Braking intention recognition method, system and equipment and medium

Abstract

The invention provides a braking intention recognition method, system and device and a medium, and the method comprises the steps: carrying out the preprocessing and clustering processing of driving data, so as to obtain a clustering data set; classifying the driving scene data according to the scene parameters to obtain a classified scene data set; performing selection combination on the clustering data set and the classification scene data set to obtain a selection feature set; performing performance evaluation on the final prediction model to obtain an optimal time step, determining a corresponding feature selection set based on the optimal time step, inputting the corresponding feature selection set into the final prediction model to perform data prediction so as to output a prediction braking mode, and performing corresponding prediction braking intensity based on the prediction braking mode; according to the method, the recognition capability and accuracy of the model in a complex driving scene can be enhanced, and meanwhile, the accuracy of the output final braking strength is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of braking intention recognition, and specifically relates to a braking intention recognition method, system, equipment and medium. Background Art

[0002] With the rapid development of intelligent transportation systems and autonomous driving technology, accurately identifying and predicting the driver's braking intention is of great significance for improving driving safety, optimizing vehicle driving strategies, and formulating corresponding energy recovery management strategies. Traditional braking intention recognition methods mainly rely on single sensor data or simple threshold judgments, which often cannot fully reflect the complex driving environment and driver behavior characteristics, resulting in limited recognition accuracy.

[0003] In recent years, with the continuous advancement of big data technology and machine learning algorithms, the braking intention recognition method based on multi-source data fusion has gradually become a research hotspot. This type of method can more comprehensively capture various information during driving by collecting and processing data from the vehicle's own sensors, navigation systems, etc., thereby achieving more accurate braking intention recognition. However, in practical applications, how to effectively integrate and utilize this data, and how to build appropriate recognition models for different driving scenarios, are still the main challenges currently faced.

[0004] In the prior art, although some braking intention recognition methods have attempted to use the vehicle's driving data and navigation data for analysis, these methods often lack detailed division and targeted processing of different driving scenarios. For example, some methods simply divide the braking mode into several fixed modes without considering the potential impact of changes in the driving environment on the driver's braking mode at the next moment. In addition, existing studies usually use the characteristic parameters of the previous second to predict the braking intention of the next second. This method does not take into account the continuity and timing dependence of the driver's behavior, and the braking intensity of the next moment predicted by the existing braking intention recognition method is inaccurate and has low precision. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a braking intention recognition method, system, device and medium, which are used to solve the technical problems in the prior art.

[0006] On the one hand, the present invention provides the following technical solution, a braking intention recognition method, comprising: Acquire driving data of the vehicle within a preset historical time period, and perform preprocessing and clustering processing on the driving data to obtain a cluster data set; Acquire driving scene data of the vehicle within a preset historical time period, and classify the driving scene data according to scene parameters to obtain a classified scene data set; Selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set; Acquire training data and train a preset prediction model according to the training data to obtain a final prediction model, perform performance evaluation on the final prediction model to obtain an optimal time step, determine a corresponding feature selection set based on the optimal time step and input the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode; A road surface limit braking intensity of the vehicle is calculated based on the driving data, and a final braking intensity is determined based on the road surface limit braking intensity and the predicted braking intensity.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention first obtains the driving data of the vehicle within a preset historical time period, pre-processes and clusters the driving data to obtain a clustering data set; then obtains the driving scene data of the vehicle within the preset historical time period, and classifies the driving scene data according to the scene parameters to obtain a classified scene data set; then the clustering data set and the classified scene data set are selected and combined to obtain a selected feature set; then the training data are obtained and the preset prediction model is trained according to the training data to obtain a final prediction model, the performance of the final prediction model is evaluated to obtain an optimal time step, and the corresponding feature is determined based on the optimal time step. A selection set is selected and the corresponding feature selection set is input into the final prediction model to predict data to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode is obtained; finally, the road surface limit braking intensity of the vehicle is calculated based on the driving data, and the final braking intensity is determined based on the road surface limit braking intensity and the predicted braking intensity. The present invention integrates multi-scenario data and personalized feature selection to enhance the recognition ability and accuracy of the model in complex driving scenarios, and determines the optimal time step to improve the recognition performance of the model. After that, the present invention determines the road surface limit braking intensity and fully considers the influence of different roads on the braking intensity to improve the accuracy of the output final braking intensity.

