Vehicle early lane changing intention prediction method based on improved KAN network

Abstract

The invention discloses a vehicle early lane changing intention prediction method based on an improved KAN network. The method comprises the steps of collecting target vehicle driving data, dividing positive and negative samples according to the target vehicle driving data, obtaining a lane changing data set according to the divided positive and negative samples, and constructing a vehicle early lane changing intention prediction model by using the lane changing data set and performing training. Inputting continuous driving data of a to-be-detected target vehicle collected in real time into the trained vehicle early lane change intention prediction model for processing to obtain a lane change probability of the to-be-detected target vehicle, and finally comparing the lane change probability with a preset lane change probability threshold to obtain whether the to-be-detected target vehicle changes a lane or not. According to the method, the KAN network based on the penalty lane changing weight loss function is adopted to process the lane changing data set, the perception ability of the unmanned vehicle to the environment in the mixed traffic environment is effectively improved, sufficient response time is provided for the unmanned vehicle, and therefore the driving safety of the unmanned vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation, and particularly relates to a method for predicting early lane-changing intention of vehicles based on an improved KAN network. Background Art

[0002] With the rapid progress of technology, driverless technology is booming with each passing day. However, for a long time in the foreseeable future, a complex situation will be formed on traffic roads where manned vehicles and driverless vehicles share road rights together. In such a background, accurately identifying the lane-changing behavior of manned vehicles or driverless vehicles will undoubtedly become a crucial means to ensure road traffic safety.

[0003] Currently, in-vehicle sensor technology has gradually become perfect. New energy vehicles at home and abroad, whether high-end luxury models or popular household cars, are mostly equipped with advanced sensor devices such as vision and radar. These sensors are like the "eyes" and "ears" of the vehicle, and can obtain the environmental information around the driverless vehicle in real time and accurately. Whether it is the driving speed and distance of the vehicle ahead, or the traffic flow in the surrounding lanes, they can be clearly captured. At the same time, the deep integration of big data and the transportation field has also played a huge role. The driving behavior data of historical vehicles on the road has been comprehensively collected and sorted, and driverless vehicles can easily obtain these valuable data resources. These rich road traffic information provides a solid data support for in-depth data analysis and accurate vehicle behavior prediction.

[0004] However, the current existing technologies have obvious deficiencies. Most research and applications mainly focus on pursuing the accuracy of predicting vehicle lane changes, but seriously ignore the key factor of lane-changing time. In the actual road traffic scenario, the advance amount of lane-changing time plays a crucial role. When the prediction time advance amount is too small, the reaction time of the driverless vehicle will be greatly compressed. On a high-speed road, the driverless vehicle is suddenly informed that the vehicle ahead is about to change lanes, but there is not enough time to make a reasonable reaction, which is very likely to lead to serious traffic accidents such as rear-end collisions and collisions, greatly affecting the driving safety of the driverless vehicle. The data related to lane-changing time is of uneven quality, with inconsistent formats and standards, insufficient representativeness, untimely updates, difficult to reflect new traffic changes, and there are also privacy and security risks, and there is a risk of data leakage. Summary of the Invention

[0005] Aiming at the deficiencies in the background art, the present invention proposes a method for predicting early lane-changing intention of vehicles based on an improved KAN network, which solves the technical problem that the prior art ignores vehicle early lane-changing prediction based on lane-changing time.

[0006] 1. A method for predicting vehicle's early lane-changing intention based on improved KAN network S1. Collect driving data of several target vehicles through the on-board sensors of the unmanned vehicle, and divide the driving data of all target vehicles into positive samples and negative samples.

[0007] S2. Preprocess the driving data according to the divided positive samples and negative samples to obtain time series data of several samples, and aggregate all the time series data to obtain a lane change data set.

[0008] S3. Construct a vehicle early lane-changing intention prediction model, input the lane-changing data set into the vehicle early lane-changing intention prediction model using a penalized lane-changing weight loss function for training, and obtain a trained vehicle early lane-changing intention prediction model.

