Wireless channel environment identification method and system, and storage medium

Abstract

The invention discloses a wireless channel environment identification method and system, and a storage medium, and the method comprises the steps: building a wireless channel detection device, simulating the pilot frequency receiving and transmitting of a communication device, and collecting the position information and pilot frequency data sets in different scenes in real time; and the wireless channel environment identification model is adopted to directly identify the channel transfer function set and the position information in the normal communication process in each scene, so that the wireless channel environment identification result in each scene is obtained, and the efficiency and the accuracy of channel environment identification are improved.

Description

Technical Field

[0001] The present invention relates to the field of communications, and more specifically, to a method, system and storage medium for identifying a wireless channel environment. Background Art

[0002] The performance of wireless communication systems is significantly affected by the channel environment in which they are located. Given the complexity and variability of the wireless channel environment, the use of optimal receiving algorithms and parameter configurations suitable for specific channel scenarios can significantly improve the overall performance of wireless communication systems. Therefore, accurately identifying various channel scenarios is crucial to improving system performance.

[0003] In the prior art, the data collected by the pilot signal is transformed, rather than directly using the frequency domain baseband information of the pilot for identification and classification. In addition, a corresponding channel detection system has not yet been established to collect actual measurement data. At the same time, the existing technology mainly relies on traditional statistical algorithms to identify the channel environment, rather than using artificial intelligence methods for identification and classification. Summary of the invention

[0004] In order to overcome the defect of low accuracy of channel environment recognition in various scenarios in the prior art, the present invention provides a wireless channel environment recognition method, system and storage medium.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] The present invention provides a method for identifying a wireless channel environment, comprising:

[0007] The wireless channel detection device simulates the pilot signal transmission and reception of the communication device and collects the position information and pilot signal data sets in different scenarios in real time, and determines the corresponding channel transfer function set according to the pilot signal data set in each scenario;

[0008] The position information and channel transfer function set corresponding to each scene are input into a preset wireless channel environment identification model to obtain the wireless channel environment identification result in each scene.

[0009] Preferably, the wireless channel detection device includes a wireless transmitting device and a wireless receiving device; the wireless transmitting device includes a radio transmitter, a power supply device, a power amplifier, a transmitting end GPS antenna, a transmitting antenna and a transmitting end mobile device; the wireless receiving device includes a radio receiver, a receiving end GPS antenna, a receiving end mobile device and a receiving antenna.

[0010] Preferably, the wireless channel detection device simulates the pilot signal transmission and reception of the communication device and collects the location information and pilot signal data sets in different scenarios in real time, including:

[0011] Install the wireless transmitting device and the wireless receiving device on the transmitting end test vehicle and the receiving end test vehicle respectively, and synchronize the initial test broadband signal frequency points of the wireless transmitting device and the wireless receiving device;

[0012] When the transmission signal of the wireless transmitting device is transmitted to the receiving signal of the wireless receiving device normally, the mobile devices of the transmitting end test vehicle and the receiving end test vehicle are used to collect the location information and pilot data set corresponding to the initial broadband signal frequency point in the current scene in real time;

[0013] Changing the current test broadband signal frequency, synchronizing the current test broadband signal frequency of the wireless transmitting device and the wireless receiving device until all test broadband signal frequencies are tested, completing the pilot signal transmission and reception of the wireless channel detection device simulating the communication device and collecting the position information and pilot signal data set corresponding to all test broadband signal frequencies in the current scenario in real time;

[0014] Change different scenarios until all test broadband signal frequencies in all scenarios are tested, complete the pilot transmission and reception of the wireless channel detection equipment to simulate the communication equipment, and collect the location information and pilot data sets in different scenarios in real time.

[0015] Preferably, the transmission of the transmission signal of the wireless transmitting device to the reception signal of the wireless receiving device includes:

[0016] The radio transmitter generates a detection signal and performs secondary frequency conversion on the detection signal, amplifies the power of the secondary frequency-converted detection signal through a power amplifier, and then transmits the detection signal through a transmitting antenna;

[0017] The radio receiver samples and down-converts the detection signal collected by the receiving antenna to obtain a received signal;

[0018] Preferably, the scenes include different scenes such as flat terrain scenes, mountainous terrain scenes and urban scenes.

