Vehicle-mounted wireless module data transmission control method based on interference detection

Abstract

The present application relates to the technical field of data processing, in particular to a vehicle-mounted wireless module data transmission control method based on interference detection. The strength of interference is determined by calculating the signal-to-noise ratio of the received signal, and the type of interference is distinguished by power spectrum density analysis, i.e. same frequency, adjacent frequency and random noise. Based on the center power, adjacent frequency power, power variance and historical packet loss rate, the frequency band is evaluated. Then, combined with the future driving path of the vehicle, a graph neural network prediction model is constructed, the frequency band state value of the future path node is predicted in advance by using the node adjacency relationship and historical interference characteristics, and an optimal frequency band switching scheme is selected to minimize the transmission stability and switching times, plan the global optimal switching scheme in advance, avoid frequent switching, improve the communication stability and anti-interference ability, and improve the vehicle-mounted wireless module data transmission quality in the vehicle networking environment.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a data transmission control method for vehicle-mounted wireless modules based on interference detection. Background Technology

[0002] In vehicle-to-everything (V2X) communication, in-vehicle wireless modules are susceptible to various interference factors during operation. Communication systems typically rely on signal-to-noise ratio (SNR) or power detection to roughly determine the presence of interference, but these methods are insufficient for accurately identifying the specific type of noise.

[0003] On the one hand, current conventional methods can usually only distinguish between strong and weak interference, lacking a detailed analysis of the nature of the interference. For example, they cannot accurately distinguish whether the interference is co-channel interference concentrated in the communication center frequency band, adjacent channel interference leaked to adjacent channels, or randomly distributed broadband noise.

[0004] On the other hand, most existing methods rely on instantaneous measurement results. The randomness and non-stationarity of noise often lead to large fluctuations in detection results, which can easily result in misjudgments, causing untimely or excessive frequency switching.

[0005] In addition, some systems rely solely on a single indicator (such as power mean) when identifying noise, lacking the utilization of fluctuation characteristics such as power variance, making it difficult to effectively reflect the distribution characteristics of random noise.

[0006] It should be noted that if interference exists and frequency band switching is performed based on the instantaneous spectrum, it will lead to two problems:

[0007] 1. The interference identification is not comprehensive enough, and it can only detect some types of interference, resulting in inaccurate frequency band switching decisions;

[0008] Second, the switching process relies heavily on real-time results from nodes, which can easily lead to frequent switching between adjacent road segments, causing oscillations, instantaneous packet loss, and communication instability. Summary of the Invention

[0009] To address the aforementioned shortcomings of existing technologies, this invention provides a data transmission control method for vehicle-mounted wireless modules based on interference detection. This method effectively solves the problems of incomplete interference identification and communication instability and oscillation caused by frequent switching in existing vehicle-mounted wireless communication technologies.

[0010] To achieve the above objectives, the present invention provides the following technical solution:

[0011] This invention provides a data transmission control method for an in-vehicle wireless module based on interference detection, comprising at least:

[0012] Calculate the signal-to-noise ratio of the vehicle-mounted wireless module to determine if interference exists;

[0013] If interference exists:

[0014] Calculate the power spectral density of the received signal, identify interference sources in the frequency band, and determine the interference type, which includes:

[0015] Co-channel interference, adjacent channel interference, and random noise interference;

[0016] Obtain the center band power, average power, power variance, and historical packet loss rate for each frequency band, and select a frequency band from the remaining available frequency bands for data transmission.

[0017] Collect data on the future driving path of the vehicle and divide it into nodes. Then, introduce a graph neural network to predict the state value of each node's frequency band under the future driving path.

[0018] To develop frequency band switching solutions for future driving routes for in-vehicle wireless modules.

[0019] Furthermore, the method for determining adjacent channel interference is as follows:

[0020] Calculate the power spectral density of the upper and lower adjacent frequency bands of the target signal, and take the maximum value of the two as the maximum leakage power.

[0021] Calculate the ratio of the center frequency band power to the maximum leakage power;

[0022] The presence of adjacent channel interference is determined based on the ratio.

[0023] Furthermore, the method for determining the random noise interference is as follows:

[0024] Calculate the power variance within the center frequency band;

[0025] If the variance is greater than the random noise variance threshold, then random noise interference is determined to exist.

[0026] Furthermore, the method for selecting a frequency band from the remaining available frequency bands for data transmission is as follows:

[0027] Obtain the remaining available frequency bands and create a frequency band list;

[0028] Calculate the center band power, average power, power variance, and historical packet loss rate for each frequency band, calculate the deviation of each of the four, and then calculate the overall state value for each frequency band.

