Driving behavior prediction method, driving behavior prediction model training method and device

Abstract

The present invention relates to a driving behavior prediction method, a driving behavior prediction model training method and a device. The method comprises: inputting a to-be-processed vehicle parameter into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; the target driving behavior prediction model is obtained by: deleting a vehicle parameter corresponding to an Nth target vehicle parameter category determined according to an importance value of the vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; obtaining a first target vehicle parameter sample of the same vehicle parameter category as the vehicle parameter sample to train the driving behavior prediction model to obtain a corresponding prediction accuracy; obtaining a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy to train the driving behavior prediction model to obtain a target driving behavior prediction model. Based on the target driving behavior prediction model, driving behavior can be accurately predicted.

Description

Technical Field

[0001] The present invention relates to the field of system technology, and in particular to a driving behavior prediction method, a driving behavior prediction model training method and a device. Background Art

[0002] Driving behavior prediction plays an important role in various fields, including intelligent transportation and autonomous driving. Related technologies use driving parameters such as speed and acceleration collected by vehicle sensors to train a driving behavior prediction model. This trained driving behavior prediction model is then used to perform driving behavior prediction.

[0003] However, the prediction results of the driving behavior prediction model obtained according to the related technical methods have low accuracy. Correspondingly, the driving behavior prediction based on the driving behavior prediction model has low accuracy in predicting the driving behavior. Summary of the Invention

[0004] One of the purposes of the present invention is to provide a driving behavior prediction method, a driving behavior prediction model training method and a device to solve the problem in the driving behavior prediction method in the related art that the driving behavior prediction model needs to process a large amount of data and the driving behavior prediction results have low accuracy; the second purpose is to provide a driving behavior prediction model training method; the third purpose is to provide a driving behavior prediction device; the fourth purpose is to provide a driving behavior prediction model training device; the fifth purpose is to provide an electronic device; and the sixth purpose is to provide a vehicle.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A driving behavior prediction method, comprising:

[0007] Get the vehicle parameters to be processed of the vehicle;

[0008] Inputting the parameters of the vehicle to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle;

[0009] The target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and using the second target vehicle parameter sample to train the driving behavior prediction model to obtain the target driving behavior prediction model.

[0010] A driving behavior prediction method, applied to the cloud, comprising:

[0011] Receive vehicle parameters to be processed sent by the vehicle;

[0012] Inputting the parameters of the vehicle to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle;

[0013] The target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and using the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0014] A driving behavior prediction method, applied to a vehicle, comprising:

[0015] Get the parameters of the vehicle to be processed;

[0016] Sending the vehicle parameters to be processed to a cloud, so that the cloud can input the vehicle parameters to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle;

[0017] The target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and using the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0018] A driving behavior prediction model training method, comprising:

[0019] Deleting a vehicle parameter corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on an importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2;

[0020] respectively obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples;

[0021] A second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy is obtained, and the preset driving behavior prediction model is trained using the second target vehicle parameter sample to obtain a target driving behavior prediction model.

[0022] A driving behavior prediction device, comprising:

[0023] A first acquisition module is used to acquire vehicle parameters to be processed of the vehicle;

[0024] A second acquisition module is used to input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein, the target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and using the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0025] A driving behavior prediction device, applied to the cloud, comprising:

[0026] A third acquisition module is used to receive the vehicle parameters to be processed sent by the vehicle;

[0027] A fourth acquisition module is used to input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein, the target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample of the same vehicle parameter category as that of each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and using the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0028] A driving behavior prediction device, applied to a vehicle, comprising:

[0029] A fifth acquisition module, used to acquire parameters of the vehicle to be processed;

[0030] A sending module is used to send the vehicle parameters to be processed to the cloud so that the cloud inputs the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and using the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0031] A driving behavior prediction model training device, comprising:

[0032] a sixth acquisition module, configured to delete the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2;

[0033] a seventh acquisition module, configured to respectively acquire first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and use the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples;

[0034] The eighth acquisition module is used to obtain a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and use the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0035] An electronic device comprises: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement at least one of the above-mentioned driving behavior prediction method and driving behavior prediction model training method.

[0036] A computer-readable storage medium, when the instructions in the storage medium are executed by the processor of a mobile terminal, enables the mobile terminal to execute the above-mentioned driving behavior prediction method and driving behavior prediction model training method.

[0037] A vehicle comprises the electronic device described above.

[0038] Beneficial effects of the present application: In this embodiment, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2, then N=1,2……n, based on this embodiment, n loop operations are implemented, wherein, in the Nth loop, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain The N+1th vehicle parameter sample is obtained. Thus, through n loop operations, n vehicle parameter samples corresponding to the vehicle parameters of the target vehicle parameter category can be obtained. For these n vehicle parameter samples, as well as the vehicle parameter samples before the vehicle parameters are deleted in the first loop operation, a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample is obtained, and the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample, and a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy is obtained. The second target vehicle parameter sample is determined based on the importance value of the vehicle parameter category and the corresponding prediction accuracy. Compared with the methods of the related art, the target driving behavior prediction model obtained based on the second target vehicle parameter sample of this embodiment has a high prediction result accuracy. When the vehicle parameters to be processed are input into the target driving behavior prediction model, the predicted driving behavior obtained has a high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flowchart of a driving behavior prediction method provided in an embodiment of the present invention;

[0040] Figure 2 A flowchart of the steps of a method for obtaining a prediction accuracy corresponding to a first target vehicle parameter sample provided in an embodiment of the present invention;

[0041] Figure 3 A flowchart of a driving behavior prediction method provided in an embodiment of the present invention;

[0042] Figure 4A flowchart of a driving behavior prediction method provided in an embodiment of the present invention;

[0043] Figure 5 A flowchart of a driving behavior prediction model training method provided in an embodiment of the present invention;

[0044] Figure 6 A flowchart of another driving behavior prediction model training method provided in an embodiment of the present invention;

[0045] Figure 7 A flowchart of a method for obtaining an updated feature set provided in an embodiment of the present invention;

[0046] Figure 8 A flowchart of a method for predicting driving behavior based on a driving behavior prediction system provided in an embodiment of the present invention;

[0047] Figure 9 A schematic structural diagram of a driving behavior prediction device provided in an embodiment of the present invention;

[0048] Figure 10 A schematic structural diagram of another driving behavior prediction device provided in an embodiment of the present invention;

[0049] Figure 11 Schematic diagram of the structure of another driving behavior prediction device provided in an embodiment of the present invention;

[0050] Figure 12 A schematic structural diagram of a driving behavior prediction model training device provided in an embodiment of the present invention;

[0051] Figure 13 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0053] Reference Figure 1 , shows a flowchart of a driving behavior prediction method provided in an embodiment of the present invention. The vehicle is equipped with at least one sensor and components, and specifically includes the following steps:

[0054] Step 101: Obtain vehicle parameters to be processed.

[0055] The vehicle parameters to be processed include vehicle parameters related to the driving behavior prediction result. For example, the vehicle parameters may include at least one of the following: vehicle operating parameters, in-vehicle environment parameters, and out-vehicle environment parameters.

[0056] For example, vehicle operating parameters may include vehicle speed, engine speed, engine coolant temperature, transmission oil temperature, and other vehicle operating parameters collected by body sensors. In-vehicle environmental parameters may include: the driver's driving time, the number of people in the vehicle during driving, whether music was played during driving, and whether people in the vehicle were talking during driving. External environmental parameters may include weather, time of day, road conditions, and congestion on the road. Road conditions may include whether the vehicle is traveling on an urban road, a rural road, a highway, or other roads.

[0057] Step 102 : Input the parameters of the vehicle to be processed into the target driving behavior prediction model to obtain the predicted driving behavior of the vehicle.

[0058] For example, the target driving behavior prediction model may be a classification model, and the target driving behavior prediction model is used to obtain a predicted driving behavior based on input parameters of the vehicle to be processed. For example, the target driving behavior prediction model may be a random forest model.

[0059] The target driving behavior prediction model is obtained by the following method:

[0060] Step A1: Delete the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample.

[0061] Among them, the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2.

[0062] The Nth target vehicle parameter category may include one or more vehicle parameter categories. The importance value is used to represent the importance of the vehicle parameter category to the driving behavior prediction result.

[0063] For example, if n is equal to 2, this step includes: deleting the vehicle parameters corresponding to the first target vehicle parameter category from the first vehicle parameter sample to obtain the second vehicle parameter sample; then deleting the vehicle parameters corresponding to the second target vehicle parameter category from the second vehicle parameter sample to obtain the third vehicle parameter sample.

[0064] For another example, if n is equal to 3, this step includes: deleting the vehicle parameters corresponding to the first target vehicle parameter category from the first vehicle parameter sample to obtain the second vehicle parameter sample; then deleting the vehicle parameters corresponding to the second target vehicle parameter category from the second vehicle parameter sample to obtain the third vehicle parameter sample; then deleting the vehicle parameters corresponding to the third target vehicle parameter category from the third vehicle parameter sample to obtain the fourth vehicle parameter sample.

[0065] The vehicle parameter category may include at least one of the following: vehicle operating parameters, in-vehicle environment parameters, and out-vehicle environment parameters.

[0066] For example, the importance value of the deleted Nth target vehicle parameter category is less than the importance values ​​of other vehicle parameter categories except the Nth target vehicle parameter category in the Nth vehicle parameter sample; for example, the importance value is used to characterize the importance of the vehicle parameter category corresponding to the importance value to the driving behavior prediction result.

[0067] For example, the Nth vehicle parameter sample is input into the driving behavior prediction model to obtain the importance value of each vehicle parameter category in the Nth vehicle parameter sample; for example, the driving behavior prediction model can be a random forest model.