[0008] Preferably, the step of preprocessing and clustering the driving data to obtain a clustered data set includes: Identifying duplicate data and mutation data in the driving data, and removing the duplicate data and mutation data to obtain removed data; Identify missing values ​​in the removed data, and fill the missing values ​​using linear interpolation to obtain filled data; Extracting the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate from the filling data, and determining whether the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate are 0 at a certain moment; If at a certain moment the brake pedal travel value, the brake pedal travel value change rate, the accelerator pedal travel value, and the accelerator pedal travel value change rate are 0, the data at the corresponding moment is filtered out from the filling data to obtain the filtered data; The K-means clustering method is used to cluster the screening data to obtain a clustered data set.

[0009] Preferably, the step of acquiring the driving scene data of the vehicle within a preset historical time period and classifying the driving scene data according to scene parameters to obtain a classified scene data set includes: Obtain the vehicle's current speed limit, distance to the speed limit sign, distance to the traffic light, status of the next traffic light, uphill and downhill status, and real-time longitudinal distance to the vehicle in front within a preset historical time period to obtain driving scenario data; Identify the validity of the driving scene data, and classify the driving scene data according to the validity of the driving scene data and a preset judgment rule to obtain single scene data, wherein the preset judgment rule is: Following vehicle scenario: When the real-time longitudinal vehicle distance is valid, the vehicle is in the following vehicle scenario; Traffic light scenario: When the distance to the traffic light and the next traffic light status are valid, the vehicle is in the traffic light scenario; Downhill scenario: When the uphill and downhill states are valid, the vehicle is in an uphill and downhill scenario; Speed ​​limit sign scenario: When the distance to the speed limit sign and the current speed limit are valid, the vehicle is in the speed limit sign scenario; The single scene data are combined to obtain a classified scene data set, wherein the classified scene data set includes full scene, speed limit sign + uphill and downhill + traffic light scene, speed limit sign + uphill and downhill + following car scene, speed limit sign + uphill and downhill scene, speed limit sign scene, uphill and downhill + following car scene, speed limit sign + following car scene, and no scene.

[0010] Preferably, the step of selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set includes: Adding the clustered data set sample to each scene in the classified scene data set to obtain an added data set; The Spearman correlation coefficient method is used to perform a correlation test on the added data set in each scenario, and the added data set that passes the test is used as the selected feature set.

[0011] Preferably, the step of performing performance evaluation on the final prediction model to obtain an optimal time step, determining a corresponding feature selection set based on the optimal time step and inputting the corresponding feature selection set into the final prediction model to perform prediction data to output a predicted braking mode includes: Input the selected feature set corresponding to the first k time steps into the final prediction model and calculate the SHAP value corresponding to each time step; Select the time step corresponding to the maximum SHAP value as the optimal time step K; The selected feature set corresponding to the first K time steps is extracted and input into the final prediction model to predict data, so as to output the predicted braking mode.

[0012] Preferably, the step of calculating the road surface limit braking intensity of the vehicle based on the driving data, and determining the final braking intensity based on the road surface limit braking intensity and the predicted braking intensity comprises: Calculate vehicle load based on the driving data Braking force : ; ; In the formula, For vehicle weight, is the dynamic load of the car, is the braking strength moment, is the wheel moment of inertia, is the wheel speed, is the wheel radius; Calculate the road friction value based on the driving data : ; In the formula, Indicates the vehicle speed; Based on the vehicle load With the braking force Calculate road braking value : ; Based on the road braking value Friction value with the road surface Determine the fitting formula: ; In the formula, denote the first, second, and third fitting parameters, respectively; Obtain a standard road surface set and determine a road surface braking value and a road surface friction value corresponding to each road surface in the standard road surface set, and input the road surface braking value and the road surface friction value corresponding to each road surface into a fitting formula to update and fit the first, second, and third fitting parameters to obtain a first updated fitting parameter , the second update fitting parameters , third update fitting parameters ; Update the fitting parameters based on the first , the second update fitting parameters , third update fitting parameters Determine the final braking intensity.

[0013] Preferably, the first update fitting parameter , the second update fitting parameters , third update fitting parameters The steps to determine the final braking intensity include: Update the fitting parameters based on the first , the second update fitting parameters , third update fitting parameters Determine the initial braking value : ; Calculate the similarity between the current vehicle's driving road and each road in the standard road set, and use the similarity and initial braking value to calculate the similarity between the current vehicle's driving road and each road in the standard road set. Calculate and update the braking value : ; In the formula, Indicates the current vehicle's driving road surface and the standard road surface concentration The similarity of the road surface, Indicates the number of concentrated pavement of standard pavement; Based on the updated braking value Braking value with road surface Determine the road surface limit for braking intensity : ; In the formula, represents the fusion weight; Limit the braking intensity based on the road surface The predicted braking intensity Determine the final braking strength : .

