Driving style determination method and device, electronic equipment and vehicle

Abstract

The invention relates to a driving style determination method and device, electronic equipment and a vehicle, relates to the technical field of data processing, and at least solves the technical problems of low accuracy and poor recognition efficiency of recognition of the driving style of the vehicle in the related art, and the method comprises the steps: obtaining the driving data of a plurality of data types of the vehicle; determining a first frequency domain characteristic parameter corresponding to the driving data of each data type in the driving data of the multiple data types, wherein the first frequency domain characteristic parameter is used for representing a frequency domain characteristic; inter-sequence correlation learning is carried out on the frequency domain characteristic parameters corresponding to the driving data of the multiple data types, second frequency domain characteristic parameters corresponding to the driving data of each data type are determined, and the second frequency domain characteristic parameters are fused with the frequency domain characteristics of the driving data of other data types; and determining a driving style corresponding to the vehicle based on the second frequency domain characteristic parameter corresponding to the driving data of each data type. The method is used for improving the accuracy and the recognition efficiency of recognizing the driving style.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and particularly relates to a method, apparatus, electronic device, and vehicle for determining a driving style. Background Art

[0002] With the rapid development of the vehicle industry, intelligentization, interconnection, and full utilization of data resources have become new trends in the development of the vehicle industry. Intelligent Connected Vehicles (ICVs) are becoming the focus, and the Advanced Driving Assistance System (ADAS) is a core component of ICVs. It can not only significantly improve the convenience and safety of a driver driving a vehicle, but also enhance the ability to mitigate safety risks and provide safety protection for the driver. Currently, accurately identifying the driving style plays a crucial role in the continuous development of ADAS. Identifying the driving style helps to optimize the functions of ADAS to make adaptive adjustments according to the individual behavior of the driver and promote a personalized driving experience. For example, by identifying the driving style, ADAS can adjust the power output, braking response, and other control parameters of the vehicle to suit the individual driving habits of the driver.

[0003] However, traditional methods for identifying the driving style mainly rely on methods such as questionnaires, simulated driving scenarios, and subjective human evaluations. These methods are prone to serious subjective biases, thus affecting the accuracy of driving style identification. In addition, these methods are resource-intensive, requiring a large amount of manpower for data collection and analysis. Once the driving style is classified, it is relatively difficult to modify and adjust later, thus affecting the feasibility of long-term driving style identification. Therefore, the current accuracy of identifying the driving style is low and the identification efficiency is poor. Summary of the Invention

[0004] The present application provides a method, apparatus, electronic device, and vehicle for determining a driving style to at least solve the technical problems of low accuracy and poor identification efficiency in identifying the driving style in the related art. The technical solution of the present application is as follows: According to a first aspect provided by the present application, a driving style determination method is provided. The method includes: obtaining driving data of multiple data types of a vehicle; determining a first frequency domain feature parameter corresponding to the driving data of each data type among the driving data of multiple data types, where the first frequency domain feature parameter is used to characterize the frequency domain feature; by performing inter-sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types, determining a second frequency domain feature parameter corresponding to the driving data of each data type, where the second frequency domain feature parameter integrates the frequency domain features of the driving data of other data types; and determining the driving style corresponding to the vehicle based on the second frequency domain feature parameter corresponding to the driving data of each data type.

[0005] According to the above technical means, the present application can determine the first frequency domain feature parameter corresponding to the driving data of each data type among the driving data of multiple data types of the vehicle to characterize the frequency domain feature of the driving data of each data type. Then, perform inter-sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types. In this way, the frequency domain features of the driving data of other data types can be integrated into the frequency domain feature parameters corresponding to the driving data of each data type, and the second frequency domain feature parameter corresponding to the driving data of each data type can be obtained. Based on this, the correlation between the driving data of each data type and the driving data of other data types can be characterized by the second frequency domain feature parameter corresponding to the driving data of each data type. Furthermore, based on the second frequency domain feature parameter corresponding to the driving data of each data type, the driving style corresponding to the vehicle can be determined. That is, by integrating the frequency domain features of the driving data of multiple data types, and based on the frequency domain feature parameters obtained after integration, the characteristics of the driving data of the vehicle can be reflected from the overall data, so that the driving style of the vehicle as a whole can be determined more accurately and efficiently.

[0006] In a possible implementation manner, the first frequency domain feature parameter includes a spectrum parameter; the above determination of the first frequency domain feature parameter corresponding to the driving data of each data type among the driving data of multiple data types includes: for the driving data of each data type among the driving data of multiple data types, performing a frequency domain transformation on the driving data of each data type based on a preset frequency to obtain the spectrum parameter corresponding to the driving data of each data type.