[0019] Preferably, the pilot data set in each scenario is subjected to digital signal processing to obtain a corresponding channel impulse response, and the channel impulse response is subjected to computational processing to obtain a signal transfer function set corresponding to the pilot data set in each scenario.

[0020] Preferably, the wireless channel environment recognition model uses a distance metric learning method to match different scene types and their corresponding channel transfer function sets.

[0021] Preferably, the preset wireless channel environment identification model is a SVM model or other artificial intelligence classification algorithms.

[0022] The present invention also provides a wireless channel environment identification system for implementing the above method, comprising:

[0023] The data acquisition module collects the location information and pilot data sets in different scenarios in real time through the wireless channel detection device, and determines the corresponding channel transfer function set according to the pilot data set in each scenario;

[0024] The identification module inputs the location information and channel transfer function set corresponding to each scene into a preset wireless channel environment identification model to obtain the wireless channel environment identification result under each scene.

[0025] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program implements the above-mentioned wireless channel environment identification method when executed by a processor.

[0026] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0027] The present invention proposes a wireless channel environment identification method, system and storage medium, builds a wireless channel detection device to collect position information and pilot data sets in real time in different scenarios, and adopts a wireless channel environment identification model to directly identify the channel transfer function set and position information in the normal communication process in each scenario, thereby obtaining the wireless channel environment identification result in each scenario, and improving the efficiency and accuracy of channel environment identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a flowchart of the wireless channel environment identification method described in Example 1;

[0029] Figure 2 This is a schematic diagram of the structure of the wireless channel detection device described in Example 2;

[0030] Figure 3 This is a flowchart of the wireless channel environment identification method described in Example 2;

[0031] Figure 4 This is a schematic diagram of the structure of the wireless channel environment identification system described in Example 3. DETAILED DESCRIPTION

[0032] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0033] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0034] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0035] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0036] Example 1

[0037] This embodiment provides a method for identifying a wireless channel environment. Figure 1 As shown, including:

[0038] The wireless channel detection device collects the location information and pilot data sets in different scenarios in real time, and determines the corresponding channel transfer function set according to the pilot data set in each scenario;

[0039] The position information and channel transfer function set corresponding to each scene are input into a preset wireless channel environment identification model to obtain the wireless channel environment identification result in each scene.

[0040] In a specific embodiment, the wireless channel detection device includes a wireless transmitting device and a wireless receiving device; the wireless transmitting device includes a radio transmitter, a power supply device, a power amplifier, a transmitting end GPS antenna, a transmitting antenna and a transmitting end mobile device; the wireless receiving device includes a radio receiver, a receiving end GPS antenna, a receiving end mobile device and a receiving antenna.

[0041] In a specific embodiment, the location information and pilot data sets in different scenarios are collected in real time by a wireless channel detection device, including:

[0042] Install the wireless transmitting device and the wireless receiving device on the transmitting end test vehicle and the receiving end test vehicle respectively, and synchronize the initial test broadband signal frequency points of the wireless transmitting device and the wireless receiving device;

[0043] When the transmission signal of the wireless transmitting device is transmitted to the receiving signal of the wireless receiving device normally, the mobile devices of the transmitting end test vehicle and the receiving end test vehicle are used to collect the location information and pilot data set corresponding to the initial broadband signal frequency point in the current scene in real time;

[0044] Changing the current test broadband signal frequency, synchronizing the current test broadband signal frequency of the wireless transmitting device and the wireless receiving device until all test broadband signal frequencies are tested, and completing the real-time collection of the position information and pilot data set corresponding to all test broadband signal frequencies in the current scenario by the wireless channel detection device;

[0045] Change different scenarios until all test broadband signal frequency points in all scenarios are tested, and the wireless channel detection device collects the location information and pilot data sets in different scenarios in real time.