[0014] In a second aspect, the present invention provides the following technical solution, a braking intention recognition system, the system comprising: A processing module, used to obtain the driving data of the vehicle within a preset historical time period, and perform preprocessing and clustering processing on the driving data to obtain a cluster data set; A classification module is used to obtain driving scene data of the vehicle within a preset historical time period, and classify the driving scene data according to scene parameters to obtain a classified scene data set; A combination module, used for selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set; A prediction module, used to obtain training data and train a preset prediction model according to the training data to obtain a final prediction model, perform performance evaluation on the final prediction model to obtain an optimal time step, determine a corresponding feature selection set based on the optimal time step and input the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode; The output module is used to calculate the road surface limit braking intensity of the vehicle based on the driving data, and determine the final braking intensity based on the road surface limit braking intensity and the predicted braking intensity.

[0015] In a third aspect, the present invention provides the following technical solution: a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the braking intention recognition method as described above when executing the computer program.

[0016] In a fourth aspect, the present invention provides the following technical solution: a storage medium having a computer program stored thereon, wherein the computer program implements the braking intention recognition method as described above when executed by a processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flowchart of a braking intention recognition method provided in Embodiment 1 of the present invention; Figure 2 A structural block diagram of a braking intention recognition system provided in Embodiment 2 of the present invention; Figure 3A schematic diagram of the hardware structure of a computer provided in another embodiment of the present invention.

[0019] The embodiments of the present invention will be further described below with reference to the accompanying drawings. DETAILED DESCRIPTION

[0020] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the embodiments of the present invention, and should not be construed as limiting the present invention.

[0021] Embodiment 1 In the first embodiment of the present invention, Figure 1 As shown, a braking intention recognition method includes: S1. Acquire driving data of a vehicle within a preset historical time period, and perform preprocessing and clustering processing on the driving data to obtain a cluster data set; The driving data here at least includes brake pedal travel value, brake pedal travel value change rate, accelerator pedal travel value, accelerator pedal travel value change rate, vehicle speed, acceleration and deceleration, etc.; At the same time, the data collection and transmission standards here follow the "GB / T 32960 Technical Specifications for Electric Vehicle Remote Service and Management System", and the data transmission architecture is: vehicle terminal → enterprise platform → government public platform → national public platform. Through the on-board diagnostic system (OBD) interface, the vehicle speed, speed, GPS positioning and other information are collected and uploaded to the server in real time for storage in the big data center. The data collection frequency is set to 10Hz, and the collected data fields include: data collection time, vehicle status, charging status, vehicle speed, cumulative mileage, total voltage, total current, SOC (battery power status), brake pedal travel value, brake pedal travel value change rate, accelerator pedal travel value, accelerator pedal travel value change rate, acceleration and deceleration.

[0022] Wherein, the step S1 comprises: S11, identifying repeated data and mutation data in the driving data, and removing the repeated data and mutation data to obtain removed data.

[0023] S12, identifying missing values ​​in the removed data, and filling the missing values ​​using a linear interpolation method to obtain filled data; Specifically, the algorithms for identifying duplicate data and mutant data here are commonly used algorithms in the prior art, so they are not repeated in this embodiment. At the same time, when determining the fill-in data, it is also necessary to filter the data according to the vehicle status and charging status, and eliminate the data during the charging process and the data when the vehicle is not started.

[0024] S13. Extract the brake pedal stroke value, brake pedal stroke value change rate, accelerator pedal stroke value, and accelerator pedal stroke value change rate from the filled data, and determine whether the brake pedal stroke value, brake pedal stroke value change rate, accelerator pedal stroke value, and accelerator pedal stroke value change rate are 0 at a certain moment.

[0025] S14, if at a certain moment the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate are 0, then the data at the corresponding moment is filtered out from the filling data to obtain the filtered data; Specifically, when the brake pedal travel value, the brake pedal travel value change rate, the accelerator pedal travel value, and the accelerator pedal travel value change rate are 0 at a certain moment, it means that the vehicle is in a gliding state at that moment, and there is no braking or acceleration process.

[0026] S15, clustering the screened data using a K-means clustering method to obtain a clustered data set; Specifically, in this example, the preferred clusters are 4 categories, including: ① Mild pre-braking category: the brake pedal stroke value and the change rate are small, the vehicle is running smoothly, and the driver lightly touches the brake pedal to fine-tune the vehicle speed; ② Slow braking category: the brake pedal stroke value is large, the change rate is small, and the driver presses the brake pedal relatively slowly; ③ Emergency braking category: the brake pedal stroke value and the change rate are both large, and the driver quickly and forcefully presses the brake pedal; ④ Braking adjustment category: the brake pedal stroke value is large, the change rate is negative, and the driver is in a state of releasing the brake pedal.