[0068] For example, the parameter n can be determined based on the vehicle parameter category in the first vehicle parameter sample and the preset number of remaining vehicle parameter categories. For example, the vehicle parameter category in the first vehicle parameter sample is k, and there is 1 Nth target vehicle parameter category deleted each time, then n=k-1. Therefore, after executing the operation of this step, the remaining vehicle parameter category in the (n+1)th vehicle parameter sample is 1.

[0069] For example, the parameter n can also be set according to user needs under the constraint that the remaining vehicle parameter categories in the (n+1)th vehicle parameter sample are greater than zero. For example, it can be set to 2, 3, or other values.

[0070] For example, the driving behavior prediction model in this embodiment can be a random forest model. The random forest model can not only be used for classification and regression processing, but also can perform feature importance analysis on the vehicle parameters of the input model. Specifically, when using the random forest model for importance analysis, the importance analysis is mainly performed from two levels: accuracy and node impurity. Furthermore, when analyzing the importance, the random forest model uses the bootstrap sampling method to sample the parameter features of the vehicle parameters and calculates their out-of-bag error (OOB) value. The parameter features are then replaced according to the OOB value, and the OOB value of the random forest is calculated again based on the replaced parameter features. By comparing the OOB values ​​of the parameter features of each vehicle parameter, the importance value of the vehicle parameter is obtained.

[0071] For example, the importance value of the Nth target vehicle parameter category is less than the importance values ​​of other vehicle parameter categories except the Nth target vehicle parameter category in the Nth vehicle parameter sample.

[0072] Correspondingly, in this step, the vehicle parameters deleted from the Nth vehicle parameter sample are the vehicle parameters of the Nth target vehicle parameter category whose importance values ​​are smaller than the importance values ​​of other vehicle parameter categories.

[0073] For example, there may be one or more Nth target vehicle parameter categories. The importance values ​​of the vehicle parameter categories in the Nth vehicle parameter sample are sorted according to their importance values, and then a preset number of vehicle parameter categories with smaller importance values ​​are determined as the Nth target vehicle parameter category. For example, if there is only one Nth target vehicle parameter category, the vehicle parameter category with the smallest importance value may be determined as the Nth target vehicle parameter sample.

[0074] For example, vehicle parameter categories include vehicle speed, engine speed, engine coolant temperature, transmission oil temperature, road conditions of the vehicle's driving road, number of people in the vehicle, and weather. The Nth vehicle parameter sample including these vehicle parameters is input into the driving behavior prediction model, and the important parameters obtained are from high to low: vehicle speed, engine speed, transmission oil temperature, road conditions of the vehicle's driving road, number of people in the vehicle, weather, and engine coolant temperature. The engine coolant temperature can then be determined as the Nth target vehicle parameter category and deleted to update the Nth vehicle parameter sample and obtain the N+1th vehicle parameter sample. The vehicle parameters included in the N+1th vehicle parameter sample are vehicle speed, engine speed, transmission oil temperature, road conditions of the vehicle's driving road, number of people in the vehicle, and weather.

[0075] This step performs n loop operations. For each parameter N, where N∈[1,n], and n is an integer greater than or equal to 2, in the Nth loop operation, the importance value of each vehicle parameter category in the Nth vehicle parameter sample is obtained. A preset number of vehicle parameter categories are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample. If the importance value of the Nth target vehicle parameter category is less than the importance values ​​of all other vehicle parameter categories in the Nth vehicle parameter sample, then in each loop operation, the vehicle parameters of the Nth target vehicle parameter category with a smaller importance value than the importance values ​​of the other vehicle parameter categories are deleted. In other words, the importance value of the vehicle parameter category in the N+1th vehicle parameter sample obtained based on this embodiment is greater than the importance value of the deleted vehicle parameter category. Based on this, the prediction accuracy is subsequently calculated for the first target vehicle parameter sample using the same parameter category as each vehicle parameter sample, and a second target vehicle parameter sample is determined based on the prediction accuracy. The prediction accuracy of the vehicle parameter category in the second target vehicle parameter sample is equal to the preset prediction accuracy and has a relatively large importance value. Using the second target vehicle parameter samples including these vehicle parameter categories to train the driving behavior prediction model can avoid the influence of vehicle parameters with relatively small importance values ​​on the prediction results, thereby ensuring the prediction accuracy.

[0076] This embodiment implements recursive deletion of vehicle parameter categories from vehicle parameter samples. For example, if the driving behavior prediction model is a random forest model, this step can be combined with the random forest model to perform recursive feature elimination to remove redundant features with low importance values, thereby obtaining features with high importance values ​​for driving behavior prediction results. The redundant features are, in this embodiment, vehicle parameters with low importance values.

[0077] Step A2: respectively obtain first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and use the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples.

[0078] For example, the vehicle parameter samples include vehicle speed, engine speed, transmission oil temperature, road conditions of the vehicle, number of people in the vehicle, and weather. The first target vehicle parameter samples also include vehicle parameters of these vehicle parameter categories.

[0079] A first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample is obtained, and a driving behavior prediction model is trained using the first target vehicle parameter sample to obtain a prediction accuracy corresponding to the first target vehicle parameter sample. Furthermore, the driving behavior prediction model is trained using the first target vehicle parameter sample to obtain a trained first driving behavior prediction model, and the prediction accuracy of the trained first driving behavior prediction model is determined as the prediction accuracy corresponding to the first target vehicle parameter sample.

[0080] In which, during each loop, while executing the loop operation, a first target vehicle parameter sample of the same vehicle parameter category as the vehicle parameter sample is obtained, and the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; or after the loop is completed, the first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample is obtained respectively, and the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample.

[0081] Step A3: Obtain a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and use the second target vehicle parameter sample to train a driving behavior prediction model to obtain a target driving behavior prediction model.

[0082] The second target vehicle parameter sample can be a vehicle parameter sample selected from the vehicle parameter samples and having the same vehicle parameter category as the first target vehicle parameter sample, or it can be a separately obtained parameter sample having the same vehicle parameter category as the first target vehicle parameter sample. In other words, in this step, it is sufficient to ensure that the vehicle parameter category of the second target vehicle parameter sample is the same as the vehicle parameter category of the first target vehicle parameter sample corresponding to the preset prediction accuracy.

[0083] Driving behavior prediction is crucial for improving traffic safety, optimizing traffic flow, and assisting intelligent driving. Related technologies utilize driving parameters such as speed and acceleration captured by vehicle sensors, directly using these parameters to train a driving behavior prediction model, resulting in a trained driving behavior prediction model. Driving behavior prediction is then performed based on this trained driving behavior prediction model. However, some of these driving parameters may negatively impact the accuracy of the model's prediction results. According to related technical methods, directly training a driving behavior prediction model using these parameters results in low prediction accuracy. Consequently, driving behavior prediction based on this driving behavior prediction model also results in low accuracy.

[0084] For example, the method of this embodiment can be applied to the cloud, to a vehicle, or to a system consisting of a cloud and a vehicle. In the case where the method is applied to the cloud, the vehicle parameters to be processed obtained by the cloud are data collected by the vehicle and sent to the cloud.

[0085] In this embodiment, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2, then N=1,2...n, based on this embodiment, n loop operations are implemented, wherein, in the Nth loop, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample. Vehicle parameter samples, thus, through n loop operations, n vehicle parameter samples corresponding to the deleted target vehicle parameter category can be obtained. For these n vehicle parameter samples, as well as the vehicle parameter samples before the vehicle parameters are deleted in the first loop operation, a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample is obtained, and the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample, and a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy is obtained. The second target vehicle parameter sample is determined based on the importance value of the vehicle parameter category and the corresponding prediction accuracy. Compared with the methods of related technologies, the target driving behavior prediction model obtained based on the second target vehicle parameter sample of this embodiment has a high prediction result accuracy. When the vehicle parameters to be processed are input into the target driving behavior prediction model, the predicted driving behavior obtained has a high accuracy.

[0086] For example, features used for model training can be screened using embedding or filtering methods. Specifically, the embedding method selects feature subsets by adjusting the weights of feature variables. This method has varying applicability for different algorithm models. The filtering method evaluates the importance of individual feature variables and then filters features with importance below a set threshold, thereby achieving feature screening. However, the threshold is typically set based on experience, and the resulting feature subset may not necessarily be optimal. However, these feature screening methods are relatively simple, and the resulting feature subsets may not be optimal. Models trained based on these feature subsets may have low prediction accuracy. The second target vehicle parameter samples of this embodiment are derived by combining the importance and prediction accuracy of vehicle parameters to driving behavior prediction results. Compared to methods that select parameters based on their importance and train a driving behavior prediction model using the selected parameters, the second target vehicle parameter samples of this embodiment meet both importance and prediction accuracy requirements. Based on the second target vehicle parameter samples, a target driving behavior prediction model can be accurately derived.

[0087] Compared to obtaining the importance value of the vehicle parameter feature, then filtering out the features that meet the threshold conditions according to the set importance value threshold, constructing a feature subset based on the filtered features, and training the model based on the feature subset. The present application does not need to use the importance value threshold, which is a parameter that needs to be set according to subjective experience. The present embodiment obtains vehicle parameter samples of a combination of multiple vehicle parameter category samples according to the importance value, and then filters the combination of multiple vehicle parameter category samples according to the prediction accuracy calculated by the first target vehicle parameter sample to obtain a second target vehicle parameter sample corresponding to the preset prediction accuracy. The second target vehicle parameter sample thus obtained is not affected by subjective experience, and has higher availability and accuracy. Correspondingly, the prediction result of the target driving behavior prediction model obtained based on the second target vehicle parameter sample is more accurate.

[0088] In one embodiment of the present application, in step A2, the method of using the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample may include sub-steps A21 to A24:

[0089] Sub-step A21 : obtaining multiple groups of sample combinations including training sets and test sets based on the first target vehicle parameter samples.

[0090] There are multiple first target vehicle parameter samples, and the multiple first target vehicle parameter samples can be divided multiple times to obtain multiple groups of sample combinations including training sets and test sets.