[0029] Remove states with values ​​greater than the threshold, sort them in ascending order, and build a new list of available frequency bands.

[0030] Select the frequency band corresponding to the minimum state value and record it as a candidate frequency band. This candidate frequency band is used for the current data transmission.

[0031] Furthermore, the method for predicting the state value of each node's frequency band under the future driving path using a graph neural network is as follows:

[0032] An adjacency matrix is ​​constructed based on the spatial relationships between nodes, where:

[0033] If two nodes are adjacent road segments or have spatial connections, then set the corresponding position in the adjacency matrix to 1; otherwise, set it to 0.

[0034] For each node, construct the node feature matrix X, including:

[0035] Historical interference indicators, historical average speed, historical weather data, historical time, and gridded location points;

[0036] Adjusting the weights of node eigenvectors in the node feature matrix of a graph neural network:

[0037] For each node frequency band, calculate the variance of the interference impact index in historical data and define adaptive weighting coefficients;

[0038] Introducing adaptive weight coefficients into a graph neural network to aggregate neighbor information:

[0039] After aggregation by an L-layer graph neural network, the output is a node feature matrix;

[0040] Update the parameters of the graph neural network based on the gradient of the loss function, iterating until the loss converges or the preset number of rounds is reached;

[0041] Input the nodes of the future path, the expected vehicle speed, the expected weather, and the expected time. Using the trained graph neural network parameters, the predicted state value of each node for each frequency band is output through forward propagation.

[0042] Based on the predicted state value, the candidate frequency band to be used by each node is obtained.

[0043] Furthermore, the frequency band switching scheme refers to the path frequency band scheme that minimizes the total number of state values ​​of multiple nodes and the frequency of switching in the future.

[0044] Furthermore, the method for determining co-channel interference is as follows:

[0045] Calculate the center frequency power of the received signal within the center frequency band of the target signal;

[0046] Calculate the average power of the adjacent frequency bands on both sides of the center frequency band;

[0047] Calculate the ratio of the center frequency band power to the average power;

[0048] If the ratio is greater than a preset second threshold, then it is determined that there is significant co-channel interference.

[0049] Furthermore, the method for determining the signal-to-noise ratio is as follows:

[0050] The average power of the received signal is calculated based on the target signal demodulated at the receiving end and the time length of the statistical power.

[0051] Based on the received signal and target signal actually received by the receiver, calculate the sum of the power of white noise and interference signal;

[0052] The signal-to-noise ratio is obtained by dividing the average power by the sum of the powers.

[0053] The technical solution provided by this invention has the following advantages compared with the known prior art:

[0054] The strength of interference is determined by calculating the signal-to-noise ratio of the received signal, and the type of interference (co-channel, adjacent channel, random noise) is identified through power spectral density analysis. The quality of frequency bands is evaluated based on center power, adjacent channel power, power variance, and historical packet loss rate. Then, combined with the future driving path of the vehicle, a graph neural network prediction model is constructed. By utilizing the adjacency relationship of nodes and historical interference characteristics, the frequency band state values ​​of future path nodes are predicted in advance, and an optimal frequency band switching scheme is selected to minimize transmission stability and the number of switching operations. The globally optimal switching scheme is planned in advance to avoid frequent switching, improve communication stability and anti-interference capability, and improve the data transmission quality of the vehicle wireless module in the vehicle-to-everything (V2X) environment. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0056] Figure 1 This is a flowchart of the overall method of the present invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0058] The present invention will be further described below with reference to embodiments.

[0059] Example 1 (see Figure 1 A data transmission control method for vehicle-mounted wireless modules based on interference detection, comprising at least:

[0060] During vehicle operation, the wireless module operates in a complex environment, and its signal may be affected by various interference sources such as co-channel, adjacent channel, and noise. Detecting the interference intensity helps determine the presence of interference, which is beneficial for subsequent interference source identification. Therefore:

[0061] Calculate the signal-to-noise ratio of the wireless module:

[0062]

[0063] in, Indicates the signal-to-noise ratio. This represents the received signal power (obtained by the power detection circuit before demodulation in the wireless module). It describes the average power of the target signal (the effective signal that the vehicle-mounted wireless module expects to receive and that belongs to its own communication link) measured by the vehicle-mounted wireless module at the receiving end. In vehicle-to-everything (V2X) communication, the target signal originates from the transmission of the intended communication object (the vehicle ahead, the base station, and the roadside unit). This represents the sum of the power of white noise and interference signals, describing the power of the receiver within its receiving bandwidth, excluding the target signal. It may include: thermal noise (the receiver's own electronic noise); adjacent channel interference (energy leaked from adjacent signals); co-channel interference (energy emitted by other devices on the same frequency); and random noise (wideband noise in the environment).