[0027] S2, obtaining driving scene data of the vehicle within a preset historical time period, and classifying the driving scene data according to scene parameters to obtain a classified scene data set; Wherein, the step S2 comprises: S21, obtaining the current speed limit of the vehicle, the distance from the speed limit sign, the distance from the traffic light, the state of the next traffic light, the uphill and downhill state, and the real-time longitudinal distance from the vehicle in front within a preset historical time period to obtain driving scene data; S22, identifying the validity of the driving scene data, and classifying the driving scene data according to the validity of the driving scene data and a preset judgment rule to obtain single scene data, wherein the preset judgment rule is: Following vehicle scenario: When the real-time longitudinal vehicle distance is valid, the vehicle is in the following vehicle scenario; Traffic light scenario: When the distance to the traffic light and the next traffic light status are valid, the vehicle is in the traffic light scenario; Downhill scenario: When the uphill and downhill states are valid, the vehicle is in an uphill and downhill scenario; Speed ​​limit sign scenario: When the distance to the speed limit sign and the current speed limit are valid, the vehicle is in the speed limit sign scenario; Specifically, invalid data indicates that the vehicle's sensors failed to recognize the corresponding scene representation.

[0028] S23, combining the single scene data to obtain a classified scene data set, wherein the classified scene data set includes the full scene, speed limit sign + uphill and downhill + traffic light scene, speed limit sign + uphill and downhill + following car scene, speed limit sign + uphill and downhill scene, speed limit sign scene, uphill and downhill + following car scene, speed limit sign + following car scene, and no scene; Since the vehicle may be in multiple scenarios at the same time during operation, in actual situations, the operation of the vehicle at one moment may include a combination of multiple scenarios. According to statistical results, the situation where four scenarios exist at the same time accounts for more than 50%, the situation where three or more scenarios exist at the same time accounts for about 95%, and the situation where a single scene accounts for less than 0.5%. Based on this phenomenon, the classified scenario data set is divided into the following categories: full scenario, speed limit sign + uphill and downhill + traffic light scenario, speed limit sign + uphill and downhill + following vehicle scenario, speed limit sign + uphill and downhill scenario, speed limit sign scenario, uphill and downhill + following vehicle scenario, speed limit sign + following vehicle scenario, and no scenario. The full scenario means that following vehicle, traffic light, uphill and downhill, and speed limit scenarios exist at the same time; By identifying and dividing different scene combinations, targeted driving features can be extracted according to the characteristics of each scene. In the process of multi-scene data fusion, it is ensured that the driving behavior characteristics in each scene are fully explored and utilized, thereby providing more accurate input data for subsequent models.

[0029] S3, selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set; Wherein, the step S3 comprises: S31, adding the clustering data set sample to each scene in the classification scene data set to obtain an added data set; Specifically, for all scenarios, by adding different types and quantities of data in the clustering dataset to each scenario, the added dataset can be obtained. S32, using the Spearman correlation coefficient method to perform a correlation test on the added data set in each scenario, and using the added data set that passes the test as the selected feature set; Specifically, by performing correlation tests, redundant features with high correlation are eliminated to ensure the independence and validity of the input variables. This step helps to reduce the dimension of the feature set, improve the accuracy and efficiency of the feature set, and provide a more optimized feature set for the model.

[0030] S4, obtaining training data and training a preset prediction model according to the training data to obtain a final prediction model, performing performance evaluation on the final prediction model to obtain an optimal time step, determining a corresponding feature selection set based on the optimal time step and inputting the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode; Wherein, the step S4 comprises: S41, inputting the selected feature set corresponding to the first k time steps into the final prediction model and calculating the SHAP value corresponding to each time step; Specifically, the SHAP value (Shapley Additive Explanations) is an explanation method based on cooperative game theory, which is used to quantify the contribution of each feature to model prediction. Specifically, the SHAP value can calculate the average marginal contribution of each feature in all possible combinations, thereby assigning an importance score to each feature. For each k value, the average SHAP value of the selected feature set for each time step in the first k time steps is calculated and used as the SHAP value corresponding to each time step, so as to evaluate the impact of different time steps on model prediction and reveal the contribution of different time steps to braking mode prediction.

[0031] S42, selecting the time step corresponding to the maximum SHAP value as the optimal time step K; Specifically, the importance of features changes significantly with the change of time steps, and the contribution of features at certain time points is particularly prominent. Usually, time steps close to the prediction moment may contain more directly relevant information, but may also have complex negative effects on the model due to short-term fluctuations. Considering the SHAP value and model accuracy, the optimal time step is selected. Experiments have found that when K=7, the prediction accuracy and stability of the model are optimal.