[0091] Sub-step A22: training a driving behavior prediction model according to the sample combination to obtain a first prediction accuracy corresponding to the sample combination.

[0092] Specifically, for each sample combination, a driving behavior prediction model is trained according to a training set and a test set in the sample combination to obtain a trained first driving behavior prediction model, and a first prediction accuracy of the first driving behavior prediction model is obtained.

[0093] Sub-step A23: Obtain an average prediction accuracy of multiple first prediction accuracies.

[0094] Sub-step A24, obtaining the prediction accuracy corresponding to the first target vehicle parameter sample according to the average prediction accuracy.

[0095] The average prediction accuracy value can be directly determined as the prediction accuracy corresponding to the first target vehicle parameter sample. Alternatively, a preset correction coefficient can be obtained and the average prediction accuracy value can be corrected according to the preset correction coefficient to obtain the prediction accuracy corresponding to the first target vehicle parameter sample.

[0096] In this embodiment, multiple groups of sample combinations including training sets and test sets are obtained based on the first target vehicle parameter samples. The driving behavior prediction model is trained separately through each sample combination including the training set and the test set to obtain first prediction accuracies corresponding to the sample combinations. These first prediction accuracies are prediction accuracies that correspond one-to-one to the sample combinations. According to the average value of these first prediction accuracies, the prediction accuracy that can reflect the overall prediction effect of multiple groups of sample combinations can be accurately obtained, and it is used as the prediction accuracy corresponding to the first target vehicle parameter sample, which has the characteristic of high accuracy of the obtained first target vehicle parameter sample.

[0097] In one embodiment of the present application, there are multiple first target vehicle parameter samples. Sub-step A21 of obtaining multiple training set and test set sample combinations based on the first target vehicle parameter samples may include: traversing the first target vehicle parameter samples, and for each traversal operation, determining the traversed first target vehicle parameter sample as the test set and determining the remaining first target vehicle parameter samples as the training set, thereby obtaining sample combinations corresponding to the traversal operation.

[0098] For example, refer to Figure 2There are k first target vehicle parameter samples. In the first traversal operation, the first first target vehicle parameter sample can be determined as the test set, and the other k-1 first target vehicle parameter samples can be determined as the training set, resulting in a sample combination 1 that includes the test set and the training set. In the second traversal operation, the second first target vehicle parameter sample can be determined as the test set, and the other k-1 first target vehicle parameter samples can be determined as the training set, resulting in a sample combination 2 that includes the test set and the training set. This is analogous to obtaining k sample combinations that include the training set and the test set. In the kth sample combination, the kth first target vehicle parameter sample is used as the test set, and the other second vehicle test samples are used as the training set.

[0099] After obtaining k sample combinations including training sets and test sets, each sample combination is used to train a preset driving behavior prediction module to obtain a first prediction accuracy corresponding to the sample combination, and then the k first prediction accuracies are averaged to obtain a prediction accuracy corresponding to the first target vehicle parameter sample.

[0100] According to the aforementioned embodiment, a sample combination including a training set and a test set is used to determine the prediction accuracy calculation of the first target vehicle parameter sample. In this embodiment, a plurality of sample combinations including a training set and a test set are obtained through a traversal method, whereby each first target vehicle parameter sample is sequentially used as a test set, and the first target vehicle parameter samples other than the traversed first target vehicle parameter sample are used as training sets, thereby obtaining multiple groups of sample combinations. The prediction accuracy of the first target vehicle parameter sample is obtained based on the sample combination, achieving cross-validation of the prediction accuracy corresponding to the vehicle parameter category in the first target vehicle parameter sample, and dynamically and accurately obtaining the prediction accuracy when using the vehicle parameters of the vehicle parameter category in the first target vehicle parameter sample for driving behavior prediction.

[0101] The Nth vehicle parameter sample is updated based on the importance of the vehicle parameter category, achieving recursive feature elimination of the vehicle parameters in the Nth vehicle parameter sample. In combination with the method of this embodiment, a combination of recursive feature elimination and cross-validation methods is achieved. Based on the combined processing of recursive feature elimination and cross-validation methods, the prediction accuracy corresponding to the vehicle parameter category can be dynamically evaluated, and the stability of the obtained prediction accuracy can be ensured. Based on the combined processing of recursive feature elimination and cross-validation methods, a second target vehicle parameter sample including the optimal vehicle parameter category can be screened. Based on the second target vehicle parameter sample, a feature subset with optimal prediction performance can be obtained. Based on the second target vehicle parameter sample, a driving behavior prediction model is trained, and the obtained target driving behavior prediction model has high prediction accuracy.

[0102] In one embodiment of the present application, sub-step A22 may include sub-steps A221 and A222:

[0103] Sub-step A221 : training a driving behavior prediction model based on the sample combination to obtain a first driving behavior prediction model, as well as the precision and recall of the first driving behavior prediction model.

[0104] The precision of the first driving behavior prediction model reflects the accuracy of the first driving behavior prediction model in predicting positive examples and is used to characterize the proportion of true positive examples. A positive example is a prediction result that is predicted as a driving behavior, and a true positive example is a prediction result that is correctly predicted as a driving behavior. The recall rate of the first driving behavior prediction model reflects the first driving behavior prediction model's ability to capture positive examples.

[0105] Sub-step A222: obtaining a first prediction accuracy corresponding to the sample combination according to the precision and recall.

[0106] The first prediction accuracy is used to evaluate the performance of the first driving behavior prediction model, which can comprehensively reflect the size of the precision and recall rate. For example, the first prediction accuracy F corresponding to the sample combination can be obtained according to the following method:

[0107]

[0108] The precision rate is the proportion of true positive examples in the samples predicted as positive examples; is the recall rate, which is the proportion of actual positive examples that are correctly predicted as positive examples.

[0109] The higher the first prediction accuracy F-value, the better the balance between precision and recall of the first driving behavior prediction model, and the better the performance of the first driving behavior prediction model. For example, an F-value of 1 indicates that both the model's precision and recall have reached 100%. A lower F-value indicates that either the precision or recall of the first driving behavior prediction model is lower.

[0110] In this embodiment, a driving behavior prediction model is trained based on the sample combination to obtain a first driving behavior prediction model, as well as the precision and recall of the first driving behavior prediction model. A first prediction accuracy corresponding to the sample combination is obtained based on the precision and recall. The first prediction accuracy can accurately reflect the prediction performance of the first driving behavior prediction model obtained when the driving behavior prediction model is trained using the training set and validation set in the sample combination.

[0111] In one embodiment of the present application, during the loop operation of step A1, when N=1, the Nth vehicle parameter sample can be obtained through sub-steps A11 to A15:

[0112] Sub-step A11: obtaining a driving behavior time series and at least one vehicle operating parameter time series.

[0113] Wherein, when N=1, the Nth vehicle parameter sample is the vehicle parameter sample processed by the first cycle operation.

[0114] The types of vehicle operating parameters in different vehicle operating parameter time series are different from each other.

[0115] The vehicle operating parameters may be parameters collected by body sensors, for example, the vehicle operating parameters may include vehicle speed, engine speed, engine coolant temperature, transmission oil temperature, etc.

[0116] The vehicle operating parameter time series includes multiple vehicle operating parameters sorted in chronological order; the driving behavior time series includes multiple driving behaviors sorted in chronological order. The vehicle operating parameter collection times in the vehicle operating parameter time series correspond to the driving behavior collection times in the driving behavior time series. For example, the vehicle operating parameter time series includes vehicle operating parameters from time T1, T2, T3, and Tn; correspondingly, the driving behavior time series includes driving behaviors from time T1, T2, T3, and Tn.

[0117] Sub-step A12, respectively obtaining the time difference correlation coefficient between the driving behavior time series and the time series of each vehicle operating parameter.

[0118] The time difference correlation coefficient is used to characterize whether the changes in the vehicle operating parameters in the vehicle operating parameter time series are ahead of or lag behind the changes in the driving behavior in the driving behavior time series, or whether the changes are simultaneous with the driving behavior in the driving behavior time series.

[0119] Sub-step A13: obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series according to the time difference correlation coefficient.

[0120] Among them, the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series.

[0121] The time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior in the driving behavior time series changes, indicating that the change in the vehicle operating parameters in the target vehicle operating parameter time series causes the change in the driving behavior in the time series. Therefore, these vehicle operating parameters are the vehicle operating parameters that have a greater impact on driving behavior. Training a driving behavior prediction model based on these vehicle operating parameters results in a more accurate model prediction result.

[0122] Sub-step A14: obtaining vehicle operating parameter samples of the same vehicle parameter category as the target vehicle operating parameter time series.

[0123] For example, if the vehicle operating parameters in the target vehicle operating parameter sequence include vehicle speed and engine speed, the vehicle speed and engine speed are obtained, and a vehicle operating parameter sample is constructed based on the obtained vehicle speed and engine speed.

[0124] Sub-step A15: constructing the Nth vehicle parameter sample based on the vehicle operation parameter sample.

[0125] Other vehicle parameter samples related to driving behavior can be obtained, and these vehicle parameter samples are combined with the vehicle operation parameter samples to determine the Nth vehicle parameter sample.

[0126] Related technologies capture on-board sensor data and then predict driving behavior based on graph convolutional network (GCN) and least squares regularized multi-manifold neural network (LSTMNN) models. However, these prediction methods often have high data requirements and are sensitive to noise. Furthermore, these prediction models are like black boxes, making the prediction results difficult to interpret. Furthermore, these related methods do not assess the temporal relationship between on-board sensor data and driving behavior, but instead make predictions based directly on all sensor data. However, on-board sensor data that aligns with driving behavior or lags behind is meaningless for predictions, affecting accuracy and increasing the amount of data the model must process.