[0007] According to the above technical means, the present application can perform a frequency domain transformation on the driving data of each data type among the driving data of multiple data types based on a preset frequency, so as to obtain the spectrum parameter corresponding to the driving data of each data type. In this way, through the frequency domain transformation, the obtained time-domain driving data can be transformed into the frequency domain, and based on the frequency domain data (i.e., the spectrum parameter), the data characteristics can be determined more accurately, and the data can be processed more accurately and efficiently.

[0008] In a possible implementation manner, the above-mentioned determination of the second frequency-domain feature parameter corresponding to the driving data of each data type by performing inter-sequence correlation learning on the first frequency-domain feature parameters corresponding to the driving data of multiple data types includes: constructing a spectrum vector corresponding to the driving data of multiple data types based on the first frequency-domain feature parameters corresponding to the driving data of each data type; determining a query vector, a key vector, and a value vector corresponding to the spectrum vector; determining the attention weight between the query vector and the key vector; and determining the second frequency-domain feature parameter corresponding to the driving data of each data type based on the attention weight and the value vector.

[0009] According to the above technical means, the present application can construct a spectrum vector corresponding to the driving data of multiple data types based on the first frequency-domain feature parameters corresponding to the driving data of each data type. The overall features corresponding to the driving data of multiple data types are represented by the spectrum vector. Furthermore, inter-sequence correlation learning is performed through the attention mechanism to determine the query vector, the key vector, and the value vector corresponding to the spectrum vector, and then the attention weight between the query vector and the key vector is determined. Thus, the second frequency-domain feature parameter corresponding to the driving data of each data type is determined based on the attention weight and the value vector. In this way, through the attention mechanism, inter-sequence correlation learning can be efficiently and accurately performed on the frequency-domain feature parameters corresponding to the driving data of multiple data types, so as to obtain the second frequency-domain feature parameters that incorporate the driving data of other data types, and more accurately represent the frequency-domain features of the driving data of multiple data types.

[0010] In a possible implementation manner, the above-mentioned determination of the driving style corresponding to the vehicle based on the second frequency-domain feature parameter corresponding to the driving data of each data type includes: processing the second frequency-domain feature parameter corresponding to the driving data of each data type through causal dilated convolution to obtain a processed second frequency-domain feature parameter; performing sampling processing on the processed second frequency-domain feature parameter to obtain a sampled second frequency-domain feature parameter; performing time-domain transformation on the sampled second frequency-domain feature parameter to obtain a time-domain feature parameter corresponding to the sampled second frequency-domain feature parameter; and determining the driving style corresponding to the vehicle based on the time-domain feature parameter.

[0011] According to the above technical means, the present application can process the second frequency domain characteristic parameters corresponding to the driving data of each data type by means of causal dilated convolution, perform sampling processing on the processed second frequency domain characteristic parameters, and then perform time domain transformation on the sampled second frequency domain characteristic parameters to obtain the time domain characteristic parameters corresponding to the sampled second frequency domain characteristic parameters. Furthermore, based on the time domain characteristic parameters, the driving style corresponding to the vehicle is determined. In this way, by performing causal dilated convolution, sampling and other processing on the frequency domain data (second frequency domain characteristic parameters), efficient and accurate data processing can be performed on the data in the frequency domain. Then, the finally obtained frequency domain data (sampled second frequency domain characteristic parameters) is transformed into the time domain to obtain the corresponding time domain characteristic parameters. Thus, more accurate data analysis can be performed based on the time domain characteristic parameters, and the driving style corresponding to the vehicle can be accurately determined.

[0012] In a possible implementation manner, the above determining the driving style corresponding to the vehicle based on the time domain characteristic parameters includes: performing data segmentation processing on the time domain characteristic parameters based on a preset step length to obtain a plurality of data segments; determining the characteristic data corresponding to each data segment in the plurality of data segments, where the characteristic data includes at least one of the following: average value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter, and the driving fluctuation parameter is used to indicate the degree of deviation of the data from the reference value; determining the driving style corresponding to the vehicle based on the characteristic data corresponding to each data segment in the plurality of data segments.

[0013] According to the above technical means, the present application can perform data segmentation processing on the determined time domain characteristic parameters based on a preset step length to obtain a plurality of data segments, and then determine the characteristic data (average value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter) corresponding to each data segment in the plurality of data segments. Thus, by analyzing the characteristic data corresponding to each data segment, the driving style corresponding to the vehicle can be determined. In this way, by determining the characteristic data corresponding to each data segment, the personal preference of the driver when driving the vehicle can be accurately characterized by the characteristic parameters, and thus the driving style corresponding to the vehicle can be accurately determined.