[0032] S43, extracting the selected feature set corresponding to the first K time steps and inputting it into the final prediction model to perform prediction data, so as to output a predicted braking mode; Specifically, the final prediction model here is the XGBoost model, and the output predicted braking mode is the braking type clustered above. Different braking intensities can be assigned according to different braking types.

[0033] S5. Calculate a road surface limit braking intensity of the vehicle based on the driving data, and determine a final braking intensity based on the road surface limit braking intensity and the predicted braking intensity.

[0034] Wherein, the step S5 comprises: S51, calculating vehicle load based on the driving data Braking force : ; ; In the formula, For vehicle weight, is the dynamic load of the car, is the braking strength moment, is the wheel moment of inertia, is the wheel speed, is the wheel radius.

[0035] S52: Calculating road friction value based on the driving data : ; In the formula, Indicates the vehicle's speed.

[0036] S53, based on the vehicle load With the braking force Calculate road braking value : .

[0037] S54, based on the road braking value Friction value with the road surface Determine the fitting formula: ; In the formula, Represent the first, second and third fitting parameters respectively.

[0038] S55, obtaining a standard road surface set and determining a road surface braking value and a road surface friction value corresponding to each road surface in the standard road surface set, and inputting the road surface braking value and the road surface friction value corresponding to each road surface into a fitting formula to perform update fitting on the first, second and third fitting parameters, so as to obtain a first updated fitting parameter , the second update fitting parameters , third update fitting parameters ; Specifically, each road surface in the standard road surface set has a corresponding first updated fitting parameter, a second updated fitting parameter, and a third updated fitting parameter.

[0039] S56, based on the first update fitting parameters , the second update fitting parameters , third update fitting parameters Determine the final braking intensity.

[0040] Wherein, the step S51 comprises: S561, based on the first update fitting parameters , the second update fitting parameters , third update fitting parameters Determine the initial braking value : ; Specifically, each road surface in the standard road surface set has a corresponding initial braking value.

[0041] S562, calculating the similarity between the current vehicle's driving road surface and each road surface in the standard road surface set, and based on the similarity and the initial braking value Calculate and update the braking value : ; In the formula, Indicates the current vehicle's driving road surface and the standard road surface concentration The similarity of the road surface, Indicates the number of concentrated pavement of standard pavement; Specifically, the similarity here can be determined by fuzzy logic processing to determine the pixel degree, so as to deal with the driver's operation certainty and fuzziness. The fuzzy logic processing here is a common method of fuzzy controller in the prior art, so it will not be described here.

[0042] S563: Based on the updated braking value Braking value with road surface Determine the road surface limit for braking intensity : ; In the formula, represents the fusion weight; In this embodiment, the fusion weight is 0.3.

[0043] S564: Limiting braking intensity based on the road surface The predicted braking intensity Determine the final braking strength : ; Specifically, the final braking intensity here has several ranges. When the final braking intensity is between 0-0.25, it is low-intensity braking. When the final braking intensity is between 0.25-0.5, it is medium-intensity braking. When the final braking intensity is between 0.5-0.75, it is high-intensity braking. When the final braking intensity is greater than 0.75, it is emergency braking.

[0044] The braking intention recognition method provided in the first embodiment of the present invention first obtains the driving data of the vehicle within a preset historical time period, pre-processes and clusters the driving data to obtain a cluster data set; then obtains the driving scene data of the vehicle within the preset historical time period, and classifies the driving scene data according to the scene parameters to obtain a classified scene data set; then the cluster data set and the classified scene data set are selected and combined to obtain a selected feature set; then the training data is obtained and the preset prediction model is trained according to the training data to obtain a final prediction model, the performance of the final prediction model is evaluated to obtain an optimal time step, and the corresponding feature is determined based on the optimal time step. A selection set is selected and the corresponding feature selection set is input into the final prediction model to predict data to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode is obtained; finally, the road surface limit braking intensity of the vehicle is calculated based on the driving data, and the final braking intensity is determined based on the road surface limit braking intensity and the predicted braking intensity. The present invention integrates multi-scenario data and personalized feature selection to enhance the recognition ability and accuracy of the model in complex driving scenarios, and determines the optimal time step to improve the recognition performance of the model. After that, the present invention determines the road surface limit braking intensity and fully considers the influence of different roads on the braking intensity to improve the accuracy of the output final braking intensity.