[0127] In this embodiment, by calculating the time difference correlation coefficient between the driving behavior time series and the time series of each vehicle operating parameter, and screening the vehicle operating parameters based on the time difference correlation coefficient, the vehicle operating parameters that are ahead of the driving behavior changes can be screened out. Based on this, it is possible to overcome the problems in related technologies of not screening sensor data, resulting in low accuracy of prediction results, and a large amount of data that the model needs to process.

[0128] In related technologies, driving behavior prediction can also be achieved based on a weighted graph convolutional neural network (WGCN) model. Specifically, a feature matrix and a local map are generated for each vehicle. The weighted feature matrix and the local map, encoded using a convolutional neural network (CNN), are then combined to form a graph. This graph is then input into an edge-augmented graph convolutional neural network (GCN). The GCN then extracts features from the input data, and the interactive features output by the GCN are fed into a Transformer model for model training. The fully connected layers of the model then generate predictions of the vehicle's driving behavior. This method relies on the GCN model to perform predictions, requiring large amounts of data to process, placing high demands on computing power, and making interpretation of the prediction results difficult. Furthermore, this method fails to consider the time lag between vehicle data and driving behavior. For example, the random forest model requires less computing power than the GCN. In this embodiment, a random forest model can be used to predict driving behavior, reducing computing power requirements.

[0129] In practical applications, changes in vehicle sensor data and driving behavior have a temporal relationship. Within the entire time series, two or more sequences may exhibit leading, coincident, or lagging relationships. In this embodiment, the vehicle operating parameters may be data collected by vehicle sensors. Based on the time-difference correlation coefficient of this embodiment, sensor data that precedes changes in driving behavior can be obtained. Determining vehicle operating parameters that precede driving behavior changes using the time-difference correlation coefficient eliminates vehicle operating parameters that lag behind driving behavior or change simultaneously with driving behavior, thereby preventing these vehicle operating parameters from interfering with the prediction results.

[0130] Using the time difference correlation coefficient to analyze driving behavior time series and vehicle operating parameter time series, the vehicle operating parameter time series is used as the baseline sequence, and the driving behavior time series as the analysis sequence. The time difference correlation coefficient quantifies the temporal relationship between the two sequences, accurately determining whether the analysis sequence leads, lags, or is consistent with the baseline sequence. This method offers the advantages of easy-to-understand principles, simple calculations, and accurate and reliable results.

[0131] During the training process of the driving behavior prediction model, vehicle operating parameters that precede driving behavior changes are screened. Training samples for training the driving behavior prediction model are constructed based on the screened vehicle operating parameters. The driving behavior prediction model is trained based on these training samples, resulting in a target driving behavior prediction model with relatively high prediction accuracy. In other words, by screening vehicle operating parameters obtained through sensor acquisition and other methods using the method of this embodiment, sensor data that precedes driving behavior changes can be screened out. Based on this screened sensor data, the prediction accuracy of the target driving behavior prediction model can be improved.

[0132] Furthermore, based on the aforementioned analysis, in this embodiment, by obtaining the time difference correlation coefficient between the driving behavior time series and each vehicle operating parameter time series, at least one target vehicle operating parameter time series is obtained from at least one vehicle operating parameter time series based on the time difference correlation coefficient. In the target vehicle operating parameter time series, changes in the vehicle operating parameters occur earlier than changes in the driving behavior time series. Therefore, the vehicle operating parameters in the target vehicle operating parameter time series are parameters that cause changes in driving behavior and have a significant impact on driving behavior. Based on the vehicle operating parameter samples thus obtained, an Nth vehicle parameter sample having a significant impact on driving behavior can be obtained, where N=1. Therefore, this embodiment obtains the vehicle parameter sample having a significant impact on driving behavior processed in the first loop operation. Based on this vehicle parameter sample, n subsequent loop operations are performed, and the vehicle parameter samples obtained in each loop operation are all parameter samples having a significant impact on driving behavior. Based on this vehicle parameter sample, a second target vehicle parameter sample is obtained. A driving behavior prediction model is trained based on the second target vehicle parameter sample, resulting in a target driving behavior prediction model with high prediction accuracy. In addition, based on this embodiment, it is possible to know which vehicle parameter categories have vehicle operating parameters that change ahead of driving behavior, and these parameters are vehicle operating parameters that have an important impact on driving behavior prediction. By constructing the Nth vehicle parameter sample based on the vehicle operating parameters of these vehicle parameter categories and conducting subsequent training processing of the driving behavior prediction model, the interpretability of the model can be achieved.

[0133] In addition, compared with the method of obtaining the importance value of the feature, then filtering out the features that meet the threshold conditions according to the set importance value threshold, constructing a feature subset based on the features thus filtered out, and training the model based on the feature subset, this embodiment avoids the problem that this method calculates the importance of all features at one time and then filters the features according to the importance value threshold, resulting in the filtered features being unable to reflect the correlation between the features, which will affect the effectiveness of feature selection.

[0134] In one embodiment of the present application, obtaining the time difference correlation coefficient between the driving behavior time series and each vehicle operating parameter time series in sub-step A12 may include sub-step A121:

[0135] In sub-step A121, for each vehicle operating parameter time series, the time difference correlation coefficient between the driving behavior time series and the vehicle operating parameter time series is obtained according to the following method:

[0136]

[0137] in, is the time lag order The corresponding time difference correlation coefficient is used to characterize the time interval between the change of driving behavior in the time series of driving behavior and the change of vehicle operating parameters in the time series of vehicle operating parameters; is the time series of vehicle operating parameters The vehicle operating parameters at the time, is the average value of all vehicle operating parameters in the vehicle operating parameter time series; is the driving behavior time series in The characteristic value of driving behavior at the moment, is the average value of the characteristic values ​​of all driving behaviors in the driving behavior time series, For the total time.

[0138] In this embodiment, the time difference correlation coefficient can reflect the degree of correlation between the driving behavior in the driving behavior time series and the vehicle operating parameters in the vehicle operating parameter time series. Based on the calculation method for obtaining the time difference correlation coefficient in this embodiment, the time difference correlation coefficient corresponding to different time intervals can be accurately obtained, and then the time intervals corresponding to the different degrees of correlation between the driving behavior in the driving behavior time series and the vehicle operating parameters in the vehicle operating parameter time series can be accurately obtained. Based on the time interval, the vehicle operating parameters that are ahead of the driving behavior changes in the driving behavior time series can be accurately obtained.

[0139] In one embodiment of the present application, the time lag order If there are multiple target vehicle operating parameter time series, then in sub-step A13, at least one target vehicle operating parameter time series is obtained from at least one vehicle operating parameter time series according to the time difference correlation coefficient, which may include sub-steps A131 to A133.

[0140] Sub-step A131, for each vehicle operating parameter time series, respectively obtain the time difference correlation coefficient corresponding to each lag order. For example, the lag order The value range of ,in, The maximum delay number is equal to the preset prediction time. The prediction time can be set according to user needs. For example, if you need to predict the driving behavior after 1 minute, the preset time can be set to 1 minute. Within the value range of , multiple time lag orders are obtained according to the preset value interval.

[0141] Sub-step A132, obtaining the maximum time difference correlation coefficient corresponding to each vehicle operating parameter time series, and the time lag order corresponding to the maximum time difference correlation coefficient.

[0142] Among them, the time difference correlation coefficient can reflect the correlation between the driving behavior in the driving behavior time series and the vehicle operating parameters in the vehicle operating parameter time series. The maximum time difference correlation coefficient is It is the maximum value of the correlation between the driving behavior in the driving behavior time series and the vehicle operating parameters in the vehicle operating parameter time series, and its corresponding time lag order is , is the time lag order when the correlation is maximum.

[0143] The time lag order is used to characterize the time interval between changes in driving behavior in the driving behavior time series and changes in vehicle operating parameters in the vehicle operating parameter time series. If it is less than zero, it means that the driving behavior is ahead of the change of vehicle operating parameters. If it is greater than zero, it means that the driving behavior is lagging behind the changes in vehicle operating parameters. is equal to 0, which means that the changes of the two are basically the same. In other words, if the lag order Greater than zero, indicating that the change of vehicle operating parameters is ahead of the change of driving behavior, and the lag order is greater than zero The corresponding vehicle operating parameters are the vehicle operating parameters that have the greatest impact on changes in driving behavior

[0144] In sub-step A133 , the vehicle operating parameter time series corresponding to the largest time difference correlation coefficient and having a lag order greater than zero is determined as the target vehicle operating parameter time series.

[0145] In this embodiment, if the time difference correlation coefficient is the largest, it indicates that the correlation between driving behavior and vehicle operating parameters is the greatest. The time lag order corresponding to the largest time difference correlation coefficient is the time lag order when the correlation between driving behavior and vehicle operating parameters is the greatest. If this time lag order is greater than zero, it indicates that changes in the vehicle operating parameters precede changes in driving behavior when the correlation is greatest. Furthermore, the target vehicle operating parameter time series with a time lag order greater than zero corresponding to the largest time difference correlation coefficient is the vehicle operating parameter time series in which the vehicle operating parameters precede changes in driving behavior when the correlation is greatest. Constructing a vehicle operating parameter sample based on this target vehicle operating parameter time series ensures that the vehicle operating parameters in the vehicle operating parameter sample are those that cause changes in driving behavior, those that precede changes in driving behavior, those that lag behind changes in driving behavior or that change in sync with driving behavior, and those that eliminate interference from vehicle sensor data on the prediction results. These vehicle operating parameters significantly influence driving behavior. Based on these vehicle operating parameters, an Nth vehicle parameter sample is constructed, and a target driving behavior prediction model is obtained based on the Nth vehicle parameter sample. The prediction results of the target driving behavior prediction model are highly accurate. Furthermore, based on the target vehicle operating parameter time series obtained in this embodiment, it is possible to determine which vehicle operating parameters significantly influence the driving behavior prediction results, thereby achieving interpretability of the driving behavior prediction model.