[0014] In a possible implementation manner, the above determining the driving style corresponding to the vehicle based on the characteristic data corresponding to each data segment in the plurality of data segments includes: determining the driving style corresponding to each data segment by means of the K-means clustering algorithm based on the characteristic data corresponding to each data segment in the plurality of data segments; determining the driving style corresponding to the vehicle based on the driving style corresponding to each data segment.

[0015] According to the above technical means, the present application can perform clustering processing on the feature data corresponding to each data segment through the K-means clustering algorithm, so as to analyze and determine the driving style corresponding to each data segment. Then, based on the driving style corresponding to each data segment, an overall analysis is performed on the driving data of multiple data types of the vehicle, so as to accurately determine the overall driving style corresponding to the vehicle.

[0016] In a possible implementation manner, the driving styles include: a conservative driving style, a normal driving style, and an aggressive driving style; the determining the driving style corresponding to the vehicle based on the driving style corresponding to each data segment includes: determining the number of data segments corresponding to each driving style; determining a driving style evaluation parameter based on the number of data segments corresponding to each driving style and the weight parameter of each driving style; and determining the driving style corresponding to the vehicle based on the driving style evaluation parameter.

[0017] According to the above technical means, the present application can determine the number of data segments corresponding to each driving style, and then determine the driving style evaluation parameter of the vehicle based on the number of data segments corresponding to each driving style and the weight parameter of each driving style. Thus, based on the driving style evaluation parameter, the driving style corresponding to the vehicle is determined. In this way, based on the driving style corresponding to each data segment, the number of data segments corresponding to each driving style included in the driving data of multiple data types of the vehicle can be determined. Thus, an overall analysis is performed on the driving style of each data segment, and the overall driving style evaluation parameter of the vehicle is obtained, so as to accurately determine the overall driving style corresponding to the vehicle.

[0018] In a possible implementation manner, the determining the driving style corresponding to the vehicle based on the driving style evaluation parameter includes: when the driving style evaluation parameter is greater than or equal to the first preset parameter and less than the second preset parameter, determining that the driving style corresponding to the vehicle is a conservative driving style, and the second preset parameter is greater than the first preset parameter; when the driving style evaluation parameter is greater than or equal to the second preset parameter and less than the third preset parameter, determining that the driving style corresponding to the vehicle is a normal driving style, and the third preset parameter is greater than the second preset parameter; and when the driving style evaluation parameter is greater than or equal to the third preset parameter, determining that the driving style corresponding to the vehicle is an aggressive driving style.

[0019] According to the above technical means, the present application can set multiple parameter intervals and the driving style corresponding to each parameter interval, so as to accurately determine the specific driving style corresponding to the vehicle by judging the parameter interval to which the driving style evaluation parameter belongs.

[0020] In a possible implementation manner, obtaining driving data of multiple data types of a vehicle includes: obtaining basic data of multiple data types collected by sensors of the vehicle, where the multiple data types include: vehicle power data and driving operation data, and the vehicle power data includes at least one of the following: driving speed, longitudinal acceleration, lateral acceleration, motor speed, motor torque, and the driving operation data includes at least one of the following: accelerator pedal angle, brake pedal angle, steering angle; preprocessing the basic data of multiple data types to obtain preprocessed driving data of multiple data types, and the preprocessing includes at least one of the following: abnormal data error correction processing, missing data filling processing, data noise reduction processing.

[0021] According to the above technical means, the present application can preprocess the basic data of multiple data types collected by sensors of the vehicle to obtain preprocessed driving data of multiple data types. In this way, based on the real data collected by the sensors of the vehicle, the driving style of the vehicle can be analyzed and determined more accurately.

[0022] In a possible implementation manner, the basic data of multiple data types includes data collected by the vehicle in multiple driving environments, and the multiple driving environments include: highway, urban road, commercial block, residential road, rural road.

[0023] According to the above technical means, the present application can obtain the basic data of multiple data types corresponding to each driving environment in multiple driving environments by obtaining the data collected by the vehicle in multiple driving environments. Thus, when performing data analysis, more comprehensive data analysis can be performed based on the data in multiple driving environments. Thus, the driving style of the vehicle can be determined more accurately.