[0045] Embodiment 2 like Figure 2 As shown, in the second embodiment of the present invention, a braking intention recognition system is provided, and the system includes: Processing module 1, used to obtain the driving data of the vehicle within a preset historical time period, and perform preprocessing and clustering processing on the driving data to obtain a cluster data set; Classification module 2, used to obtain driving scene data of the vehicle within a preset historical time period, and classify the driving scene data according to scene parameters to obtain a classified scene data set; A combination module 3, used for selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set; Prediction module 4, used for acquiring training data and training a preset prediction model according to the training data to obtain a final prediction model, performing performance evaluation on the final prediction model to obtain an optimal time step, determining a corresponding feature selection set based on the optimal time step and inputting the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode; An output module 5, configured to calculate a road surface limit braking intensity of the vehicle based on the driving data, and determine a final braking intensity based on the road surface limit braking intensity and the predicted braking intensity; The processing module 1 comprises: A removal submodule, used for identifying duplicate data and mutation data in the driving data, and removing the duplicate data and mutation data to obtain removed data; A filling submodule is used to identify missing values ​​in the removed data and fill the missing values ​​using a linear interpolation method to obtain filled data; an extraction submodule, for extracting the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate from the filled data, and determining whether the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate are 0 at a certain moment; a screening submodule, for filtering out the data at the corresponding moment from the filling data if the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate are 0 at a certain moment, so as to obtain the screening data; The clustering submodule is used to perform clustering processing on the screening data using a K-means clustering method to obtain a clustered data set.

[0046] The classification module 2 comprises: The acquisition submodule is used to obtain the current speed limit of the vehicle, the distance from the speed limit sign, the distance from the traffic light, the state of the next traffic light, the uphill and downhill state, and the real-time longitudinal distance from the vehicle in front within a preset historical time period to obtain driving scene data; The classification submodule is used to identify the validity of the driving scene data, and classify the driving scene data according to the validity of the driving scene data and a preset judgment rule to obtain single scene data, wherein the preset judgment rule is: Following vehicle scenario: When the real-time longitudinal vehicle distance is valid, the vehicle is in the following vehicle scenario; Traffic light scenario: When the distance to the traffic light and the next traffic light status are valid, the vehicle is in the traffic light scenario; Downhill scenario: When the uphill and downhill states are valid, the vehicle is in an uphill and downhill scenario; Speed ​​limit sign scenario: When the distance to the speed limit sign and the current speed limit are valid, the vehicle is in the speed limit sign scenario; The combination submodule is used to combine single scene data to obtain a classified scene data set, wherein the classified scene data set includes full scene, speed limit sign + uphill and downhill + traffic light scene, speed limit sign + uphill and downhill + following car scene, speed limit sign + uphill and downhill scene, speed limit sign scene, uphill and downhill + following car scene, speed limit sign + following car scene, and no scene.

[0047] The combined module 3 comprises: An adding submodule, used for adding the clustering data set sample to each scene in the classification scene data set to obtain an added data set; The test submodule is used to perform a correlation test on the added data set in each scenario using the Spearman correlation coefficient method, and the added data set that passes the test is used as the selected feature set.

[0048] The prediction module 4 comprises: A SHAP value submodule, used to input the selected feature set corresponding to the first k time steps into the final prediction model and calculate the SHAP value corresponding to each time step; The time step submodule is used to select the time step corresponding to the maximum SHAP value as the optimal time step K; The prediction submodule is used to extract the selected feature set corresponding to the first K time steps and input it into the final prediction model to predict data, so as to output the predicted braking mode.

[0049] The output module 5 comprises: A first calculation submodule is used to calculate the vehicle load based on the driving data Braking force : ; ; In the formula, For vehicle weight, is the dynamic load of the car, is the braking strength moment, is the wheel moment of inertia, is the wheel speed, is the wheel radius; The second calculation submodule is used to calculate the road friction value based on the driving data : ; In the formula, Indicates the vehicle speed; The third calculation submodule is used to calculate the vehicle load based on the vehicle load. With the braking force Calculate road braking value : ; A fourth calculation submodule, for submodules, for calculating the braking value of the road surface based on the road surface braking value. Friction value with the road surface Determine the fitting formula: ; In the formula, denote the first, second, and third fitting parameters, respectively; The fitting submodule is used to obtain a standard road surface set and determine a road surface braking value and a road surface friction value corresponding to each road surface in the standard road surface set, and input the road surface braking value and the road surface friction value corresponding to each road surface into a fitting formula to update and fit the first, second and third fitting parameters to obtain a first updated fitting parameter , the second update fitting parameters , third update fitting parameters ; An output submodule for updating fitting parameters based on the first , the second update fitting parameters , third update fitting parameters Determine the final braking intensity.