[0046] In a specific embodiment, the transmission of a transmission signal of a wireless transmitting device to a reception signal of a wireless receiving device includes:

[0047] The radio transmitter generates a detection signal and performs secondary frequency conversion on the detection signal, amplifies the power of the secondary frequency-converted detection signal through a power amplifier, and then transmits the detection signal through a transmitting antenna;

[0048] The radio receiver samples and down-converts the detection signal collected by the receiving antenna to obtain a received signal;

[0049] In a specific embodiment, the scenes include flat terrain scenes, mountainous terrain scenes and urban areas scenes.

[0050] In a specific embodiment, the pilot data set in each scenario is subjected to digital signal processing to obtain a corresponding channel impulse response, and the channel impulse response is processed by calculation to obtain a signal transfer function set corresponding to the pilot data set in each scenario.

[0051] In a specific embodiment, the wireless channel environment recognition model uses a distance metric learning method to match different scene types and their corresponding channel transfer function sets.

[0052] In a specific embodiment, the preset wireless channel environment identification model is a SVM model or other artificial intelligence classification algorithms.

[0053] In this embodiment, a channel detection platform is constructed using a general software radio platform, and a channel detection waveform based on a communication pilot sequence is designed to collect data from actual outdoor scenes. Then, channel tests are carried out in different environments such as urban areas, mountainous areas and flat areas, and a large amount of measured data is collected. Using these measured data, the channel impulse response is extracted by digital signal processing technology, and the signal transfer function is further obtained. These signal transfer functions and their corresponding scene types are used as training and test data for machine learning models. This embodiment builds a machine learning framework based on PyTorch, and uses a support vector machine (SVM) algorithm to train the data. During the training process, 80% of the data is randomly selected for model training, and the remaining 20% ​​of the data is used for testing. The test results show that the recognition accuracy of the model reaches more than 90%. The scene fast recognition technology of this embodiment, which combines pilot sequence detection, measured data collection and machine learning algorithm, can effectively identify the scene type according to the characteristics of the wireless channel. This technology can assist communication radio stations in waveform adjustment and decision making, thereby optimizing communication performance.

[0054] Example 2

[0055] Based on the wireless channel environment identification method described in Example 1, this embodiment lists a specific example as follows:

[0056] A general software radio platform is used to build a channel detection system to collect channel characteristics under different environments. The channel measurement system uses the ZC signal simulation pilot as the detection signal, simulates the pilot sequence, takes the software radio device as the core hardware device, and designs the channel detection waveform through software programming. The main performance parameters of the measurement system are shown in the following table. The measurement system can be divided into hardware and software parts. The hardware part realizes the sending and receiving of the detection signal, including software radio transmitter / receiver, GPS antenna, power supply module, notebook, wide-band power amplifier, transmitting antenna / receiving antenna, etc.; the software part completes the driving of the hardware equipment and the generation, real-time processing and storage control of the test signal, and provides a visual control and display interface for testers.

[0057] The software radio transmitter / receiver is the core of the measurement system. The software radio transmitter imports / generates the baseband detection signal and up-converts the baseband detection signal through secondary frequency conversion (baseband to intermediate frequency, intermediate frequency to radio frequency). The software radio receiver samples and down-converts the signal collected by the antenna to obtain the baseband signal and stores it on the notebook disk. In addition, the software radio receiver can also process the received signal in real time, obtain some key channel parameters and display them visually.

[0058] like Figure 2 As shown, at the transmitting end, the test signal output by the software radio transmitter is amplified by a power amplifier and then transmitted through a transmitting antenna; at the receiving end, the receiver periodically collects the measurement data under the synchronization signal and records the GPS information of the collection time and collection location.

[0059] Based on the soft channel detector, a high-precision V2V wireless channel measurement system that can meet the simultaneous movement of the transmitter and receiver is constructed with a car as the platform. The channel detector is very small and can be placed on the car. The power supply is a 12V uninterruptible power supply (UPS). The channel detector needs to be calibrated and synchronized before starting the measurement, so a GPS antenna is configured on the roof of the Rx and Tx cars for synchronization.

[0060] Because the channel detector is small enough, it can be placed in the back seat of a car to perform channel measurements in a real environment. At the same time, both the transmitter and receiver parts of the channel detector are connected to a laptop computer via an Ethernet cable. The transmitter part controls the transmitter to transmit the ZC signal through the human-computer interaction interface, and imports the Tx location information into the laptop computer through GPS and records it. The receiver part can see the received waveform in real time, store the detected channel data, and record the location information obtained by GPS.