[0024] According to the second aspect provided by the present application, a driving style determination device is provided, and the driving style determination device includes: an acquisition module and a processing module; the acquisition module is used to acquire driving data of multiple data types of a vehicle; the processing module is used to determine a first frequency domain feature parameter corresponding to the driving data of each data type in the driving data of multiple data types, and the first frequency domain feature parameter is used to characterize the frequency domain feature; the processing module is further used to determine a second frequency domain feature parameter corresponding to the driving data of each data type by performing sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types, and the second frequency domain feature parameter integrates the frequency domain features of the driving data of other data types; the processing module is further used to determine the driving style corresponding to the vehicle based on the second frequency domain feature parameter corresponding to the driving data of each data type.

[0025] In a possible implementation manner, the first frequency-domain feature parameter includes a spectrum parameter; the processing module is specifically configured to perform frequency-domain transformation on the driving data of each data type among the driving data of multiple data types based on a preset frequency to obtain the spectrum parameter corresponding to the driving data of each data type.

[0026] In a possible implementation manner, the processing module is specifically configured to construct a spectrum vector corresponding to the driving data of multiple data types based on the first frequency-domain feature parameter corresponding to the driving data of each data type; the processing module is specifically configured to determine a query vector, a key vector, and a value vector corresponding to the spectrum vector; the processing module is specifically configured to determine the attention weight between the query vector and the key vector; the processing module is specifically configured to determine the second frequency-domain feature parameter corresponding to the driving data of each data type based on the attention weight and the value vector.

[0027] In a possible implementation manner, the processing module is specifically configured to process the second frequency-domain feature parameter corresponding to the driving data of each data type through causal dilated convolution to obtain the processed second frequency-domain feature parameter; the processing module is specifically configured to perform sampling processing on the processed second frequency-domain feature parameter to obtain the sampled second frequency-domain feature parameter; the processing module is specifically configured to perform time-domain transformation on the sampled second frequency-domain feature parameter to obtain the time-domain feature parameter corresponding to the sampled second frequency-domain feature parameter; the processing module is specifically configured to determine the driving style corresponding to the vehicle based on the time-domain feature parameter.

[0028] In a possible implementation manner, the processing module is specifically configured to perform data segmentation processing on the time-domain feature parameter based on a preset step size to obtain multiple data segments; the processing module is specifically configured to determine the feature data corresponding to each data segment among the multiple data segments, and the feature data includes at least one of the following: average value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter, and the driving fluctuation parameter is used to indicate the degree of deviation of the data from the reference value; the processing module is specifically configured to determine the driving style corresponding to the vehicle based on the feature data corresponding to each data segment among the multiple data segments.

[0029] In a possible implementation manner, the processing module is specifically configured to determine the driving style corresponding to each data segment through the K-means clustering algorithm based on the feature data corresponding to each data segment among the multiple data segments; the processing module is specifically configured to determine the driving style corresponding to the vehicle based on the driving style corresponding to each data segment.

[0030] In a possible implementation manner, driving styles include: a conservative driving style, a normal driving style, and an aggressive driving style; a processing module, specifically configured to determine the number of data segments corresponding to each driving style; a processing module, specifically configured to determine a driving style evaluation parameter based on the number of data segments corresponding to each driving style and the weight parameter of each driving style; a processing module, specifically configured to determine the driving style corresponding to the vehicle based on the driving style evaluation parameter.

[0031] In a possible implementation manner, the processing module is specifically configured to determine that the driving style corresponding to the vehicle is a conservative driving style when the driving style evaluation parameter is greater than or equal to a first preset parameter and less than a second preset parameter, and the second preset parameter is greater than the first preset parameter; the processing module is specifically configured to determine that the driving style corresponding to the vehicle is a normal driving style when the driving style evaluation parameter is greater than or equal to the second preset parameter and less than a third preset parameter, and the third preset parameter is greater than the second preset parameter; the processing module is specifically configured to determine that the driving style corresponding to the vehicle is an aggressive driving style when the driving style evaluation parameter is greater than or equal to the third preset parameter.

[0032] In a possible implementation manner, an acquisition module is specifically configured to acquire basic data of multiple data types collected by sensors of the vehicle, and the multiple data types include: vehicle power data and driving operation data. The vehicle power data includes at least one of the following: driving speed, longitudinal acceleration, lateral acceleration, motor speed, motor torque. The driving operation data includes at least one of the following: accelerator pedal angle, brake pedal angle, steering angle; the processing module is further configured to preprocess the basic data of the multiple data types to obtain preprocessed driving data of the multiple data types, and the preprocessing includes at least one of the following: abnormal data error correction processing, missing data filling processing, data noise reduction processing.