[0146] In one embodiment of the present application, constructing the Nth vehicle parameter sample based on the vehicle operation parameter sample in sub-step A15 may include sub-steps A151 and A152:

[0147] Sub-step A151, obtaining vehicle interior environment parameters and vehicle exterior environment parameters.

[0148] For example, in-vehicle environmental parameters may include: driving time, number of people in the vehicle during driving, whether music is played during driving, whether people in the vehicle are talking during driving, etc. External environmental parameters may include weather, time of driving, road conditions, and congestion conditions of the road. The road conditions may include whether the road is an urban road, a rural road, a highway, or other roads.

[0149] Sub-step A152, constructing the Nth vehicle parameter sample based on the vehicle operation parameter sample, the vehicle interior environment parameters, and the vehicle exterior environment parameters.

[0150] For example, the Nth vehicle parameter sample includes vehicle driving data collected by on-board sensors, external vehicle environment parameters obtained from a public natural driving data set, and internal vehicle environment parameters.

[0151] In related art, the training samples used for driving behavior prediction can be vehicle operating parameters such as speed. However, driving behavior is not only affected by vehicle operating parameters, but also by external environmental parameters such as weather and traffic flow on the road. Furthermore, internal environmental parameters such as the driver's in-vehicle environment are closely related. Related art ignores the impact of both the external and internal environments on driving behavior. For example, related art models train based on data collected by on-board sensors, without considering data from the in-vehicle environment during driving. Related art can also predict driving behavior based on a GCN model, converting vehicle positions into an image format. However, driving scenarios are highly variable. For example, in multi-vehicle scenarios, changes in external environmental parameters such as relative speeds between vehicles significantly impact driving behavior. However, this method in related art fails to consider external environmental factors, which can lead to inaccurate prediction results. This embodiment, however, considers vehicle operating parameters as well as external and internal environmental parameters. The Nth vehicle parameter sample obtained using this embodiment's method can improve the prediction accuracy of the target driving behavior prediction model.

[0152] In this embodiment, an Nth vehicle parameter sample is constructed based on vehicle operating parameter samples, in-vehicle environmental parameters, and external environmental parameters. The vehicle parameters in the Nth vehicle parameter sample include not only vehicle operating parameters that change ahead of driving behavior and significantly influence driving behavior prediction results, but also in-vehicle and external environmental parameters. This Nth vehicle parameter sample accurately reflects the impact of the in-vehicle and external environmental environments on driving behavior. Compared to related art methods that use vehicle operating parameters such as vehicle speed to construct a sample, the Nth vehicle parameter sample in this embodiment includes a more comprehensive set of vehicle parameters, resulting in a more accurate prediction result for the target driving behavior prediction model.

[0153] In one embodiment of the present application, there are multiple second target vehicle parameter samples. Sub-step A3 uses the second target vehicle parameter samples to train a driving behavior prediction model to obtain a target driving behavior prediction model, which may include sub-steps A31 to A33:

[0154] Sub-step A31 : Use each second target vehicle parameter sample to train a driving behavior prediction model, and obtain a driving behavior prediction model corresponding to each second target vehicle parameter sample.

[0155] Sub-step A32, respectively obtaining the determination coefficient of the driving behavior prediction model corresponding to each second target vehicle parameter sample.

[0156] Among them, the coefficient of determination is used to characterize the model fitting goodness of the driving behavior prediction model.

[0157] In sub-step A33, the driving behavior prediction model corresponding to the preset determination coefficient is determined as the target driving behavior prediction model.

[0158] From the driving behavior prediction model corresponding to the determination coefficient, a driving behavior prediction model corresponding to a preset determination coefficient is obtained.

[0159] In this embodiment, the determination coefficient of the driving behavior prediction model corresponding to each second target vehicle parameter sample is obtained respectively; the determination coefficient is used to characterize the model fitting goodness of the driving behavior prediction model; the model parameters of the driving behavior prediction model are optimized according to the determination coefficient, which can prevent the driving behavior prediction model from overfitting, and the driving behavior prediction model corresponding to the preset determination coefficient is determined as the target driving behavior prediction model, which can avoid the overfitting of the target driving behavior prediction model and ensure the availability of the target driving behavior prediction model and the accuracy of the prediction results.

[0160] In one embodiment of the present application, the vehicle parameter to be processed is of the same vehicle parameter category as the second target vehicle parameter sample. Thus, it is ensured that the vehicle parameter to be processed input into the target driving behavior prediction model is of the same vehicle parameter category as the second target vehicle parameter sample used in the model training process, thereby ensuring the accuracy of the driving behavior prediction result. In addition, because the vehicle parameter category of the second target vehicle parameter sample is obtained by screening a plurality of vehicle parameter categories based on importance values ​​and prediction accuracy, therefore, by obtaining the vehicle parameter to be processed that is of the same vehicle parameter category as the second target vehicle parameter sample and performing driving behavior prediction based on the vehicle parameter to be processed, fewer types of vehicle parameters need to be collected and processed, and the amount of data is small, thereby ensuring the accuracy of the prediction result and improving processing efficiency.

[0161] Reference Figure 3 A driving behavior prediction method according to an embodiment of the present application, applied to the cloud, may include the following steps:

[0162] Step 201: Receive vehicle parameters to be processed sent by the vehicle.

[0163] The vehicle parameters to be processed sent by the vehicle may include vehicle operating parameters and in-vehicle environmental parameters collected by vehicle sensors, or may include vehicle operating parameters, in-vehicle environmental parameters and out-vehicle environmental parameters.

[0164] Step 202: Input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method:

[0165] From Delete the first The vehicle parameters corresponding to the target vehicle parameter category are obtained Vehicle parameter samples, where The vehicle parameter sample consists of vehicle parameters of multiple vehicle parameter categories. The target vehicle parameter category is based on Determining the importance value of each vehicle parameter category in the vehicle parameter sample; , is an integer greater than or equal to 2;

[0166] respectively obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples;

[0167] A second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy is obtained, and the driving behavior prediction model is trained using the second target vehicle parameter sample to obtain a target driving behavior prediction model.

[0168] Among them, if the parameters to be processed collected by the vehicle include vehicle operating parameters and in-vehicle environmental parameters, the cloud can obtain the external environmental parameters, and then input the vehicle parameters to be processed consisting of vehicle operating parameters, in-vehicle environmental parameters and external environmental parameters into the target driving behavior prediction model to obtain the predicted driving behavior of the vehicle.

[0169] In this embodiment, the vehicle obtains the vehicle parameters to be processed and transmits them to the cloud. The cloud then performs driving behavior prediction based on the target driving behavior prediction model. Based on the aforementioned analysis, the target driving behavior prediction model obtained in this embodiment has a high prediction accuracy. Inputting the vehicle parameters to be processed into the target driving behavior prediction model yields highly accurate predicted driving behavior. Furthermore, the cloud typically has higher computing power than the vehicle. Executing the above operations in the cloud to obtain predicted driving behavior offers the advantages of high processing efficiency and conserving the vehicle's computing resources.

[0170] Reference Figure 4 A driving behavior prediction method according to an embodiment of the present application, applied to a vehicle, may include the following steps:

[0171] Step 301: Obtain vehicle parameters to be processed.

[0172] The vehicle parameters to be processed may include vehicle operating parameters and vehicle interior environment parameters collected by vehicle sensors, or may include vehicle operating parameters, vehicle interior environment parameters, and vehicle exterior environment parameters.

[0173] In step 302, the vehicle parameters to be processed are sent to the cloud, so that the cloud can input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle. The target driving behavior prediction model is obtained by the following method:

[0174] Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined based on an importance value of each vehicle parameter category in the Nth vehicle parameter sample; , n is an integer greater than or equal to 2;

[0175] respectively obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples;

[0176] A second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy is obtained, and the second target vehicle parameter sample is used to train a preset driving behavior prediction model to obtain a target driving behavior prediction model.

[0177] Among them, if the parameters to be processed collected by the vehicle include vehicle operating parameters and in-vehicle environmental parameters, the cloud can obtain the external environmental parameters, and then input the vehicle parameters to be processed consisting of vehicle operating parameters, in-vehicle environmental parameters and external environmental parameters into the target driving behavior prediction model to obtain the predicted driving behavior of the vehicle.

[0178] In this embodiment, the vehicle acquires the vehicle parameters to be processed and transmits them to the cloud, which then performs driving behavior prediction based on the target driving behavior prediction model. Based on the aforementioned analysis, the target driving behavior prediction model obtained in this embodiment has a high degree of prediction accuracy. Inputting the vehicle parameters to be processed into the target driving behavior prediction model yields highly accurate predicted driving behavior. Furthermore, since the cloud typically has higher computing power than the vehicle, executing the aforementioned operations in the cloud to predict driving behavior offers the advantages of high processing efficiency and conserving the vehicle's computing resources.

[0179] Reference Figure 5 A driving behavior prediction model training method according to an embodiment of the present application may include the following steps:

[0180] Step 401, delete the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2.

[0181] The importance value is used to represent the importance of the vehicle parameter category to the driving behavior prediction results.

[0182] Step 402 : Obtain first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and use the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples.

[0183] Among them, the importance value of the deleted vehicle parameter category is less than the importance values ​​of other vehicle parameter categories except the deleted vehicle parameter category; the importance value is used to represent the importance of the vehicle parameter category to the driving behavior prediction result.

[0184] Step 403 : Obtain a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy, and use the second target vehicle parameter sample to train a preset driving behavior prediction model to obtain a target driving behavior prediction model.