[0061] The test site is a typical environment, and the hardware environment test equipment includes a test vehicle, software radio, receiving and transmitting antennas, antenna racks, laptops, and low-power amplifiers. The general test method is to select different terrains to send and receive signals and collect original signals for subsequent analysis.

[0062] The specific test steps are as follows:

[0063] Transport the antenna, instruments and supporting equipment to the vehicle and transport them to the appropriate test site;

[0064] Connect the transmitting antenna to the power amplifier and fix it on the roof, then connect the power amplifier to the transmitting software radio and connect it with a laptop;

[0065] Fix the receiving antenna to the antenna rack and install it on the roof;

[0066] Connect the receiving antenna to the software radio in the car through a feeder, and then connect the software radio to the computer through an Ethernet cable;

[0067] Use the vehicle power supply to power it and turn it on;

[0068] Configure software radio waveform parameters: mainly select broadband signal frequency;

[0069] Drive the vehicle to the typical terrain that has been explored in the early stage, stop the vehicle to perform static channel measurement, or ensure that both vehicles are moving to perform dynamic channel measurement;

[0070] Select the initial test frequency, and the transmitter starts transmitting signals;

[0071] Turn on the receiving software radio and confirm that the received signal is normal;

[0072] The transmitter and receiver select other predetermined test frequencies, and the transmitter is changed to the corresponding frequency;

[0073] Collect IQ signal data of the receiving software radio through a laptop and save the records;

[0074] Record the detailed GPS location and altitude information of the transceiver corresponding to the test data, take photos of the environment, record the temperature and weather conditions, and save these data and the test data in the same folder and name it;

[0075] Change the frequency and test location respectively. After testing all frequency points in all terrains, shut down and power off the equipment, organize the equipment on the vehicle, and end the test.

[0076] like Figure 3As shown, machine learning is used to achieve rapid perception of scenes in multiple regions. First, the waveform is designed and data is collected in the field through the channel measurement system. The channel transfer function set under different environments is obtained through calculation and processing, and 80% of the data is taken out to form a training data set. Subsequently, the SVM classifier is used for training. The classification categories include flat terrain scenes, mountainous terrain scenes, and urban scenes. The training equipment uses a graphics card host with NVIdia3090Ti for training. 112,600 sets of data are used for training, the training time is 180s, and then 28,200 sets of data are used for testing, with an accuracy rate of more than 90%.

[0077] In order to mine relevant information from a variety of data. When training these machine learning models, a suitable distance (similarity) metric can ensure stable and good performance. In the era of big data, it is unreasonable to manually design a suitable distance metric for each data set of a learning task. In view of this situation, distance metric learning can automatically learn a suitable distance metric for machine learning tasks. This paper uses distance metric learning to identify the channel transfer function obtained after Fourier inverse transformation of flat areas, mountainous terrains and urban areas, where the input is the delayed power spectrum.

[0078] Distance metric learning aims to handle data distribution by learning appropriate distance metrics from training sets, and has become a very active research field in recent years. Based on the learned distance metric, the classification model can minimize the distance between examples of the same category and maximize the distance between examples of different categories. In most distance metric learning methods, an optimization problem for metric learning is generally constructed by parameterizing the Mahalanobis distance metric matrix. In detail, first, corresponding examples are constructed for each training example, and pairwise penalty constraints or triple penalty constraints are constructed based on these example pairs. Then, the parameterized Mahalanobis distance metric matrix can be optimized by penalizing the large distance of similar example pairs and the small distance of non-similar example pairs. In practical applications, distance metric learning has a wide range of applications, such as face recognition, image retrieval, behavior recognition, and environmental perception.