[0033] In a possible implementation manner, the basic data of the multiple data types includes data collected by the vehicle in multiple driving environments, and the multiple driving environments include: highways, urban roads, commercial blocks, residential roads, rural roads.

[0034] According to a third aspect provided by the present application, there is provided an electronic device, including: a processor and a memory; wherein, the memory is used to store one or more programs, and the one or more programs include computer execution instructions. When the electronic device runs, the processor executes the computer execution instructions stored in the memory, and the electronic device executes the method according to the first aspect and any possible implementation manner thereof.

[0035] According to the fourth aspect provided by the present application, there is provided a computer-readable storage medium. When the computer instructions stored in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device executes the method according to the first aspect and any possible implementation manner thereof.

[0036] According to the fifth aspect provided by the present application, there is provided a computer program product. The computer program product includes computer instructions. When the computer instructions run on an electronic device, the electronic device executes the method according to the first aspect and any possible implementation manner thereof.

[0037] According to the sixth aspect provided by the present application, there is provided a vehicle. The vehicle includes a driving style determination device as described in the second aspect, and the vehicle is used to implement the method according to the first aspect and any possible implementation manner thereof.

[0038] According to the seventh aspect provided by the present application, there is provided a driving style determination system. The driving style determination system includes: a sensor and a controller; the sensor is used to obtain driving data of multiple data types of the vehicle; the controller is used to determine a first frequency domain feature parameter corresponding to the driving data of each data type in the driving data of multiple data types, and the first frequency domain feature parameter is used to characterize the frequency domain feature; the controller is further used to determine a second frequency domain feature parameter corresponding to the driving data of each data type by performing sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types, and the second frequency domain feature parameter integrates the frequency domain features of the driving data of other data types; the controller is further used to determine the driving style corresponding to the vehicle based on the second frequency domain feature parameter corresponding to the driving data of each data type.

[0039] It should be noted that for the technical effects brought by any implementation manner in the second aspect to the seventh aspect, reference may be made to the technical effects brought by the corresponding implementation manner in the first aspect, which will not be elaborated herein.

[0040] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. Description of the Drawings

[0041] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application, and do not constitute an improper limitation to the present application.

[0042] Figure 1 is a schematic structural diagram of a driving style determination system shown according to an exemplary embodiment; Figure 2 is a schematic structural diagram of another driving style determination system shown according to an exemplary embodiment; Figure 3 It is a flowchart of a driving style determination method shown according to an exemplary embodiment; Figure 4 It is a flowchart of another driving style determination method shown according to an exemplary embodiment; Figure 5 It is a flowchart of another driving style determination method shown according to an exemplary embodiment; Figure 6 It is a flowchart of another driving style determination method shown according to an exemplary embodiment; Figure 7 It is a flowchart of another driving style determination method shown according to an exemplary embodiment; Figure 8 It is a flowchart of another driving style determination method shown according to an exemplary embodiment; Figure 9 It is a block diagram of a driving style determination device shown according to an exemplary embodiment; Figure 10 It is a block diagram of an electronic device shown according to an exemplary embodiment; Figure 11 It is a schematic structural diagram of a computer system shown according to an exemplary embodiment. Detailed implementation manners

[0043] In order to enable those of ordinary skill in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0044] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order different from those illustrated or described here. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application. On the contrary, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0045] The driving style determination method provided by the embodiments of the present application can be applied to a driving style determination system. Figure 1 It is a schematic structural diagram of a driving style determination system shown according to an exemplary embodiment. As Figure 1 shown, the driving style determination system includes: sensor 11 and controller 12.

[0046] The sensor 11 is used to collect vehicle power data and driving operation data. The vehicle power data includes at least one of the following: driving speed, longitudinal acceleration, lateral acceleration, motor speed, motor torque. The driving operation data includes at least one of the following: accelerator pedal angle, brake pedal angle, steering angle.

[0047] In a possible implementation, the sensor 11 can be a speed sensor, an acceleration sensor, a rotation speed sensor, an angle sensor, etc.

[0048] In a possible implementation, the sensor 11 can collect basic data of various data types of the vehicle, and process the collected basic data to obtain processed driving data of various data types.

[0049] The controller 12 can be a vehicle controller or a driving style controller, etc. The driving style controller is used to control and adjust the driving style of the vehicle.

[0050] In a possible implementation, the controller 12 can process the basic data collected by the sensor 11 to obtain processed driving data of various data types.

[0051] In a possible implementation, the controller 12 can determine the first frequency domain characteristic parameter corresponding to the driving data of each data type in the driving data of various data types. The first frequency domain characteristic parameter is used to characterize the frequency domain characteristic.