[0009] S4. Collect continuous driving data from the previous frames to the current frame of several target vehicles to be tested in real time, and input the data into the trained vehicle early lane change intention prediction model for processing to obtain the lane change probability of each target vehicle to be tested in real time. The lane change probability of each target vehicle to be tested is compared with a preset lane change probability threshold to obtain in real time whether each target vehicle to be tested has changed lanes.

[0010] Based on the real-time information obtained on whether all target vehicles under test change lanes, the unmanned vehicle makes corresponding driving adjustments, such as changing lanes and slowing down.

[0011] The step S1 is specifically as follows: S11. Collecting driving data of each frame of several target vehicles through the on-board sensor of the unmanned vehicle.

[0012] S12. Classify the target vehicles into lane-changing vehicles and non-lane-changing vehicles according to the driving data.

[0013] S13. For vehicles that change lanes, the continuous driving data consisting of the target vehicle's lane change completion time and N-1 frames before the lane change completion time is used as a positive sample, and the continuous driving data from 2N-1 frames before the lane change completion time to the previous N frames is used as a negative sample; for vehicles that do not change lanes, any continuous N frames of driving data are used as negative samples.

[0014] The target vehicle is the vehicle in front of the unmanned vehicle on the left or right adjacent lane. The driving data includes speed, left longitudinal distance, right longitudinal distance and lateral distance; the left longitudinal distance is the longitudinal distance between the target vehicle and the vehicle in front of the target vehicle on the left adjacent lane; the right longitudinal distance is the longitudinal distance between the target vehicle and the vehicle in front of the target vehicle on the right adjacent lane; the lateral distance is the distance between the center point of the target vehicle and the center line of the lane. The lane change completion time is the time when the center point of the target vehicle coincides with the lane line on one side of the target vehicle.

[0015] The specific steps of step S2 are as follows: S21. Divide the samples into time series windows in the way that the sliding window size is M frames and the sliding step is one frame, obtaining N - M + 1 time series. The N - M + 1 obtained time series are combined to obtain the time series data of a sample.

[0016] S22. Process in the same way as in step S21 to obtain the time series data of several samples. The time series data of all samples are aggregated to obtain a lane - changing dataset.

[0017] The vehicle early lane - changing intention prediction model in step S3 is a KAN network.

[0018] The penalty lane - changing weight loss function in step S3 is set according to the following formula: L i + (p t + ) = -∑ x=1 N-M+1 αlog(p ix + ) L j - (p t - ) = -∑ y=1 N-M+1 log(1 - p jy - ) α = e (-max(0，d-Φ(epoch-1)-γ)) Φ(epoch - 1)=1 / k∑ f=1 k G f epoch-1 Among them, both i and x represent indices; L i + () is the loss function of the i - th positive sample; p ix + represents the lane - changing probability predicted by the x - th time series in the i - th positive sample of the vehicle early lane - changing intention prediction model; α represents the weight; both j and y represent indices; L j - () is the loss function of the j - th negative sample; p jy -Denotes the lane - changing probability obtained from the y - th time series in the j - th negative sample of the vehicle early lane - changing intention prediction model; e represents a constant; max( ) represents taking the maximum value; d represents the difference between the index of the last frame of the x - th time series and the index of the lane - changing completion moment; epoch represents the index of the training round and also represents the epoch - th training; Φ(epoch - 1) represents the mean of the lane - changing time differences of all positive samples in the (epoch - 1) - th training round; γ represents a hyperparameter; k represents the total number of positive samples; f represents an index; G f epoch-1 Denotes the lane - changing time difference of the f - th positive sample in the (epoch - 1) - th training round.

[0019] Specifically, the lane - changing time difference is the difference between the index of the time series when the first lane - changing probability in each positive sample is not less than the preset lane - changing probability threshold and the index of the last time series.