[0064] for :

[0065]

[0066] in, This represents the integration time window, describing the duration of the statistical power. This indicates the target signal demodulated by the receiving end;

[0067]

[0068] in, This indicates the actual received signal at the receiving end, including the target signal, white noise, and interference signals.

[0069] Therefore, it can be based on the signal-to-noise ratio. Compare with the threshold:

[0070] If signal-to-noise ratio If the value is greater than or equal to the threshold, it indicates that the current interference is low or non-existent and does not affect normal wireless communication transmission.

[0071] If signal-to-noise ratio If the value is less than the threshold, it indicates that the current interference is strong and affects normal wireless communication transmission.

[0072] Therefore, if the interference is determined to be strong based on the aforementioned signal-to-noise ratio, then in order to suppress the interference and ensure normal wireless communication transmission, it is necessary to identify the existing interference source and determine the actual type of interference. This involves:

[0073] First, calculate the power spectral density of the received signal:

[0074]

[0075] in, Indicates frequency time The power spectral density describes the energy intensity of the received signal at different frequencies. Indicates receiving signal, This represents the Fourier transform, used to convert time-domain signals to the frequency domain;

[0076] Then the source of interference can be identified, including:

[0077] In wireless communication, co-channel interference is one of the most common types of interference. Its danger lies in the fact that the interfering signal and the target signal share the same center frequency band, making it impossible for the receiver to distinguish between them using filters. The presence of abnormal energy concentrated in the center frequency band is identified by detecting the center frequency band power, and the center frequency band power is calculated as follows:

[0078]

[0079] in, This represents the center frequency power obtained within the center bandwidth. This indicates the center frequency of the target signal of the vehicle-mounted wireless module, which is used to determine the reference frequency point for power calculation and to determine whether it is within the center frequency range. An energy anomaly occurred. Indicates the signal bandwidth, limits the integration range, and prevents interference energy from mixing into adjacent frequencies. The center frequency band is ( , );

[0080] Compare the center band power with the center band power threshold. If it is greater than the center band power threshold... This indicates that its energy superimposed on the target signal, resulting in a significant increase in the power of the center frequency band. However, it cannot be ruled out that the power of the surrounding adjacent frequency bands is also high. Therefore, it is difficult to determine whether the frequency band in question is subject to co-channel interference.

[0081] Furthermore, if the power in the center frequency band is higher than that in the adjacent frequency band, it can be inferred that the energy is abnormally concentrated, indicating that there may be co-channel interference. Therefore, to further determine whether it affects the adjacent frequency band, calculations are performed using... The center frequency band (the bandwidth where the target signal is located) The adjacent frequency bands on both sides (upper adjacent frequency band: , Next adjacent frequency band: , Average power:

[0082]

[0083] in, It represents the average power of the adjacent frequency bands on both sides of the center frequency band, reflecting the degree of interference on the adjacent frequency bands;

[0084] Furthermore, based on the center frequency band power With average power The ratio of the power threshold to the center frequency band is considered. If this ratio is greater than the power threshold, it indicates that the interference is mainly concentrated in the center frequency band, while adjacent frequency bands are unaffected, thus classifying it as co-channel interference. Conversely, if the ratio is less than the power threshold, it is not co-channel interference. It is worth noting that the power threshold for the center frequency band is used to determine the interference level. With average power The ratio judgment can be based on the average power. As a reference, avoid relying solely on This can lead to misjudgments of high-noise environments, thus hindering the improvement of judgment and recognition accuracy.

[0085] Therefore, if it is not co-channel interference, then adjacent-channel interference needs to be determined, which leads to:

[0086] The maximum leakage power in the upper and lower adjacent channel bands of the center frequency band is measured. :

[0087]

[0088] in, This indicates that the maximum of the two integral values ​​is taken, selecting the most severe adjacent channel interference. Indicates the adjacent frequency bands on the center frequency standard ( , Integrate and accumulate the power spectral density within the range. This indicates the adjacent frequency bands under the center frequency standard ( , The power spectral density within the range is integrated and accumulated.