[0079] However, despite great progress in these areas, most existing metric learning methods can be regarded as nearest neighbor distance metric learning, which is used to improve the performance of nearest neighbor classification models, while there are few studies on non-nearest neighbor distance metric learning. For the nearest neighbor distance metric learning problem, we first learn a suitable distance metric as a distance measurement standard. Then, the nearest neighbor classification model is constructed using the learned distance metric. In the prediction stage, in order to classify the test example, the distance from all training examples to the test example needs to be calculated. Then, these calculated distances are sorted to select the closest multiple training examples. The label of the test example is determined by the common label in the closest training example. Therefore, it can be seen that for the nearest neighbor distance metric learning, its disadvantage is that it needs to store and access all training examples to classify the test example. In order to overcome the disadvantages of the nearest neighbor distance metric learning, non-nearest neighbor metric learning models can be studied.

[0080] Unlike the nearest neighbor classification model, the support vector machine is an excellent non-nearest neighbor learning algorithm with complete theoretical support and a large amount of mature research work, and does not have the shortcomings of the fc nearest neighbor model described above. For the support vector machine model, a set of positive and negative parallel hyperplanes can be constructed as the boundaries of positive and negative examples respectively. When training the support vector machine model, by maximizing the edge distance (Euclidean distance) of this set of positive and negative parallel hyperplanes, a separation hyperplane (maximum edge hyperplane) can be explicitly learned as a classifier. Classifying the test example only requires projecting it into the same feature space of the separation hyperplane. Compared with the nearest neighbor classification model, the support vector machine model does not need to calculate the distance from the test example to all training examples. The category to which the test example belongs is determined by the positive or negative side of its projection to the separation hyperplane, rather than by the common labels in the A closest training examples around it.

[0081] SVM is an excellent learning algorithm in pattern recognition and statistical learning. It is proposed to solve pattern classification and regression problems. For the standard support vector machine, the basic idea is to find the maximum margin hyperplane, that is, the splitting hyperplane, which can separate the examples of two classes from each other with a fixed-size margin. In the standard support vector machine, two parallel hyperplanes can be constructed, each hyperplane is close to one class of examples and far away from the other class of examples. The splitting hyperplane can be obtained by maximizing the distance between the two parallel hyperplanes. That is, when two parallel hyperplanes with the maximum margin are found, the hyperplane located in the middle of their edges is the splitting hyperplane to be found.

[0082] However, most of the existing research is oriented to the SVM model based on Euclidean distance, while there are relatively few studies on the SVM model based on distance metric learning. For the SVM based on Euclidean distance, the Euclidean distance is used as the distance measurement method in the process of training the SVM model. As we all know, the Euclidean distance only considers the feature similarity of the input examples, while ignoring the label information of the examples. Since the distance measurement method of the training process is defined as the Euclidean distance, the distance between the positive example and the negative example is calculated based on the similarity of the entire example. For these examples that are close to each other, even if they have different labels, they will be classified to the same side of the segmentation hyperplane. Therefore, using the Euclidean distance as the distance measurement method of the support vector machine may lead to classification errors. In order to solve this problem, a convex optimization distance metric learning model can be introduced into the support vector machine, and a suitable Mahalanobis distance metric is used as the distance measurement method of the support vector machine. This specific Mahalanobis distance metric is learned through the annotation information and data features of the training examples. In addition, by using this trained specific Mahalanobis distance metric in the support vector machine, the misclassified example can be projected to the correct side of the segmentation hyperplane based on distance metric learning according to the label information of the misclassified example.

[0083] In PyTorch, SVM can be used for classification tasks, and 80% of the data in the dataset is used for training and 20% of the data is used for validation. Implementing SVM for task classification usually involves the following steps:

[0084] First, prepare the dataset for training and testing. The data should be labeled and each data point should belong to a specific category.

[0085] In PyTorch, you can use a linear SVM model to try to find a linear hyperplane to separate different categories;

[0086] Use the data set to train the SVM model and automatically find the best hyperplane during the training process;

[0087] The performance of the SVM model is evaluated using the test dataset, and the classification performance is usually measured using indicators such as recognition rate.

[0088] Example 3

[0089] This embodiment also provides a wireless channel environment identification system. Figure 4 As shown, the method for implementing embodiment 1 or 2 includes:

[0090] The data acquisition module simulates the pilot signal transmission and reception of the communication device through the wireless channel detection device and collects the position information and pilot signal data sets in different scenarios in real time, and determines the corresponding channel transfer function set according to the pilot signal data set in each scenario;

[0091] The identification module inputs the location information and channel transfer function set corresponding to each scene into a preset wireless channel environment identification model to obtain the wireless channel environment identification result under each scene.