[0020] When comparing the lane - changing probability of the target vehicle to be measured in step S4 with the preset lane - changing probability threshold, it is set according to the following formula to obtain whether the target vehicle to be measured changes lanes: f(p q ) = 1, p q ≥P th f(p q ) = 0, p q <P th Wherein, f(p q ) represents whether the q - th target vehicle to be measured changes lanes; p q represents the lane - changing probability predicted by inputting the continuous driving data of the q - th target vehicle to be measured into the trained vehicle early lane - changing intention prediction model; 1 represents changing lanes; 0 represents not changing lanes; P th represents the preset lane - changing probability threshold.

[0021] Second, a computer device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps of the above - mentioned method are implemented.

[0022] Third, a computer - readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above - mentioned method are implemented.

[0023] The innovation of the present invention lies in using the KAN network with a penalty - type lane - changing weight loss function to process the driving data of the target vehicle, which realizes improving the prediction time advance of the target vehicle's lane - changing while ensuring the correct rate of predicting the vehicle's lane - changing behavior, and has an advantage with stronger practical significance compared with the traditional method of predicting lane - changing behavior.

[0024] The beneficial effects of the present invention are as follows: The method of the present invention processes the collected driving data of the target vehicle through the KAN network based on the penalty lane-changing weight loss function. While ensuring the accuracy of predicting the lane-changing behavior of the vehicle, it improves the prediction time lead of the target vehicle's lane change, increases the reaction time to changes in the surrounding environment, has stronger practical significance compared with the traditional method of predicting lane-changing behavior, and can better ensure traffic safety in the mixed traffic environment of manned vehicles and driverless vehicles. Brief Description of the Drawings

[0025] Figure 1 It is a flowchart of the method of the present invention.

[0026] Figure 2 It is a schematic flow diagram of the vehicle early lane-changing intention prediction model of the present invention adopting the penalty lane-changing weight loss function.

[0027] Figure 3 It is a schematic diagram of the driverless vehicle observing the lane-changing behavior of the target vehicle in Embodiment 1.

[0028] Figure 4 It is a schematic diagram of the target vehicle and surrounding vehicles in Embodiment 1.

[0029] Figure 5 It is a lane-changing probability map predicted by the vehicle early lane-changing intention prediction model in Embodiment 1.

[0030] Figure 6 It is a lane-changing probability map predicted by the vehicle early lane-changing intention prediction model in Embodiment 2 and Comparative Examples 1-3. Detailed Embodiments

[0031] The following describes the present invention in more detail with reference to the drawings and embodiments. However, the present invention is not limited thereto. For those of ordinary skill in the art in this technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as within the protection scope of the present invention. The content not described in detail in this specification belongs to the prior art well-known to those of ordinary skill in the art.

[0032] Embodiment 1

[0033] As Figure 1 shown, the vehicle early lane-changing intention prediction method of this embodiment includes the following steps: S1. Collect a number of driving data of the target vehicle through the in-vehicle sensors of the driverless vehicle, and divide the positive samples and negative samples according to all the driving data of the target vehicle.

[0034] S11. Collect the driving data of each frame of a number of target vehicles through the in-vehicle sensors of the driverless vehicle; In this embodiment, the driverless vehicle collects the driving data of the target vehicle on a section of the expressway in a certain city, from a two-way 10-lane section on the west side to a two-way 3-lane section on the east side, with a length of 427 m. The total collection time is 5 minutes and 33 seconds, and the collection frame rate is 24 frames per second.

[0035] S12. Classify the target vehicles into lane-changing vehicles and non-lane-changing vehicles according to the driving data.

[0036] S13. For the lane-changing vehicles, use the continuous driving data composed of the lane-changing completion moment of the target vehicle and the N - 1 frames before the lane-changing completion moment as positive samples, and use the continuous driving data from the 2N - 1 frames before to the N frames before the lane-changing completion moment as negative samples.

[0037] For the non-lane-changing vehicles, use any continuous N frames of driving data as negative samples. In specific implementation, the label of the positive sample is 1, and the label of the negative sample is 0. In this embodiment, N is taken as 192 frames, corresponding to 8 seconds of driving data.