[0089] If the maximum leakage power If the power exceeds the leakage power threshold, it usually indicates that the interference signal has leaked into an adjacent frequency band. Therefore, the power in the center frequency band should be calculated. With maximum leakage power The ratio quantifies the severity of adjacent channel interference relative to the center signal. If the ratio is less than the leakage threshold, it indicates that there is severe adjacent channel interference; otherwise, the adjacent channel interference is slight and can be ignored.

[0090] In this embodiment, if the detection results, after passing the above technical solution, indicate that the interference is neither co-channel nor adjacent-channel, then it is necessary to further determine random noise interference and calculate the power variance of the center frequency band.

[0091]

[0092] in, This represents the power variance in the center frequency band, used to quantify power fluctuations within the center frequency band. The smaller the variance, the smoother the power distribution, indicating the absence of significant concentrated peaks. Indicates the average power in the center frequency band. The center value is used to measure power; variance calculation requires subtracting the average power of the center frequency band.

[0093] So, Compare with the random noise variance threshold; if it is greater than the random noise variance threshold... If the noise distribution is more random and the energy fluctuations are large within the frequency band, resulting in a significant increase, it can be determined as random noise interference.

[0094] Furthermore, if co-channel interference, adjacent channel interference, or random noise interference is identified through the above analysis, then the optimal frequency band is selected for data transmission to ensure normal data transmission. The method for determining the optimal frequency band is as follows:

[0095] Obtain the remaining available frequency bands and create a frequency band list;

[0096] Calculate the center band power, average power, power variance, and historical packet loss rate for each frequency band. Calculate the deviation values ​​of the four factors separately, and then calculate the state value of each frequency band together. :

[0097]

[0098] in, , These represent the corresponding thresholds. , , as well as These represent the corresponding weighting coefficients.

[0099] By calculating the difference between each data point and performing dimensionless processing (such as min-max or z-score methods), the state value of the frequency band is calculated. Bands with state values ​​greater than a set threshold are removed, and the bands are sorted from smallest to largest to build a new list of available frequency bands.

[0100] Select the frequency band corresponding to the minimum state value (reflecting the lower the interference, the better the power stability, and the higher the historical reliability of the frequency band) and record it as the optimal candidate frequency band. Use this candidate frequency band for the current data transmission.

[0101] It should be noted that the vehicle-mounted wireless module travels with the vehicle, and the vehicle will move to different locations. The interference associated with that frequency band may vary at different locations. Therefore, to predict the stability of the frequency band used by the vehicle-mounted wireless module in the future, it is necessary to predict the frequency band status values ​​at different locations along the vehicle's route, so as to pre-plan the frequency band used by the vehicle-mounted wireless module in the future. This leads to the following:

[0102] To obtain the vehicle's future driving path, the system divides the path into nodes at fixed intervals (e.g., every 100 meters), with each node representing a location point along the path. ;

[0103] An adjacency matrix is ​​constructed based on the spatial relationships between nodes, where:

[0104] If two nodes are adjacent road segments (physically continuous or connected road segments, such as road segment A and road segment B being geographically connected end to end, and a vehicle immediately enters B after traveling from A, then A and B are considered adjacent road segments) or have direct spatial connections (including: geographical proximity: two road segments are very close, such as adjacent lanes, parallel roads, ramps and main roads, where wireless signals or interference may affect each other; shared environment: two nodes are in the same area, such as within the coverage area of ​​the same base station or in the same tunnel, where there may be frequency band interference; traffic flow intersection: two nodes are not directly connected, but have common traffic flow intersections, such as overpasses or junctions, where vehicles and interference signals may be indirectly coupled), then the corresponding position in the adjacency matrix is ​​set to 1; otherwise, it is set to 0. This allows the spatial correlation of interference between neighboring nodes to be considered in subsequent training of the graph neural network.

[0105] By aggregating spatial neighborhoods, the model can identify patterns in how interference spreads rapidly from a point to the neighborhood and affects certain frequency bands, thus enabling proactive avoidance measures at nodes that have not yet been reached.