[0092] This embodiment further provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the wireless channel environment identification method of embodiment 1 or 2 is implemented.

[0093] The same or similar reference numerals correspond to the same or similar components;

[0094] The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent;

[0095] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for identifying a wireless channel environment, characterized in that: include: The wireless channel detection device simulates the pilot signal transmission and reception of the communication device and collects the position information and pilot signal data sets in different scenarios in real time, and determines the corresponding channel transfer function set according to the pilot signal data set in each scenario; The position information and channel transfer function set corresponding to each scene are input into a preset wireless channel environment identification model to obtain the wireless channel environment identification result in each scene.

2. The wireless channel environment identification method according to claim 1, characterized in that: The wireless channel detection device includes a wireless transmitting device and a wireless receiving device; the wireless transmitting device includes a radio transmitter, a power supply device, a power amplifier, a transmitting end GPS antenna, a transmitting antenna and a transmitting end mobile device; the wireless receiving device includes a radio receiver, a receiving end GPS antenna, a receiving end mobile device and a receiving antenna.

3. The wireless channel environment identification method according to claim 2, characterized in that: The wireless channel detection device simulates the pilot signal transmission and reception of the communication device in real time and collects the location information and pilot signal data sets in different scenarios, including: Install the wireless transmitting device and the wireless receiving device on the transmitting end test vehicle and the receiving end test vehicle respectively, and synchronize the initial test broadband signal frequency points of the wireless transmitting device and the wireless receiving device; When the transmission signal of the wireless transmitting device is transmitted to the receiving signal of the wireless receiving device normally, the mobile devices of the transmitting end test vehicle and the receiving end test vehicle are used to collect the location information and pilot data set corresponding to the initial broadband signal frequency point in the current scene in real time; Changing the current test broadband signal frequency, synchronizing the current test broadband signal frequency of the wireless transmitting device and the wireless receiving device until all test broadband signal frequencies are tested, and completing the real-time collection of the position information and pilot data set corresponding to all test broadband signal frequencies in the current scenario by the wireless channel detection device; Change different scenarios until all test broadband signal frequency points in all scenarios are tested, and the wireless channel detection device collects the location information and pilot data sets in different scenarios in real time.

4. The wireless channel environment identification method according to claim 3, characterized in that: The transmission signal of the wireless transmitting device is transmitted to the receiving signal of the wireless receiving device, including: The radio transmitter generates a detection signal and performs secondary frequency conversion on the detection signal, amplifies the power of the secondary frequency-converted detection signal through a power amplifier, and then transmits the detection signal through a transmitting antenna; The radio receiver samples and down-converts the detection signal collected by the receiving antenna to obtain a received signal.

5. The wireless channel environment identification method according to claim 1, characterized in that: The scenarios include flat terrain, mountainous terrain, and urban areas.

6. The wireless channel environment identification method according to claim 1, characterized in that: The pilot data set in each scenario is subjected to digital signal processing to obtain a corresponding channel impulse response, and the channel impulse response is subjected to computational processing to obtain a signal transfer function set corresponding to the pilot data set in each scenario.

7. The wireless channel environment identification method according to claim 1, characterized in that: The wireless channel environment recognition model adopts a distance metric learning method to match different scene types and their corresponding channel transfer function sets.

8. The method for identifying a wireless channel environment according to claim 1, characterized in that: The preset wireless channel environment recognition model is a SVM model or other artificial intelligence classification algorithms.

9. A wireless channel environment identification system, used to implement the method of claims 1-8, characterized in that: include: The data acquisition module simulates the pilot signal transmission and reception of the communication device in real time through the wireless channel detection device and collects the location information and pilot signal data sets in different scenarios, and determines the corresponding channel transfer function set according to the pilot signal data set in each scenario; The identification module inputs the location information and channel transfer function set corresponding to each scene into a preset wireless channel environment identification model to obtain the wireless channel environment identification result under each scene.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the wireless channel environment identification method according to any one of claims 1 to 8 is implemented.

Images