[0038] As Figure 3 and Figure 4 shown, the target vehicle is the vehicle in front on the adjacent lane to the left or right of the driverless vehicle; in specific implementation, only consider the nearest vehicle in front on each adjacent lane, and the target vehicles of the driverless vehicle include the nearest vehicle in front on the adjacent lane to the left and the nearest vehicle in front on the adjacent lane to the right. As Figure 2 shown, at the current moment T0, one of the target vehicles of the driverless vehicle is in the left lane. When it reaches the moment of T f the target lane has completely merged with the lane of the driverless vehicle.

[0039] The driving data includes speed, left longitudinal distance, right longitudinal distance, and lateral distance; the left longitudinal distance is the longitudinal distance between the target vehicle and the vehicle in front on the adjacent lane to the left of the target vehicle; the right longitudinal distance is the longitudinal distance between the target vehicle and the vehicle in front on the adjacent lane to the right of the target vehicle; the lateral distance is the perpendicular distance between the center point of the target vehicle and the center line of the lane where it is located. The lane-changing completion moment is the moment when the center point of the target vehicle coincides with the lane line on one side of the target vehicle. In specific implementation, when the target vehicle is the vehicle in the leftmost lane of the lane, the left longitudinal distance is 0, indicating that the target vehicle cannot change lanes to the left; when the target vehicle is the vehicle in the rightmost lane of the lane, the right longitudinal distance is 0, indicating that the target vehicle cannot change lanes to the right. When the target vehicle merges from the ramp into the main road, it is regarded as being in the rightmost lane. In specific implementation, all distance measurements are made according to the center point of the front side of each vehicle.

[0040] In this embodiment, a total of 3197 target vehicles are collected, including 782 positive samples (lane-changing vehicles) and 2415 negative samples (non-lane-changing vehicles).

[0041] S2. Preprocess the driving data according to the divided positive samples and negative samples to obtain time series data of several samples, and aggregate all the time series data to obtain a lane change data set.

[0042] S21. Divide the sample into time series windows by using a sliding window size of M frames and a sliding step size of one frame to obtain N-M+1 time series. The obtained N-M+1 time series are combined to obtain time series data of one sample.

[0043] In the specific implementation, the time series window division is performed according to the time frame from the front to the back. In this embodiment, the time series window division of the sample is performed in a manner of sliding window size of 72 frames (3 seconds in total) and sliding step length of one frame, and 121 time series are obtained. The obtained 121 time series are combined to obtain the time series data of one sample.

[0044] S22. The same method as step S21 is adopted to obtain time series data of several samples, and the time series data of all samples are aggregated to obtain a lane change data set.

[0045] S3. Construct a vehicle early lane-changing intention prediction model, input the lane-changing data set into the vehicle early lane-changing intention prediction model using a penalized lane-changing weight loss function for training, and obtain a trained vehicle early lane-changing intention prediction model.

[0046] The vehicle's early lane-changing intention prediction model is the KAN (Kolmogorov-Arnold) network.

[0047] The design inspiration of the KAN network used in this embodiment comes from the Kolmogorov-Arnold theorem, which shows that a multivariate continuous function can be expressed as a composite of a finite number of single-variable continuous functions and binary addition operations. Based on this theory, the KAN network introduces a learnable activation function on the network edge (i.e., weight) to improve the flexibility and expressiveness of the model.

[0048] The KAN network replaces the elements in the weight parameter matrix in the multilayer perceptron (MLP) with nonlinear activation functions with variable parameters. These activation functions are usually learnable rather than fixed, which makes the KAN network more flexible in processing data.

[0049] The core of the KAN network is to use spline functions as activation functions, which replace the linear weight parameters in traditional neural networks. The introduction of spline functions enables the KAN network to simulate complex functions with fewer parameters, enhancing the interpretability of the model.