[0052] In a possible implementation, the controller 12 can determine the second frequency domain characteristic parameter corresponding to the driving data of each data type by performing sequence - to - sequence correlation learning on the frequency domain characteristic parameters corresponding to the driving data of various data types. The second frequency domain characteristic parameter integrates the frequency domain characteristics of the driving data of other data types.

[0053] In a possible implementation, the controller 12 can determine the driving style corresponding to the vehicle based on the second frequency domain characteristic parameter corresponding to the driving data of each data type.

[0054] In a possible implementation, the controller 12 can adjust the power parameters or driving operation parameters of the vehicle based on the determined driving style corresponding to the vehicle to adjust the driving style of the vehicle.

[0055] In a possible implementation, Figure 2 is a schematic structural diagram of another driving style determination system shown according to an exemplary embodiment. As Figure 2 shown, the driving style determination system includes: a data acquisition module 21, a data pre - processing module 22, a data segment division module 23, a feature modeling module 24, and an evaluation parameter calculation module 25.

[0056] The data acquisition module 21 is used to obtain driving data of various data types of the vehicle. Specifically, it is used to obtain the basic data of various data types collected by the sensors of the vehicle.

[0057] Specifically, the data acquisition module 21 is used to collect the power data and driving operation data of the vehicle when drivers with different driving styles drive the vehicle in various driving scenarios.

[0058] The data preprocessing module 22 is used to preprocess the basic data of various data types to obtain the preprocessed driving data of various data types. The preprocessing includes at least one of the following: abnormal data error correction processing, missing data filling processing, and data noise reduction processing.

[0059] The data segment division module 23 is used to perform data segmentation processing on the data based on a preset step size to obtain multiple data segments. Specifically, the data (i.e., time domain feature parameters) can be segmented according to a preset sliding window (time interval, step size), and the statistical features and time series features of the data included in each data segment are determined.

[0060] The feature modeling module 24 is used to determine the first frequency domain feature parameters corresponding to the driving data of each data type in the driving data of various data types, and determine the second frequency domain feature parameters corresponding to the driving data of each data type by performing inter-sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of various data types, and determine the driving style corresponding to the vehicle based on the second frequency domain feature parameters corresponding to the driving data of each data type.

[0061] Specifically, first, perform frequency domain transformation on the driving data of various data types, and then use a variable-oriented frequency-aware attention mechanism to interact with the frequency domain time series variables of the driving data of each data type, and capture the dependence relationship between the frequency components in the driving data of each data type through frequency component causal dilation convolution.

[0062] The evaluation parameter calculation module 25 is used to determine the number of data segments corresponding to each driving style, and determine the driving style evaluation parameter based on the number of data segments corresponding to each driving style and the weight parameter of each driving style, and determine the driving style corresponding to the vehicle based on the driving style evaluation parameter.

[0063] For the sake of easy understanding, the following specifically introduces the driving style determination method provided by this application in combination with the accompanying drawings.

[0064] Figure 3 is a flowchart of a driving style determination method shown according to an exemplary embodiment, which is applied to an electronic device (controller), such as Figure 3 shown, and this driving style determination method includes the following S301 - S304: S301. Obtain driving data of multiple data types of the vehicle.

[0065] In the embodiments of the present application, by obtaining various relevant data involved in the vehicle driving process (such as the vehicle power data and driving operation data described below), and performing data analysis, the driving style of the driver when driving the vehicle can be determined based on the characteristics of the data.

[0066] In some embodiments, in a driving style determination method provided by the embodiments of the present application, the above step S301 may specifically include S3011 - S3012: S3011. Obtain the basic data of multiple data types collected by the sensors of the vehicle.

[0067] Among them, the multiple data types include: vehicle power data and driving operation data. The vehicle power data includes at least one of the following: driving speed, longitudinal acceleration, lateral acceleration, motor speed, motor torque. These data reflect the physical motion state of the vehicle and the performance of the power system. The driving operation data includes at least one of the following: accelerator pedal angle, brake pedal angle, steering angle. These data are directly related to the driver's operation behavior and driving style.

[0068] In a possible implementation manner, the sensors may be speed sensors, acceleration sensors, speed sensors, angle sensors, etc. That is, the sensors are used to collect the driver's driving operation data (i.e., accelerator pedal angle, brake pedal angle, steering angle) during the vehicle driving process or the response data (i.e., driving speed, longitudinal acceleration, lateral acceleration, motor speed, motor torque, etc.) generated after the vehicle responds to the driver's driving operation.