[0185] In this embodiment, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; , n is an integer greater than or equal to 2, then N=1, 2, ... n. Based on this embodiment, n loop operations are implemented, wherein in the Nth loop, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample. Thus, through n loop operations, n vehicle parameter samples corresponding to the target vehicle parameter category are obtained. For these n vehicle parameter samples, as well as the vehicle parameter samples before the vehicle parameters are deleted in the first loop operation, first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample are obtained. The first target vehicle parameter samples are used to train the driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter sample. A second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy is obtained. The second target vehicle parameter sample is determined based on the importance value of the vehicle parameter category and the corresponding prediction accuracy. Compared with the related art method, the target driving behavior prediction model obtained based on the second target vehicle parameter sample in this embodiment has a high prediction accuracy. When the vehicle parameters to be processed are input into the target driving behavior prediction model, the predicted driving behavior is highly accurate.

[0186] With the development of Internet of Vehicles technology and artificial intelligence, traditional driving behavior prediction has been unable to meet the requirements of driving behavior prediction accuracy. To solve this problem, driving behavior prediction based on artificial intelligence has emerged. In related technologies, vehicle driving information can be collected based on on-board sensors. According to the driving information and the system's lane change trajectory prediction requirements, a random forest model is trained and the vehicle lane change trajectory is predicted based on the trained model. However, this method of related technology only uses the ranking results of the feature importance of the random forest itself to extract features and trains the prediction model based on the extracted features. This processing method may not necessarily screen out the optimal feature subset, and the accuracy of the driving behavior prediction model trained based on the extracted features is not high. The driving behavior prediction model training method of this embodiment can overcome the problem of low accuracy of the driving behavior prediction model in the related art. In order to enable those skilled in the art to better understand the embodiments of the present invention, the training method of the driving behavior prediction model of this application is further illustrated by an example.

[0187] Reference Figure 6 The training method of the driving behavior prediction model of the embodiment of the present application may include the following steps:

[0188] Step S1, obtaining vehicle body sensor data and driving behavior data.

[0189] The vehicle body sensor data is equivalent to the vehicle operating parameters in the aforementioned embodiment, and the driving behavior data includes the in-vehicle environment data in the aforementioned embodiment.

[0190] Step S2: performing time difference correlation analysis on the vehicle body sensor data and the driving behavior data to obtain a time difference correlation coefficient.

[0191] The method of this step has been described in the aforementioned steps A11 to A15 and will not be repeated here.

[0192] Step S3: Obtain the vehicle body sensor data that precedes the driving behavior change based on the time difference correlation coefficient, and then proceed to step S6.

[0193] The vehicle body sensor data that precedes the driving behavior change is the vehicle body sensor data that precedes the driving behavior change. The vehicle body sensor data that precedes the driving behavior change is equivalent to the vehicle operating parameters in the target vehicle parameter time series in the aforementioned embodiment.

[0194] Step S4: Acquire vehicle interior environment parameters, and then proceed to step S6.

[0195] Step S5: Obtaining vehicle exterior environment parameters.

[0196] Step S6, constructing a vehicle parameter sample based on the vehicle body sensor data, the in-vehicle environment data, and the out-vehicle environment data that precede the changes in driving behavior, and combining the importance value analysis of the vehicle parameter category and the verification of the prediction accuracy of driving behavior prediction using the vehicle parameters corresponding to the vehicle parameter category, to screen the vehicle parameter category in the Nth vehicle parameter sample.

[0197] In this step, the Nth vehicle parameter sample is constructed using the filtered vehicle body sensor data that indicates driving behavior changes, along with in-vehicle and out-of-vehicle environmental parameters. Then, through multiple loops, the vehicle parameter categories are deleted and updated based on their corresponding importance, achieving recursive feature elimination.

[0198] For example, the driving behavior prediction model is a random forest model. Based on this step, a combination of recursive feature elimination method, random forest model processing method and cross-validation method is implemented. Based on the combination of these methods, an optimal feature subset can be obtained. The optimal feature subset is equivalent to the second target vehicle parameter sample in the aforementioned embodiment.

[0199] Furthermore, this step is implemented through a loop operation, wherein the input data set is obtained in each loop operation: , , M is the total number of features, features are vehicle parameter categories, Xi is the i-th parameter feature, and y is the predicted driving behavior category. The initial feature set is F = {}. The initial feature set F = {} is the M-th vehicle parameter sample input to the driving behavior prediction model in the first cycle.

[0200] In each loop operation, the types of vehicle parameters are sorted according to their importance values ​​to obtain a sorted feature set O{}. The following is a further exemplary description of the method of this step:

[0201] Reference Figure 7 The method for obtaining the prediction accuracy in this step may include sub-steps B1 to B4:

[0202] Sub-step B1: obtaining a feature set.

[0203] The feature set in this step is the Nth vehicle parameter sample in the aforementioned embodiment.

[0204] Sub-step B2, using the random forest model, obtains the importance value ranking results of each vehicle parameter category in the feature set, and deletes the vehicle parameters of the vehicle parameter category with the smallest importance value based on the importance value ranking results to obtain an updated feature set.

[0205] This embodiment is merely an exemplary description. In this embodiment, the vehicle parameter with the smallest importance value is deleted. Vehicle parameters of a preset number of vehicle parameter categories with relatively small importance values ​​may also be deleted.

[0206] Sub-step B3, judging whether the updated features meet the preset loop termination condition, if so, the loop is terminated, otherwise, it returns to step B1.

[0207] Sub-step B4, respectively verifying the prediction accuracy of driving behavior prediction using vehicle parameter samples of the same vehicle parameter category as each updated feature set.

[0208] For example, step S6 may further include sub-steps C1 to C3:

[0209] In sub-step C1, random forest is used to sort the feature importance of various vehicle parameters in the initial feature set F{} to obtain the sorted feature set O{}.

[0210] In sub-step C2, the feature with the smallest importance value is removed from the sorted feature set O{} to obtain an updated feature set.

[0211] From the sorted feature set O{}, remove the feature with the smallest importance value, then obtain the importance value of the feature through the random forest model, and update the feature sorting result of the set O{} according to the importance value.

[0212] Sub-step C3, repeat step C2 until all features are traversed and the feature subset with the highest prediction accuracy is selected.

[0213] A feature subset is a set of features that includes a portion of the data in a feature set.

[0214] In this embodiment, cross-validation, recursive feature elimination through loop operations, and a random forest model are combined to obtain the feature subset with the highest prediction accuracy.

[0215] Step S7: obtaining a second target vehicle parameter sample for training the driving behavior prediction model based on the screened vehicle parameter categories.

[0216] Step S8: input the second target vehicle parameter sample into the driving behavior prediction model, adjust the parameters of the driving behavior prediction model, and obtain the target driving behavior prediction model.

[0217] The prediction model in this embodiment can be a random forest model. The random forest model has two model parameters that affect its prediction accuracy and generalization error. These two model parameters include the number of decision trees in the random forest and the number of features selected by the decision tree. Specifically, these two model parameters will affect the classification accuracy of the decision tree and the correlation between the decision trees. Furthermore, by reducing the correlation between the decision trees and increasing the classification accuracy of a single decision tree, the generalization error of the random forest can be reduced. For example, the larger the value of the number of features K selected by the decision tree, the greater the classification accuracy of the single decision tree and the greater the correlation between the decision trees.

[0218] In this embodiment, a grid search method can be used to optimize the model parameters of the driving behavior prediction model. Grid search is an exhaustive search method used to train the driving behavior prediction model by traversing all values ​​of vehicle parameter samples and obtaining the optimal solution for the model parameters based on the model training results.

[0219] Specifically, when searching for the optimal solution for model parameters using a grid search method, the values ​​of the vehicle parameters can be obtained according to the following method: obtain the value range and preset step size of the vehicle parameter r, and obtain the values ​​of the parameter samples that need to be traversed based on the preset step size and value range. For example, if the value range of the vehicle parameter r is [0, 10] and the preset step size is 5, then the values ​​of the vehicle parameter r are 0, 5, and 10. Use the grid search method to solve the optimal model parameters corresponding to the vehicle parameter r. Furthermore, obtain the training effects of the model obtained by training the model using the three vehicle parameters of 0, 5, and 10 respectively, and then obtain the optimal solution by comparing the training effects of the model. This optimal solution is the model parameters of the target driving behavior prediction model.

[0220] For example, the training effect of the driving behavior prediction model can be compared by the determination coefficient. The determination coefficient is an indicator used to evaluate the regression fit of the driving behavior prediction model. Its value range is between 0 and 1. The determination coefficient calculation formula is

[0221]

[0222] in, represents the coefficient of determination, is the observed value or actual value of the i-th driving behavior, is the total number of driving behaviors; is the driving behavior prediction value, is the mean of the observed driving behavior values. For example, a preset feature value can be used to represent driving behavior, and the mean of the observed driving behavior values ​​can be the mean of the feature values ​​corresponding to the driving behavior prediction results. The numerator in this calculation formula is the sum of the squares of the differences between the observed and predicted values, i.e., the residual sum of squares, which represents the degree of error in the model prediction. The denominator is the sum of the squares of the differences between the observed and predicted values, which represents the overall fluctuation of the observed values. A higher coefficient of determination indicates a better fit for the driving behavior prediction model. When constructing the vehicle parameter sample, the influencing factors of the in-vehicle environment on the driver's driving behavior are taken into account. Specifically, in-vehicle environment information, external environment information, and vehicle body sensor data are obtained. Then, a random forest model is used to select vehicle parameter categories based on their importance. The prediction accuracy of the selected vehicle parameter categories is cross-validated. The optimal feature subset (the second target vehicle parameter sample) is then selected from the vehicle parameters. The driving behavior prediction model is trained based on this optimal feature subset, which can improve the prediction accuracy of the target driving behavior prediction model. Specifically, vehicle parameters corresponding to a preset number of vehicle parameter categories are deleted according to the importance values ​​of the vehicle parameters, thereby realizing recursive feature elimination. During the preprocessing process, features are repeatedly eliminated and models are established through a loop operation until all features are traversed. Based on the loop operation, features are eliminated and feature subsets with different numbers are generated. Finally, the optimal feature subset is selected based on the prediction effects of the models established with different feature subsets.