[0050] The KAN network has a fully connected structure, similar to the MLP, but its weights are represented as spline functions rather than real-valued weights. This design allows KAN to adopt learnable activation functions at the network edges, providing extremely high flexibility.

[0051] In this embodiment, the KAN network adopts the Kolmogorov-Arnold representation theorem, and the formula of the Kolmogorov-Arnold representation theorem is set as follows: f(x1,…,…,x n )=∑ w=1 2n+1 ξ w (∑ r=1 n ψ w,r (x r )) Where x i ∈[0,1], i = 1, 2, …, n; f(x1,…,…,x n ) represents an n-ary continuous function composed of n basis functions; ξ w represents a univariate continuous function; ψ w,r also represents a univariate continuous function; x r represents the independent variable of each basic univariate continuous function ψ w,r ; w and r are both indices; n represents the arity of the continuous function; x1 and x n both represent the independent variables of the continuous function f(x1,…,…,x n ). The Kolmogorov-Arnold representation theorem states that a multivariate continuous function on a bounded region can be represented as the sum of a series of univariate basis functions combined.

[0052] In this embodiment, KAN adopts two fully connected layers, uses a learnable activation function to replace the weight parameters at the network edges, and finally outputs the predicted lane-changing probability through the Sigmoid function. The lane-changing probability and the label are input into the improved loss function to calculate the loss function value of the current training round and the mean value of the lane-changing time differences of all positive samples in the current training round, and the mean value of the lane-changing time differences of all positive samples in the current training round is used as the weight for calculating the loss function of positive samples in the next training round.

[0053] As Figure 2 shown, the penalty lane-changing weight loss function is set as follows: L i + (p t + )=-∑ x=1 N-M+1 αlog(pix + ) L j - (p t - )=-∑ y=1 N-M+1 log(1 - p jy - ) α = e (-max(0，d-Φ(epoch-1)-γ)) Φ(epoch - 1)=1 / k∑ f=1 k G f epoch-1 where i and x both represent indices; L i + () is the loss function of the i-th positive sample; p ix + represents the lane-changing probability predicted for the x-th time series in the i-th positive sample of the vehicle early lane-changing intention prediction model; α represents the weight; j and y both represent indices; L j - () is the loss function of the j-th negative sample; p jy - represents the lane-changing probability predicted for the y-th time series in the j-th negative sample of the vehicle early lane-changing intention prediction model; e represents a constant; max( ) represents taking the maximum value; d represents the difference between the index of the last frame of the x-th time series and the index of the lane-changing completion moment; epoch represents the index of the training round and also represents the epoch-th training; Φ(epoch - 1) represents the mean value of the lane-changing time differences of all positive samples in the (epoch - 1)-th training round; γ represents a hyperparameter; k represents the total number of positive samples; f represents an index; G f epoch-1 represents the lane-changing time difference of the f-th positive sample in the (epoch - 1)-th training round.

[0054] Specifically, the lane-changing time difference is the difference between the index of the time series when the first lane-changing probability in each positive sample is not less than the preset lane-changing probability threshold and the index of the last time series (including the time series at the lane-changing completion moment). In this embodiment, the preset lane-changing probability threshold is taken as 0.7.

[0055] For each positive sample, the time series window is partitioned from the front to the back of the time frames, and the time series data is also input into the model for training from the front to the back of the time frames. The index of the last frame in the time series is used as the index of the time series; the index of the time series from 191 frames before the lane change completion moment to 120 frames before the lane change completion moment is 120, the index of the time series from 190 frames before the lane change completion moment to 119 frames before the lane change completion moment is 119, and so on. The index of the time series from 71 frames before the lane change completion moment to the lane change completion moment frame is 0. The time series of the time series data is input into the model for training in the index order of 120, 119, 118, …, 0.

[0056] Therefore, each sample has 121 time series. When input into the model, 121 consecutive lane change probabilities will be obtained. When the probability value is not less than the preset lane change probability threshold, a lane change occurs. The difference is calculated between the index of the first time series among the 121 consecutive lane change probabilities that is not less than the lane change probability threshold and the index 0 of the last time series. For example, for a sample, when the lane change probability of the time series with index 83 is greater than the lane change probability threshold for the first time, the lane change time difference is 83 - 0 = 83.