[0069] It can be understood that the basic data is the time-series data directly collected by the sensors. By collecting and analyzing the basic data of multiple data types, a multi-dimensional driving behavior model can be constructed, which can not only identify and classify driving styles, but also deeply understand the change rules of driving behaviors in different scenarios.

[0070] Table 1

[0071] Exemplarily, as shown in Table 1, while obtaining data fields such as driving speed (ESP_VehicleSpeed), lateral acceleration (ESP_LatAccel), longitudinal acceleration (ESP_LongAccel), steering angle (SAS_SteeringAngle), accelerator pedal position (PCU_AccPedl), and brake pedal angle (PCU_BrkPedlSts), data fields such as the unique identifier (id) of the vehicle and the data acquisition time (terminal_time) can also be obtained. Moreover, while obtaining the corresponding data fields, the unit of each data field can also be determined.

[0072] In some embodiments, the basic data of multiple data types includes data collected by the vehicle in multiple driving environments, and the multiple driving environments include: highways, urban roads, commercial blocks, residential roads, and rural roads.

[0073] In one possible implementation, the basic data of multiple data types corresponding to each of the multiple driving environments can be collected for data analysis to ensure the comprehensiveness and representativeness of the driving styles obtained after data analysis.

[0074] It should be noted that in these different driving environments, the driving styles demonstrated by drivers may vary significantly. Therefore, in order to comprehensively determine the driving style, the basic data of multiple data types corresponding to each of the multiple driving environments can be collected for data analysis, and moreover, by combining the data of multiple data types, the driving style can be determined more accurately.

[0075] It should be noted that the following embodiments take the data of multiple data types corresponding to multiple driving environments as examples for analysis. It can be understood that the processing and analysis processes of the data of multiple data types corresponding to each of the multiple driving environments are the same, and the data of multiple data types corresponding to each of the multiple driving environments can also be analyzed separately to determine the driving style of the vehicle in that driving environment.

[0076] In the embodiments of the present application, the present application can obtain the data collected by the vehicle in multiple driving environments, and thus obtain the basic data of multiple data types corresponding to each of the multiple driving environments. Therefore, when performing data analysis, more comprehensive data analysis can be carried out based on the data under multiple driving environments. Thus, the driving style of the vehicle can be determined more accurately.

[0077] S3012. Preprocess the basic data of multiple data types to obtain the preprocessed driving data of multiple data types.

[0078] Among them, the preprocessing includes at least one of the following: abnormal data error correction processing, missing data filling processing, and data noise reduction processing.

[0079] In a possible implementation manner, data filtering processing can also be performed on the basic data of multiple data types to screen out signal fields (such as driving speed, longitudinal acceleration, and lateral acceleration, etc.) that are closely related to identifying driving styles from the basic data of multiple data types. Among them, the driving speed can indicate the speed of the vehicle, indicating the driver's requirement for speed; the longitudinal acceleration can indicate the driver's requirement for accelerating the vehicle when driving; the lateral acceleration can indicate whether the driver will make sharp turns, quickly change lanes, etc. when driving.

[0080] In a possible implementation manner, the abnormal data error correction processing can correct data deviation by means of deletion or filling, delete the missing data or fill it by interpolation; the missing data filling processing can delete the missing data or fill it by interpolation; the data noise reduction processing can remove or weaken the noise and interference in the time series through certain algorithms, such as weighted average or filtering technology.

[0081] In a possible implementation manner, the missing data may include: possible acquisition errors during data acquisition or possible transmission anomalies during cloud data transmission. When it is determined that the data has continuous missing values, this part of the data can be deleted; if there are a small number of missing data, the forward data filling method and the average value filling method can be used to fill the missing data.

[0082] It should be noted that during the process of missing data filling processing, the missing data can be understood as: within a certain period of time, the data of a certain data type among multiple data types (such as driving speed) is not collected, and only the data of other partial data types is collected, then the data of this certain data type (such as driving speed) is the missing data. Therefore, when performing missing data filling processing, the missing data of a certain data type (such as driving speed) can be filled. Or, the missing data can also be deleted, that is, the other data collected during this period of time (that is, the data of other partial data types collected) can be deleted. This is because the data of this certain data type (such as driving speed) makes the other data collected during this period of time (that is, the data of other partial data types collected) invalid, so it can be deleted.

[0083] In a possible implementation manner, the abnormal data can be abnormal data caused by sensor acquisition errors or information transmission errors. For such data, the standard deviation method can be used to determine whether it is abnormal data. And the abnormal data can be removed by combining the box plot method.