[0223] In this embodiment, the importance ranking of all features in the initial feature set is obtained by using random forest importance evaluation, and then the feature with the smallest importance value is removed from the feature set to generate a new feature set. After traversing all features, the above operation is performed in a loop to finally obtain the feature subset with the highest classification accuracy. In this way, it can be ensured that in each round of the loop, the operation of re-evaluating the importance of the current feature set is performed, based on which the importance value of the feature is continuously adjusted. Based on this embodiment, the importance value of the feature subset can be dynamically and timely adjusted and evaluated to ensure that the prediction performance of the optimal feature subset obtained in the end is better and the data is more stable. At the same time, in order to ensure the stability of the processing results, the entire vehicle parameter category screening process is performed in a cross-validation loop. Furthermore, based on the optimal feature subset after screening, the model parameters are trained using grid search to obtain a random forest prediction model for driving behavior prediction, which has the characteristics of high accuracy of prediction results.

[0224] For example, the present application embodiment also provides a driving behavior prediction system, referring to Figure 8 The driving behavior prediction system includes a cloud and a vehicle. The following is an example of how the cloud and vehicle collaborate to achieve driving behavior prediction:

[0225] Vehicle, used to obtain vehicle parameters to be processed and send them to the cloud;

[0226] The cloud is used to input the parameters of the vehicle to be processed into the target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method:

[0227] Deleting a vehicle parameter corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on an importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2;

[0228] respectively obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples;

[0229] A second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy is obtained, and the preset driving behavior prediction model is trained using the second target vehicle parameter sample to obtain a target driving behavior prediction model.

[0230] Specifically, the cloud sends a data collection command to the vehicle. The vehicle responds to the command by collecting vehicle operating parameters and in-vehicle environment data and returns it to the cloud. The data collection command instructs the vehicle to collect optimal vehicle parameters, which can include vehicle body sensor data characteristics and in-vehicle environment characteristics.

[0231] The cloud receives the vehicle parameters collected by the vehicle side and inputs them into the target driving behavior prediction model to obtain the predicted driving behavior.

[0232] Through interactive processing between the vehicle and the cloud, a target driving behavior prediction model is obtained and driving behavior prediction is performed. Based on this embodiment, the abundant computing resources of the cloud can be fully utilized. Specifically, the optimal feature subset is selected in the cloud, and the driving behavior prediction model is trained and the model parameters are optimized. Because the cloud has abundant computing resources, the training and construction of the driving behavior prediction model is faster.

[0233] Based on the vehicle parameter categories in the optimal feature subset, the cloud sends data collection instructions to the vehicle. The vehicle, in response to these instructions, sends the vehicle sensor data and in-vehicle environment information to the cloud. The cloud then obtains the external environment information based on the vehicle parameter categories in the optimal feature subset, assembles it with the vehicle sensor data and in-vehicle environment information, and feeds the resulting processed data into the target driving behavior prediction model. Due to the cloud's abundant computing resources, it can achieve rapid and accurate predictions of driving behavior.

[0234] The driving behavior prediction model is trained in the cloud, the vehicle parameters that need to be collected are analyzed, and data collection instructions are sent to the vehicle side to finally complete the driving behavior prediction, reducing the vehicle side's requirements for computing power resources.

[0235] It should be noted that this embodiment is only an exemplary description of the driving behavior prediction system. In this embodiment, driving behavior prediction is performed collaboratively by the cloud and the vehicle. In some embodiments, the driving behavior prediction method of the above embodiment can also be performed independently by the vehicle.

[0236] like Figure 9As shown, the present invention discloses a driving behavior prediction device, which includes: a first acquisition module 501, used to obtain vehicle parameters to be processed of a vehicle; a second acquisition module 502, used to input the vehicle parameters to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy, and using the second target vehicle parameter sample to train a preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0237] Optionally, the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample, including: obtaining multiple groups of sample combinations including training sets and test sets based on the first target vehicle parameter sample; training the driving behavior prediction model based on the sample combination to obtain the first prediction accuracy corresponding to the sample combination; obtaining the average prediction accuracy of multiple first prediction accuracies; and obtaining the prediction accuracy corresponding to the first target vehicle parameter sample based on the average prediction accuracy.

[0238] Optionally, there are multiple first target vehicle parameter samples; based on the first target vehicle parameter samples, multiple groups of sample combinations including training sets and test sets are obtained, including: traversing the first target vehicle parameter samples, and for each traversal operation, determining the traversed first target vehicle parameter samples as the test set, and determining other first target vehicle parameter samples as the training set, to obtain a sample combination corresponding to the traversal operation.

[0239] Optionally, a driving behavior prediction model is trained according to the sample combination to obtain a first prediction accuracy corresponding to the sample combination, including: training a driving behavior prediction model according to the sample combination to obtain a first driving behavior prediction model, and the precision and recall rate of the first driving behavior prediction model; and obtaining the first prediction accuracy corresponding to the sample combination according to the precision and recall rate.

[0240] Optionally, in the case where N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; obtaining a time difference correlation coefficient between the driving behavior time series and each vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from the at least one vehicle operating parameter time series based on the time difference correlation coefficient; a time at which the vehicle operating parameter in the target vehicle operating parameter time series changes is earlier than a time at which the driving behavior changes in the driving behavior time series; and obtaining a vehicle operating parameter sample of the same vehicle parameter category as that in the target vehicle operating parameter time series.

[0241] Construct an Nth vehicle parameter sample based on the vehicle operation parameter sample.

[0242] Optionally, constructing the Nth vehicle parameter sample based on the vehicle operation parameter sample includes: obtaining in-vehicle environmental parameters and out-vehicle environmental parameters; and constructing the Nth vehicle parameter sample based on the vehicle operation parameter sample, the in-vehicle environmental parameters, and the out-vehicle environmental parameters.

[0243] Optionally, the importance value of the Nth target vehicle parameter category is less than the importance values ​​of other vehicle parameter categories in the Nth vehicle parameter sample.

[0244] In this embodiment, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2, then N=1,2...n, based on this embodiment, n loop operations are implemented, wherein, in the Nth loop, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample. Vehicle parameter samples, thus, through n loop operations, n vehicle parameter samples corresponding to the deleted target vehicle parameter category can be obtained. For these n vehicle parameter samples, as well as the vehicle parameter samples before the vehicle parameters are deleted in the first loop operation, a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample is obtained, and the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample, and a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy is obtained. The second target vehicle parameter sample is determined based on the importance value of the vehicle parameter category and the corresponding prediction accuracy. Compared with the methods of related technologies, the target driving behavior prediction model obtained based on the second target vehicle parameter sample of this embodiment has a high prediction result accuracy. When the vehicle parameters to be processed are input into the target driving behavior prediction model, the predicted driving behavior obtained has a high accuracy.

[0245] like Figure 10As shown, the present invention discloses a driving behavior prediction device, which is applied to the cloud. The device includes: a third acquisition module 601, which is used to receive vehicle parameters to be processed sent by a vehicle; a fourth acquisition module 602, which is used to input the vehicle parameters to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtaining a first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample, and using the first target vehicle parameter sample to train the driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter sample; obtaining a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy, and using the second target vehicle parameter sample to train a preset driving behavior prediction model to obtain a target driving behavior prediction model.

[0246] In this embodiment, the vehicle obtains the vehicle parameters to be processed and transmits them to the cloud. The cloud then performs driving behavior prediction based on the target driving behavior prediction model. Based on the aforementioned analysis, the target driving behavior prediction model obtained in this embodiment has a high prediction accuracy. Inputting the vehicle parameters to be processed into the target driving behavior prediction model yields highly accurate predicted driving behavior. Furthermore, the cloud typically has higher computing power than the vehicle. Executing the above operations in the cloud to obtain predicted driving behavior offers the advantages of high processing efficiency and conserving the vehicle's computing resources.

[0247] like Figure 11 As shown, the present invention discloses a driving behavior prediction device, which is applied to a vehicle and includes:

[0248] A fifth acquisition module 701 is used to acquire vehicle parameters to be processed; a sending module 702 is used to send the vehicle parameters to be processed to the cloud, so that the cloud can input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein, the target driving behavior prediction model is obtained by the following method: deleting the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; respectively obtain a first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample, and use the first target vehicle parameter sample to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample; obtain a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and use the second target vehicle parameter sample to train the preset driving behavior prediction model to obtain the target driving behavior prediction model.

[0249] In this embodiment, the vehicle acquires the vehicle parameters to be processed and transmits them to the cloud, which then performs driving behavior prediction based on the target driving behavior prediction model. Based on the aforementioned analysis, the target driving behavior prediction model obtained in this embodiment has a high degree of prediction accuracy. Inputting the vehicle parameters to be processed into the target driving behavior prediction model yields highly accurate predicted driving behavior. Furthermore, since the cloud typically has higher computing power than the vehicle, executing the aforementioned operations in the cloud to predict driving behavior offers the advantages of high processing efficiency and conserving the vehicle's computing resources.

[0250] like Figure 12As shown, the present invention discloses a driving behavior prediction model training device, which includes: a sixth acquisition module 801, used to delete the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; a seventh acquisition module 802, used to respectively obtain a first target vehicle parameter sample of the same vehicle parameter category as each vehicle parameter sample, and use the first target vehicle parameter sample to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter sample; an eighth acquisition module 803, used to obtain a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to a preset prediction accuracy, and use the second target vehicle parameter sample to train a preset driving behavior prediction model to obtain a target driving behavior prediction model.