[0057] In specific implementation, when training the early lane change intention prediction model of a vehicle, the closer the driving data is to the lane change completion moment, the easier it is for the model to learn the pattern between the driving data and the lane change behavior. The farther the driving data is from the lane change completion moment, the less significant the correlation is, and the more difficult it is for the model to learn the pattern from it.

[0058] S4. Continuously collect the driving data of several target vehicles to be measured from several frames before the current moment to the current moment frame in real time, and input it into the trained early lane change intention prediction model of the vehicle for processing to obtain the lane change probability of each target vehicle to be measured in real time. The lane change probability of each target vehicle to be measured is compared with the preset lane change probability threshold to obtain in real time whether each target vehicle to be measured has a lane change.

[0059] In specific implementation, the several target vehicles to be measured collected in real time are the target vehicles in front of the advancing direction of the driverless vehicle, and the specific number of target vehicles depends on the actual situation. The range of several frames before is 47 frames (including the current moment frame, a total of 2 seconds) - 239 frames (including the current moment frame, a total of 10 seconds).

[0060] When comparing the lane change probability of the target vehicle to be measured with the preset lane change probability threshold, it is set according to the following formula to obtain whether the target vehicle to be measured has a lane change: f(p q ) = 1, p q ≥P th f(p q) = 0, p q <P th where f(p q ) represents whether the q-th target vehicle to be measured has changed lanes; p q represents the lane-changing probability predicted by inputting the continuous driving data of the q-th target vehicle to be measured into the trained vehicle early lane-changing intention prediction model; 1 represents a lane change; 0 represents no lane change; P th represents the preset lane-changing probability threshold.

[0061] According to whether all target vehicles to be measured have changed lanes obtained in real time, the driverless vehicle makes corresponding driving adjustments, such as making lane changes and decelerating operations.

[0062] Such as Figure 5 shown in the result graph of the lane-changing probability predicted by the vehicle early lane-changing intention prediction model in this embodiment on the lane-changing data set collected and processed on a certain city's expressway. It can be seen from the graph that the target vehicle's lane change has been well predicted in the time series corresponding to the first 111 frames of the time series containing the lane-changing completion moment.

[0063] Embodiment 2

[0064] In this embodiment, the same method as steps S1 - S3 in Embodiment 1 is used to conduct experiments in the existing domestic and foreign traffic data set SQM1. In the SQM1 data set, the road sections where the road narrows are at 90 meters to 150 meters and 210 meters to 270 meters in the east-west direction. These driving data are deleted to avoid mislearning of the vehicle early lane-changing intention prediction model. There are 1041 target vehicles in the SQM1 data set, including 722 positive samples (lane-changing vehicles) and 319 negative samples (non-lane-changing vehicles). In this embodiment, the preset lane-changing probability threshold is taken as 0.9.

[0065] The conditions for dividing the lane-changing data set are: in this embodiment, N is taken as 120 frames, which is a total of 5 seconds of driving data. The time series window is divided in the order from the front to the back of the time frame. The sliding window size M used is 72 frames (a total of 3 seconds), and the sliding step is one frame to divide the samples into time series windows, obtaining 49 time series. The 49 obtained time series are combined to obtain the time series data of a sample.

[0066] The lane-changing probability result of the vehicle early lane-changing intention prediction model in this embodiment on the SQM1 data set is as Figure 6 shown in the curve of the KAN penalty lane-changing weight loss function in. It can be seen that 2 seconds before the lane-changing completion moment, the vehicle early lane-changing intention prediction model in this embodiment can predict a very high lane-changing probability.

[0067] Comparative Example 1

[0068] This comparative example conducts experiments in the existing domestic and foreign traffic dataset SQM1 using the same method, dataset, and lane-changing dataset division conditions as in Example 2. However, in this comparative example, the loss function is modified to the conventional cross-entropy loss function.