[0106] For each node, construct the node feature matrix X, including:

[0107] Historical interference metrics (center band power, average power, power variance, historical packet loss rate) All need to be normalized to Examples of weather data include: min-max data, historical average speed (normalized to 0-1 using min-max), and historical weather data (sunny, rainy, foggy, and snowy weather, each weather type mapped to a binary vector using one-hot encoding, e.g., sunny weather). Rainy days are Foggy days are Snowy days are ), historical time (time stamp of the corresponding node interference index, such as historical time 14:30, to prevent the training from being affected by the excessive time scale, normalized to 0-1 by dividing min-max by the maximum value) and the gridded position points;

[0108] To improve prediction accuracy, the weights of node feature vectors in the node feature matrix of the graph neural network are adaptively adjusted to address different nodes, frequency bands, and historical data stability.

[0109] For each node frequency band, calculate the variance of the interference impact index (which can be center band power, average power, power variance, or historical packet loss rate) in the historical data. The smaller the variance, the higher the interference stability of the node in that frequency band. Then, an adaptive weighting coefficient is defined. ;

[0110] Aggregating neighbor information using graph neural networks:

[0111]

[0112] in, express The node feature matrix of the layer is used for prediction in the next layer. This indicates multiplying by an adaptive weight coefficient by the node dimension. Indicates the first The layer's weight matrix represents a linear transformation, adjusts the feature dimensions, and learns the mapping relationship between node features and output predictions. express The degree matrix is ​​used for neighborhood normalization to prevent eigenvalues ​​from being too large or too small. Activation functions, such as RelU, introduce non-linearity, enabling the model to learn complex mappings. Indicates the first Layer node feature matrix, dimensions , Indicates the number of nodes. Indicates the first The feature dimension of the layer, where, When, X is the input node feature matrix, This represents the adjacency matrix with added self-loops, which is the original adjacency matrix plus the identity matrix. This indicates that the adjacency matrix is ​​symmetrically normalized.

[0113] In this embodiment, during adaptive adjustment, adaptive weight coefficients can be calculated based on the variance of different nodes in historical frequency band data, and the node feature vector of the node can be adjusted by weighting. This allows nodes with more stable frequency band states to have a greater influence during propagation, while nodes with larger frequency band fluctuations are appropriately weakened, thereby improving the overall accuracy and robustness of prediction.

[0114] Specifically, a 2-3 layer graph neural network is used to fuse the node feature matrix with neighbor information to predict the future frequency band status of nodes. This makes the prediction not only based on the historical interference of a single node, but also combined with the surrounding environment and neighboring node information.

[0115] After aggregation by an L-layer graph neural network, the output node feature matrix is ​​obtained. , This represents the output dimension of each node, corresponding to the predicted state value for each frequency band:

[0116]

[0117] in, Indicates the first The node, the first Predicted state values ​​for each frequency band;

[0118] Construct the loss function:

[0119]

[0120] in, Represents the loss function, which is achieved by minimizing... To adjust the model parameters, This represents the number of samples (e.g., the number of nodes multiplied by the number of frequency bands). Represents the historical true state value. Indicates the number of frequency bands;

[0121] The predicted state value is compared with the historical true state value to measure the deviation between the predicted value and the true value;

[0122] The parameters of the graph neural network, i.e. the weight matrix, are updated based on the gradient of the loss function, and the iteration continues until the loss converges or the preset number of rounds is reached.

[0123] Therefore, we can input the nodes of the future path, the expected vehicle speed, the expected weather, and the expected time (current vehicle time + the ratio of distance to speed to reach the future path location). Using a trained weight matrix, the graph neural network propagates forward and outputs the predicted state value for each node and each frequency band. ;

[0124] Based on the predicted state value, the candidate frequency bands that each node can use are obtained (frequency bands whose predicted state value is less than the threshold).

[0125] It should be noted that in traditional frequency band selection processes, decisions are often made independently node-by-node based on real-time measurements. That is, when a vehicle reaches a certain road segment node, the optimal candidate frequency band is selected based on the real-time interference situation at that node. However, this approach has significant shortcomings:

[0126] Due to the strong volatility of the wireless environment, the optimal candidate frequency bands of adjacent nodes may be different, or even alternate. This causes vehicles to frequently switch frequency bands in continuous road segments, resulting in oscillations. These oscillations not only increase the switching overhead of the system, but also easily cause instantaneous packet loss and communication instability.

[0127] Therefore, this embodiment predicts the frequency band state of each node in the future path based on a graph convolutional network, obtaining not only the state value of a single node but also the frequency band evolution trend of consecutive nodes in the future. Based on this, the system can perform global optimization at the path level, pre-planning a minimum handover path according to the optimal node for each node. This involves selecting the frequency band handover scheme with the least overall interference and the fewest handovers across multiple nodes in the future (not just considering the optimal frequency band of a single node, but also considering the situation of consecutive nodes, selecting a frequency band (determined from the available frequency bands) that minimizes the total interference (i.e., the total number of state values) and the number of handovers in the future path). In this way, even if the prediction result for a certain node shows a short-term better frequency band, the system can maintain a long-term stable frequency band based on the overall trend, avoiding frequent handovers (prone to oscillations) caused by local optima. Based on the determined candidate frequency bands for future optimality, the system further ensures communication stability.