[0251] In this embodiment, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample, wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2, then N=1,2...n, based on this embodiment, n loop operations are implemented, wherein, in the Nth loop, the vehicle parameters corresponding to the Nth target vehicle parameter category are deleted from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample. Vehicle parameter samples, thus, through n loop operations, n vehicle parameter samples corresponding to the deleted target vehicle parameter category can be obtained. For these n vehicle parameter samples, as well as the vehicle parameter samples before the vehicle parameters are deleted in the first loop operation, a first target vehicle parameter sample with the same vehicle parameter category as each vehicle parameter sample is obtained, and the first target vehicle parameter sample is used to train the driving behavior prediction model to obtain the prediction accuracy corresponding to the first target vehicle parameter sample, and a second target vehicle parameter sample with the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy is obtained. The second target vehicle parameter sample is determined based on the importance value of the vehicle parameter category and the corresponding prediction accuracy. Compared with the methods of related technologies, the target driving behavior prediction model obtained based on the second target vehicle parameter sample of this embodiment has a high prediction result accuracy. When the vehicle parameters to be processed are input into the target driving behavior prediction model, the predicted driving behavior obtained has a high accuracy.

[0252] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0253] The embodiment of the present invention further provides an electronic device, such as Figure 13 As shown, it includes a processor 901, a device interface 902, and a memory

[0254] 903 and bus 904;

[0255] Memory 903, used for storing computer programs;

[0256] The processor 901 is configured to implement the steps of any of the above embodiments when executing the program stored in the memory 903 .

[0257] The bus mentioned in the terminal above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. This bus can be divided into address buses, data buses, control buses, etc. For ease of illustration, the figure shows only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0258] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor.

[0259] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0260] The present invention also provides a storage medium. When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute the driving behavior prediction method and the driving behavior prediction model training method of the aforementioned embodiment.

[0261] The present application also provides a vehicle, comprising the electronic device or apparatus in the aforementioned embodiment.

[0262] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0263] The algorithm and display provided herein are not inherently related to any particular computer, virtual device or other equipment. According to the above description, it is obvious that the structure required for constructing this type of device is suitable. In addition, the present invention is not directed to any specific programming language. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the above description of specific languages ​​is for the purpose of disclosing the best mode of the present invention.

[0264] In the description provided herein, numerous specific details are described. However, it is understood that embodiments of the present invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques are not shown in detail so as not to obscure the understanding of this description.

[0265] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Accordingly, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of the invention.

[0266] Those skilled in the art will appreciate that the modules in the devices of the embodiments can be adaptively modified and installed in one or more devices different from the embodiments. The modules, units, or components of the embodiments can be combined into a single module, unit, or component, and furthermore, they can be divided into multiple sub-modules, sub-units, or sub-components. All features disclosed in this specification (including the accompanying claims, abstract, and drawings), and all processes or units of any method or device disclosed therein, can be combined in any sample combination, unless at least some of such features and / or processes or units are mutually exclusive. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0267] The various component embodiments of the present invention may be implemented in hardware, or as software modules running on one or more processors, or as a combination of these. Those skilled in the art will appreciate that in practice, a microprocessor or digital signal processor (DSP) may be used to implement some or all of the functions of some or all of the components of the sorting device according to the present invention. The present invention may also be implemented as an apparatus or device program for performing part or all of the methods described herein. Such a program implementing the present invention may be stored on a computer-readable medium or in the form of one or more signals. Such signals may be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0268] It should be noted that the above embodiments illustrate rather than limit the invention, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between brackets should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising several different elements and by means of appropriately programmed computers. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

[0269] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, apparatuses and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0270] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0271] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0272] It should be noted that the various data-related processes in the embodiments of the present application are all carried out in compliance with the corresponding data protection laws and policies of the country where they are located, and with the authorization given by the owner of the corresponding device.

Claims

1. A driving behavior prediction method, characterized in that: include: Obtaining vehicle parameters to be processed of a vehicle; inputting the vehicle parameters to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

2. The method according to claim 1, characterized in that The using the first target vehicle parameter sample to train the driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter sample includes: According to the first target vehicle parameter sample, a plurality of sample combinations including training sets and test sets are obtained; Training the driving behavior prediction model according to the sample combination to obtain a first prediction accuracy corresponding to the sample combination; Obtaining an average prediction accuracy of a plurality of first prediction accuracies; The prediction accuracy corresponding to the first target vehicle parameter sample is obtained according to the prediction accuracy average value.

3. The method according to claim 2, characterized in that There are multiple first target vehicle parameter samples; and according to the first target vehicle parameter samples, multiple groups of sample combinations including training sets and test sets are obtained, including: The first target vehicle parameter samples are traversed, and for each traversal operation, the traversed first target vehicle parameter samples are determined as a test set, and other first target vehicle parameter samples are determined as a training set to obtain a sample combination corresponding to the traversal operation.

4. The method according to claim 2, characterized in that The step of training the driving behavior prediction model according to the sample combination to obtain a first prediction accuracy corresponding to the sample combination includes: Training the driving behavior prediction model according to the sample combination to obtain a first driving behavior prediction model, as well as a precision and a recall rate of the first driving behavior prediction model; The first prediction accuracy corresponding to the sample combination is obtained according to the precision and the recall.

5. The method according to claim 1, wherein The step of constructing the Nth vehicle parameter sample according to the vehicle operation parameter sample includes: Obtaining vehicle interior and exterior environmental parameters; The Nth vehicle parameter sample is constructed according to the vehicle operation parameter sample, the in-vehicle environmental parameter, and the out-vehicle environmental parameter.

6. The method according to claim 1, characterized in that The importance value of the Nth target vehicle parameter category is smaller than the importance values ​​of other vehicle parameter categories in the Nth vehicle parameter sample except the Nth target vehicle parameter category.

7. A driving behavior prediction method, characterized in that: Applied to the cloud, the method includes: Receive vehicle parameters to be processed sent by the vehicle; The parameters of the vehicle to be processed are input into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

8. A driving behavior prediction method, characterized in that: Applied to a vehicle, the method comprises: Get the parameters of the vehicle to be processed; The vehicle parameters to be processed are sent to the cloud, so that the cloud can input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

9. A driving behavior prediction model training method, characterized in that: include: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

10. A driving behavior prediction device, characterized in that: include: A first acquisition module is used to acquire vehicle parameters to be processed of the vehicle; The second acquisition module is configured to input the vehicle parameters to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

11. A driving behavior prediction device, characterized in that: Applied to the cloud, the device includes: A third acquisition module is used to receive the vehicle parameters to be processed sent by the vehicle; A fourth acquisition module is configured to input the vehicle parameters to be processed into a target driving behavior prediction model to obtain a predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

12. A driving behavior prediction device, characterized in that: Applied to a vehicle, the device comprises: A fifth acquisition module, used to acquire parameters of the vehicle to be processed; A sending module is used to send the vehicle parameters to be processed to the cloud, so that the cloud can input the vehicle parameters to be processed into a target driving behavior prediction model to obtain the predicted driving behavior of the vehicle; wherein the target driving behavior prediction model is obtained by the following method: Deleting vehicle parameters corresponding to an Nth target vehicle parameter category from an Nth vehicle parameter sample to obtain an N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of a plurality of vehicle parameter categories, and the Nth target vehicle parameter category is determined based on the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], where n is an integer greater than or equal to 2; Obtaining first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, respectively, and using the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; obtaining second target vehicle parameter samples of the same vehicle parameter category as the first target vehicle parameter samples corresponding to a preset prediction accuracy, and using the second target vehicle parameter samples to train the driving behavior prediction model to obtain a target driving behavior prediction model; When N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operating parameter time series; the vehicle parameter categories of the vehicle operating parameters in different vehicle operating parameter time series are different from each other; respectively obtaining the time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; the time when the vehicle operating parameters in the target vehicle operating parameter time series change is earlier than the time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample with the same vehicle parameter category as that in the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample.

13. A driving behavior prediction model training device, characterized in that: include: The sixth acquisition module is used to delete the vehicle parameters corresponding to the Nth target vehicle parameter category from the Nth vehicle parameter sample to obtain the N+1th vehicle parameter sample; wherein the Nth vehicle parameter sample is composed of vehicle parameters of multiple vehicle parameter categories, and the Nth target vehicle parameter category is determined according to the importance value of each vehicle parameter category in the Nth vehicle parameter sample; N∈[1,n], n is an integer greater than or equal to 2; in the case of N=1, the Nth vehicle parameter sample is obtained by the following method: obtaining a driving behavior time series and at least one vehicle operation parameter time series; different vehicles in the vehicle operation parameter time series The vehicle parameter categories of the operating parameters are different from each other; respectively obtaining a time difference correlation coefficient between the driving behavior time series and each of the vehicle operating parameter time series; obtaining at least one target vehicle operating parameter time series from at least one vehicle operating parameter time series based on the time difference correlation coefficient; a time when the vehicle operating parameter in the target vehicle operating parameter time series changes is earlier than a time when the driving behavior changes in the driving behavior time series; obtaining a vehicle operating parameter sample of the same vehicle parameter category as the target vehicle operating parameter time series; and constructing the Nth vehicle parameter sample based on the vehicle operating parameter sample. a seventh acquisition module, configured to respectively acquire first target vehicle parameter samples of the same vehicle parameter category as each vehicle parameter sample, and use the first target vehicle parameter samples to train a driving behavior prediction model to obtain a prediction accuracy corresponding to the first target vehicle parameter samples; The eighth acquisition module is used to obtain a second target vehicle parameter sample of the same vehicle parameter category as the first target vehicle parameter sample corresponding to the preset prediction accuracy, and use the second target vehicle parameter sample to train the driving behavior prediction model to obtain a target driving behavior prediction model.

14. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to execute the instructions to implement the method according to any one of claims 1 to 9.

15. A computer-readable storage medium, characterized in that When the instructions in the storage medium are executed by a processor of a mobile terminal, the mobile terminal is enabled to execute the method according to any one of claims 1 to 9.

Images