[0069] The lane-changing probability results of the vehicle early lane-changing intention prediction model in this comparative example on the SQM1 dataset are as Figure 6 shown by the KAN_cross-entropy loss function curve in. Through Figure 6 it can be seen that the lane-changing probability predicted early by the vehicle early lane-changing intention prediction model in this comparative example before the lane-changing completion moment is much lower than that of the vehicle early lane-changing intention prediction model in Example 2.

[0070] Comparative Example 2

[0071] This comparative example conducts experiments in the existing domestic and foreign traffic dataset SQM1 using the same method, dataset, and lane-changing dataset division conditions as in Example 2. However, in this comparative example, the vehicle early lane-changing intention prediction model uses a long short-term memory network (LSTM) with a cross-entropy loss function.

[0072] The lane-changing probability results of the vehicle early lane-changing intention prediction model in this comparative example on the SQM1 dataset are as Figure 6 shown by the LSTM_punitive lane-changing weight loss function curve in. Through Figure 6 it can be seen that the lane-changing probabilities predicted in the mid-term and at the final lane-changing completion moment by the vehicle early lane-changing intention prediction model in this comparative example before the lane-changing completion moment are both lower than those of the vehicle early lane-changing intention prediction model in Example 2.

[0073] Comparative Example 3

[0074] This comparative example conducts experiments in the existing domestic and foreign traffic dataset SQM1 using the same method, dataset, and lane-changing dataset division conditions as in Example 2. However, in this comparative example, the vehicle early lane-changing intention prediction model uses a long short-term memory network (LSTM) with a punitive lane-changing weight loss function.

[0075] The lane-changing probability results of the vehicle early lane-changing intention prediction model in this comparative example on the SQM1 dataset are as Figure 6 shown by the LSTM_cross-entropy loss function curve in. Through Figure 6 it can be seen that the performance of the vehicle early lane-changing intention prediction model in this comparative example is much lower than that of the vehicle early lane-changing intention prediction model used in Example 2.

[0076] From Figure 6From the comparison of the curve results of the four early lane - changing intention prediction models for vehicles, it can be seen that the method of the present invention can greatly increase the prediction time lead and output the prediction value to assist the decision - making of driverless behavior, increase the prediction time lead and output its prediction value to assist the decision - making of driverless behavior.

[0077] The method of the present invention processes the collected driving data of the target vehicle through the KAN network based on the penalty lane - changing weight loss function. While ensuring the accuracy of the predicted lane - changing behavior of the vehicle, it improves the prediction time lead of the target vehicle's lane - changing, increases the reaction time to changes in the surrounding environment, and has stronger practical significance compared with the traditional method of predicting lane - changing behavior, and can better ensure traffic safety in the mixed - traffic environment of manned and driverless vehicles.

[0078] The present invention is not limited to the embodiments described above. The above description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention. The above - mentioned specific embodiments are only illustrative and not restrictive. Without departing from the spirit of the present invention and the scope protected by the claims, those of ordinary skill in the art can make many specific transformations in form under the inspiration of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. A method for predicting vehicle early lane-changing intention based on an improved KAN network, characterized in that: The steps include: S1. Collecting driving data of several target vehicles through the on-board sensors of the unmanned vehicle, and dividing the driving data of all target vehicles into positive samples and negative samples; S2, preprocessing the driving data according to the divided positive samples and negative samples to obtain time series data of several samples, and summarizing all the time series data to obtain a lane change data set; S3, constructing a vehicle early lane change intention prediction model, inputting the lane change data set into the vehicle early lane change intention prediction model using a penalty lane change weight loss function for training, and obtaining a trained vehicle early lane change intention prediction model; S4. Collect continuous driving data from the previous frames to the current frame of several target vehicles to be tested in real time, and input the data into the trained vehicle early lane change intention prediction model for processing to obtain the lane change probability of each target vehicle to be tested in real time. The lane change probability of each target vehicle to be tested is compared with a preset lane change probability threshold to obtain in real time whether each target vehicle to be tested has changed lanes.

2. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 1, characterized in that: The step S1 is specifically as follows: S11, collecting driving data of each frame of a plurality of target vehicles through the on-board sensor of the unmanned vehicle; S12, classifying the target vehicles into lane-changing vehicles and non-lane-changing vehicles according to the driving data; S13. For vehicles that change lanes, the continuous driving data consisting of the target vehicle's lane change completion time and N-1 frames before the lane change completion time is used as a positive sample, and the continuous driving data from 2N-1 frames before the lane change completion time to the previous N frames is used as a negative sample; for vehicles that do not change lanes, any continuous N frames of driving data are used as negative samples.

3. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 2, characterized in that: The target vehicle is the vehicle in front of the unmanned vehicle on the left or right adjacent lane; the driving data includes speed, left longitudinal distance, right longitudinal distance and lateral distance; the left longitudinal distance is the longitudinal distance between the target vehicle and the vehicle in front of the target vehicle on the left adjacent lane; the right longitudinal distance is the longitudinal distance between the target vehicle and the vehicle in front of the target vehicle on the right adjacent lane; the lateral distance is the distance between the center point of the target vehicle and the center line of the lane; the lane change completion time is the time when the center point of the target vehicle coincides with the lane line on one side of the target vehicle.

4. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 1, characterized in that: The step S2 is specifically as follows: S21, divide the sample into time series windows by using a sliding window size of M frames and a sliding step length of one frame to obtain N-M+1 time series, and combine the obtained N-M+1 time series to obtain time series data of one sample; S22. The same method as step S21 is adopted to obtain time series data of several samples, and the time series data of all samples are aggregated to obtain a lane change data set.

5. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 1, characterized in that: The vehicle early lane change intention prediction model in step S3 is a KAN network.

6. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 1, characterized in that: The penalty lane-changing weight loss function in step S3 is set according to the following formula: L i + (p t + )=-∑ x=1 N-M+1 αlog(p ix + ) L j - (p t - )=-∑ y=1 N-M+1 log(1-p jy - ) α=e (-max(0，d-Φ(epoch-1)-γ)) Φ(epoch -1)=1 / k∑ f=1 k G f epoch-1 Where i and x both represent indexes; L i + () is the loss function of the i-th positive sample; p ix + represents the lane-changing probability predicted by the xth time series in the ith positive sample; α represents the weight; j and y both represent indexes; L j - () is the loss function of the jth negative sample; p jy - represents the lane-changing probability obtained from the yth time series in the jth negative sample; e represents a constant; max() represents the maximum value; d represents the difference between the index of the last frame of the xth time series and the index of the lane-changing completion time; epoch represents the index of the training round, and also represents the training of the epoch; Φ(epoch-1) represents the mean of the lane-changing time differences of all positive samples in the epoch-1th training round; γ represents a hyperparameter; k represents the total number of positive samples; f represents an index; G f epoch-1 It represents the lane-changing time difference of the fth positive sample in the epoch-1th training round.

7. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 6, characterized in that: The lane-changing time difference is specifically the difference between the index of the time series when the first lane-changing probability in each positive sample is not less than the preset lane-changing probability threshold and the index of the last time series.

8. The method for predicting vehicle early lane change intention based on improved KAN network according to claim 1, characterized in that: When the lane change probability of the target vehicle to be tested in step S4 is compared with the preset lane change probability threshold, the following formula is used to determine whether the target vehicle to be tested has changed lanes: f(p q )=1,p q ≥P th f(p q )=0,p q <P th Among them, f(p q ) indicates whether the qth target vehicle to be tested changes lanes; p q represents the lane change probability predicted by the continuous driving data of the qth target vehicle to be tested input into the trained vehicle early lane change intention prediction model; 1 indicates that a lane change occurs; 0 indicates that no lane change occurs; P th Indicates the preset lane-changing probability threshold.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images