[0050] The output submodule comprises: A first calculation unit is used to update the fitting parameters based on the first , the second update fitting parameters , third update fitting parameters Determine the initial braking value : ; The second calculation unit is used to calculate the similarity between the current vehicle's driving road surface and each road surface in the standard road surface set, and calculate the similarity between the current vehicle's driving road surface and the initial braking value according to the similarity. Calculate and update the braking value : ; In the formula, Indicates the current vehicle's driving road surface and the standard road surface concentration The similarity of the road surface, Indicates the number of concentrated pavement of standard pavement; A third calculation unit is used to update the braking value based on the Braking value with road surface Determine the road surface limit for braking intensity : ; In the formula, represents the fusion weight; A fourth calculation unit, configured to limit the braking intensity based on the road surface The predicted braking intensity Determine the final braking strength : .

[0051] In some other embodiments of the present invention, the embodiments of the present invention provide the following technical solution: a computer device, comprising a memory 102, a processor 101, and a computer program stored in the memory 102 and executable on the processor 101, wherein the processor 101 implements the braking intention recognition method as described above when executing the computer program.

[0052] Specifically, the processor 101 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiment of the present invention.

[0053] Among them, the memory 102 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 102 may include a hard disk drive (HDD), a floppy disk drive, a solid state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 102 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 102 may be inside or outside the data processing device. In a specific embodiment, the memory 102 is a non-volatile memory. In a specific embodiment, the memory 102 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0054] The memory 102 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 101 .

[0055] The processor 101 implements the above-mentioned braking intention recognition method by reading and executing the computer program instructions stored in the memory 102.

[0056] In some embodiments, the computer device may further include a communication interface 103 and a bus 100. Figure 3 As shown, the processor 101, the memory 102, and the communication interface 103 are connected via a bus 100 and communicate with each other.

[0057] The communication interface 103 is used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present invention. The communication interface 103 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0058] The bus 100 includes hardware, software or both, and couples the components of the computer device to each other. The bus 100 includes but is not limited to at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 100 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, bus 100 may include one or more buses. Although embodiments of the present invention describe and illustrate a particular bus, the present invention contemplates any suitable bus or interconnect.

[0059] The computer can execute the braking intention recognition method of the present invention based on the acquired braking intention recognition system, thereby realizing braking intention recognition.

[0060] In some further embodiments of the present invention, in combination with the above-mentioned braking intention recognition method, the embodiments of the present invention provide the following technical solution, a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned braking intention recognition method is implemented.

[0061] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0062] More specific examples of readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0063] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0064] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0065] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A braking intention recognition method, characterized in that: include: Acquire driving data of the vehicle within a preset historical time period, and perform preprocessing and clustering processing on the driving data to obtain a cluster data set; Acquire driving scene data of the vehicle within a preset historical time period, and classify the driving scene data according to scene parameters to obtain a classified scene data set; Selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set; Acquire training data and train a preset prediction model according to the training data to obtain a final prediction model, perform performance evaluation on the final prediction model to obtain an optimal time step, determine a corresponding feature selection set based on the optimal time step and input the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode; A road surface limit braking intensity of the vehicle is calculated based on the driving data, and a final braking intensity is determined based on the road surface limit braking intensity and the predicted braking intensity.

2. The braking intention recognition method according to claim 1, characterized in that: The step of preprocessing and clustering the driving data to obtain a clustered data set includes: Identifying duplicate data and mutation data in the driving data, and removing the duplicate data and mutation data to obtain removed data; Identify missing values ​​in the removed data, and fill the missing values ​​using linear interpolation to obtain filled data; Extracting the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate from the filling data, and determining whether the brake pedal stroke value, the brake pedal stroke value change rate, the accelerator pedal stroke value, and the accelerator pedal stroke value change rate are 0 at a certain moment; If at a certain moment the brake pedal travel value, the brake pedal travel value change rate, the accelerator pedal travel value, and the accelerator pedal travel value change rate are 0, the data at the corresponding moment is filtered out from the filling data to obtain the filtered data; The K-means clustering method is used to cluster the screening data to obtain a clustered data set.

3. The braking intention recognition method according to claim 1, characterized in that: The step of acquiring the driving scene data of the vehicle within a preset historical time period and classifying the driving scene data according to the scene parameters to obtain a classified scene data set includes: Obtain the vehicle's current speed limit, distance to the speed limit sign, distance to the traffic light, status of the next traffic light, uphill and downhill status, and real-time longitudinal distance to the vehicle in front within a preset historical time period to obtain driving scenario data; Identify the validity of the driving scene data, and classify the driving scene data according to the validity of the driving scene data and a preset judgment rule to obtain single scene data, wherein the preset judgment rule is: Following vehicle scenario: When the real-time longitudinal vehicle distance is valid, the vehicle is in the following vehicle scenario; Traffic light scenario: When the distance to the traffic light and the next traffic light status are valid, the vehicle is in the traffic light scenario; Downhill scenario: When the uphill and downhill states are valid, the vehicle is in an uphill and downhill scenario; Speed ​​limit sign scenario: When the distance to the speed limit sign and the current speed limit are valid, the vehicle is in the speed limit sign scenario; The single scene data are combined to obtain a classified scene data set, wherein the classified scene data set includes full scene, speed limit sign + uphill and downhill + traffic light scene, speed limit sign + uphill and downhill + following car scene, speed limit sign + uphill and downhill scene, speed limit sign scene, uphill and downhill + following car scene, speed limit sign + following car scene, and no scene.