[0128] Also provided:

[0129] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method.

[0130] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described above. For specific details, refer to the method described above.

[0131] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for controlling data transmission of a vehicle-mounted wireless module based on interference detection, characterized in that, The method comprises the following steps: calculating the signal-to-noise ratio of the vehicle-mounted wireless module to determine whether there is interference; if there is interference: calculating the power spectral density of the received signal to identify the interference source of the frequency band and determine the type of interference, which includes: co-channel interference, adjacent channel interference and random noise interference; obtaining the center frequency band power, average power, power variance and historical packet loss rate of each frequency band, and selecting a frequency band from the remaining available frequency bands for data transmission; collecting the future driving path division nodes of the vehicle, and introducing a graph neural network to predict the state value of each node frequency band in the future driving path; constructing a frequency band switching scheme for the future driving path for the vehicle-mounted wireless module; the judgment method of the adjacent channel interference is as follows: calculating the power spectral density of the upper adjacent frequency band and the lower adjacent frequency band of the target signal, and taking the maximum value of the two as the maximum leakage power; calculating the ratio of the center frequency band power to the maximum leakage power; judging whether there is adjacent channel interference according to the ratio.

2. The method of claim 1, wherein, the judgment method of the random noise interference is as follows: calculating the power variance in the center frequency band; if the variance is greater than the random noise variance threshold, it is determined that there is random noise interference.

3. The method of claim 1, wherein, the method for selecting a frequency band from the remaining available frequency bands for data transmission is as follows: obtain the remaining available frequency bands and establish a frequency band list; calculate the center frequency band power, average power, power variance and historical packet loss rate of each frequency band, respectively calculate the deviation value of the four, and comprehensively calculate the state value of each frequency band: eliminate the state value greater than the threshold value, sort them from small to large and construct a new available frequency band list; select the frequency band corresponding to the smallest state value as the candidate frequency band, which is used for current data transmission.

4. The method of claim 1, wherein, the method for introducing a graph neural network to predict the state value of each node frequency band in the future driving path is as follows: establish an adjacency matrix according to the spatial relationship between nodes, wherein: if two nodes are adjacent road segments or have spatial contact, the corresponding position in the adjacency matrix is 1, otherwise it is 0; for each node, construct a node feature matrix X, including: historical interference indicators, historical average speed, historical weather data, historical time and grid-based location points; adjust the weight of the node feature vector in the node feature matrix in the graph neural network: for each node frequency band, calculate the variance of the interference influence index in the historical data, and define an adaptive weight coefficient; introduce the adaptive weight coefficient into the graph neural network to aggregate neighbor information: after L-layer graph neural network aggregation, output the node feature matrix; update the graph neural network parameters according to the loss function gradient, and iterate until the loss converges or reaches the preset number of rounds; input the nodes of the future path, the predicted speed, the predicted weather and the predicted time, use the trained graph neural network parameters, and output the predicted state value of each node and each frequency band through forward propagation; obtain the candidate frequency band used by each node according to the predicted state value.

5. The method of claim 1, wherein, The frequency band switching scheme refers to: in the future multiple nodes, the path frequency band scheme with the minimum total amount of state values and the minimum switching frequency of multiple nodes.

6. The method of claim 1, wherein, the judgment method of the co-channel interference is as follows: calculating the center frequency band power of the received signal in the center frequency band of the target signal; calculating the average power of the adjacent frequency bands on both sides of the center frequency band; calculating a ratio of the center band power to the average power; if the ratio is greater than a preset second threshold, determining that there is obvious co-channel interference.

7. The method of claim 1, wherein, The method for determining the signal-to-noise ratio comprises: calculating an average power of the received signal power according to a target signal demodulated by the receiving end and a time length of the statistical power; calculating a power sum of the white noise and the interference signal according to the received signal actually received by the receiving end and the target signal; the ratio of the average power to the power sum is the signal-to-noise ratio.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor, when executing the computer program, implements the steps of the method in any one of claims 1 to 7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method in any one of claims 1 to 7.

Images