the second observed training data, (y (q) ) H represents the conjugate transpose matrix of y (q) .

7. The method of claim 5, wherein, The method further comprises: Based on the reconstructed covariance matrix and the sample covariance matrix, a second model loss function is constructed: where C y represents the reconstructed covariance matrix, represents the sample covariance matrix, X2represents the pilot training statistics, Y2represents the second observation training data, D2represents the second training data set, W2represents the second model parameters of the original second angle of departure estimation model, tr() refers to the trace of the matrix, i.e. the sum of the diagonal elements of the main diagonal of the matrix, det(C y ) represents the determinant value of the reconstructed covariance matrix; Based on the second model loss function, the second model parameter of the original second angle of departure estimation model is adjusted to obtain the pre-trained second angle of departure estimation model.

8. An apparatus for estimating a departure angle based on unsupervised learning, the apparatus comprising: The device comprises: An observation data acquisition module is configured to acquire observation sequence data transmitted by a base station; A pilot information acquisition module is configured to acquire pilot information based on a communication protocol to obtain pilot-related information; A first angle estimation module is configured to, when the pilot-related information is specific pilot information, perform angle of departure estimation on the observation sequence data and the pilot information through a pre-trained first angle of departure estimation model to obtain target angle of departure information. A second angle estimation module is configured to, in response to the pilot correlation information being pilot statistical information of the pilot information, perform angle of departure estimation on the observation sequence data and the pilot statistical information by using a pre-trained second angle of departure estimation model to obtain target angle of departure information.

9. An electronic device, comprising: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the unsupervised learning-based angle of departure estimation method in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1-9. The computer program is executed by the processor to implement the unsupervised learning-based angle of departure estimation method in any one of claims 1 to 7.

Images