4. The braking intention recognition method according to claim 1, characterized in that: The step of selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set comprises: Adding the clustered data set sample to each scene in the classified scene data set to obtain an added data set; The Spearman correlation coefficient method is used to perform a correlation test on the added data set in each scenario, and the added data set that passes the test is used as the selected feature set.

5. The braking intention recognition method according to claim 1, characterized in that: The step of performing performance evaluation on the final prediction model to obtain an optimal time step, determining a corresponding feature selection set based on the optimal time step, and inputting the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode comprises: Input the selected feature set corresponding to the first k time steps into the final prediction model and calculate the SHAP value corresponding to each time step; Select the time step corresponding to the maximum SHAP value as the optimal time step K; The selected feature set corresponding to the first K time steps is extracted and input into the final prediction model to predict data, so as to output the predicted braking mode.

6. The braking intention recognition method according to claim 1, characterized in that: The step of calculating the road surface limit braking intensity of the vehicle based on the driving data and determining the final braking intensity based on the road surface limit braking intensity and the predicted braking intensity comprises: Calculate vehicle load based on the driving data Braking force : ； ； In the formula, For vehicle weight, is the dynamic load of the car, is the braking strength moment, is the wheel moment of inertia, is the wheel speed, is the wheel radius; Calculate the road friction value based on the driving data : ； In the formula, Indicates the vehicle speed; Based on the vehicle load With the braking force Calculate road braking value : ； Based on the road braking value Friction value with the road surface Determine the fitting formula: ； In the formula, denote the first, second, and third fitting parameters, respectively; Obtain a standard road surface set and determine a road surface braking value and a road surface friction value corresponding to each road surface in the standard road surface set, and input the road surface braking value and the road surface friction value corresponding to each road surface into a fitting formula to update and fit the first, second, and third fitting parameters to obtain a first updated fitting parameter , the second update fitting parameters , third update fitting parameters ; Update the fitting parameters based on the first , the second update fitting parameters , third update fitting parameters Determine the final braking intensity.

7. The braking intention recognition method according to claim 6, characterized in that: The first update fitting parameter , the second update fitting parameters , third update fitting parameters The steps to determine the final braking intensity include: Update the fitting parameters based on the first , the second update fitting parameters , third update fitting parameters Determine the initial braking value : ； Calculate the similarity between the current vehicle's driving road and each road in the standard road set, and use the similarity and initial braking value to calculate the similarity between the current vehicle's driving road and each road in the standard road set. Calculate and update the braking value : ； In the formula, Indicates the current vehicle's driving road surface and the standard road surface concentration The similarity of the road surface, Indicates the number of concentrated pavement of standard pavement; Based on the updated braking value Braking value with road surface Determine the road surface limit for braking intensity : ； In the formula, represents the fusion weight; Limit the braking intensity based on the road surface The predicted braking intensity Determine the final braking strength : 。 8. A braking intention recognition system, characterized in that: The system comprises: A processing module, used to obtain the driving data of the vehicle within a preset historical time period, and perform preprocessing and clustering processing on the driving data to obtain a cluster data set; A classification module is used to obtain driving scene data of the vehicle within a preset historical time period, and classify the driving scene data according to scene parameters to obtain a classified scene data set; A combination module, used for selecting and combining the clustering data set and the classification scene data set to obtain a selected feature set; A prediction module, used to obtain training data and train a preset prediction model according to the training data to obtain a final prediction model, perform performance evaluation on the final prediction model to obtain an optimal time step, determine a corresponding feature selection set based on the optimal time step and input the corresponding feature selection set into the final prediction model to predict data, so as to output a predicted braking mode, and a predicted braking intensity corresponding to the predicted braking mode; The output module is used to calculate the road surface limit braking intensity of the vehicle based on the driving data, and determine the final braking intensity based on the road surface limit braking intensity and the predicted braking intensity.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the braking intention recognition method according to any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the braking intention recognition method according to any one of claims 1 to 7 is implemented.

Images