[0084] It should be noted that the box plot constructs an intuitive graphical representation using the quartiles of the data, and data above or below the data distribution boundary is regarded as outliers. It is possible to select data outside the 95th percentile and 5th percentile of the data as outliers for deletion by plotting a data visualization graph.

[0085] In a possible implementation, the abnormal data can also be multiple consecutive unchanged data. Since the probability of multiple consecutive data values remaining unchanged is small, it can be regarded as abnormal data. For such abnormal data, a downsampling method can be adopted to reduce the error caused by data redundancy.

[0086] In a possible implementation, the abnormal data can also be frequently randomly fluctuating data (i.e., noise). Since time series data often corresponds to actual data in reality, when the data is affected by frequent random fluctuations or obvious periodic changes, the Kalman filtering method can be used to smooth the data to remove or reduce the interference of these noises.

[0087] In the embodiments of the present application, the present application can preprocess the basic data of various data types collected by the sensors of the vehicle to obtain the driving data of various data types after preprocessing. In this way, based on the real data collected by the sensors of the vehicle, the driving style of the vehicle can be analyzed and determined more accurately.

[0088] S302. Determine the first frequency domain characteristic parameter corresponding to the driving data of each data type in the driving data of multiple data types.

[0089] Among them, the first frequency domain characteristic parameter is used to characterize the frequency domain characteristics.

[0090] In some embodiments, the first frequency domain characteristic parameter includes spectral parameters. In a driving style determination method provided by an embodiment of the present application, the above step S302 may specifically include: for the driving data of each data type in the driving data of multiple data types, perform a frequency domain transformation on the driving data of each data type based on a preset frequency to obtain the spectral parameter corresponding to the driving data of each data type.

[0091] In a possible implementation, based on a deep learning data feature modeling system, feature analysis and modeling can be performed on the driving data of each data type in the driving data of multiple data types by enhancing the correlation between Fourier sequences and the dependence of frequency components. The deep learning data feature modeling system can be composed of parts such as a Fourier transform structure, a frequency-aware attention mechanism, frequency component causal dilation convolution and sampling, and time domain mapping.

[0092] In a possible implementation, the driving data of each data type collected can be Fourier-transformed through Equation (1).

[0093]

[0094] Where represents the sampled data at time t in the driving data of the i-th data type, where , N is the total number of data types, represents the spectral parameter (spectral value) of the driving data of the i-th data type at the preset frequency f, and j is the imaginary unit.

[0095] It can be understood that through Fourier transform, the time-domain data signal can be converted into a frequency-domain data signal . Using frequency-domain data can represent the compactness and overall perspective of the data, and can more effectively encode time data and capture the inherent time patterns in the data.

[0096] In the embodiments of the present application, the present application can perform frequency-domain transformation on the driving data of each data type among the driving data of multiple data types based on a preset frequency, so as to obtain the spectral parameters corresponding to the driving data of each data type. In this way, through frequency-domain transformation, the obtained time-domain driving data can be transformed into the frequency domain, so that based on the frequency-domain data (i.e., spectral parameters), the data characteristics can be determined more accurately, and the data can be processed more accurately and efficiently.

[0097] S303. By performing inter-sequence correlation learning on the frequency-domain characteristic parameters corresponding to the driving data of multiple data types, determine the second frequency-domain characteristic parameter corresponding to the driving data of each data type.

[0098] Wherein, the second frequency-domain characteristic parameter incorporates the frequency-domain characteristics of the driving data of other data types.

[0099] S304. Based on the second frequency-domain characteristic parameter corresponding to the driving data of each data type, determine the driving style corresponding to the vehicle.

[0100] In a possible implementation, a frequency-aware attention mechanism can be adopted to simulate the dynamic correlation between various variables (the driving data of each data type), and then a frequency-component causal dilation convolution is used to capture the dependence relationship between frequency components (frequency-domain characteristic parameters).

[0101] In a possible implementation, after performing a frequency-domain transformation on the driving data of each data type based on a preset frequency to obtain the spectral parameters corresponding to the driving data of each data type, a variable-oriented frequency-aware attention mechanism can be used to interact with the spectral parameters corresponding to the driving data of each data type. At the same time, frequency-component causal dilation convolution is used to capture the dependencies between frequency components in each frequency-domain variable, so as to capture important similarity associations between adjacent frequencies or harmonically related frequencies.

[0102] In a possible implementation, data processing can be performed on the second frequency-domain feature parameters corresponding to the driving data of each data type through data processing methods such as causal dilation convolution and frequency-domain sampling, and then the processed data is subjected to a time-domain transformation to obtain the corresponding time-domain data (i.e., time-domain feature parameters). Furthermore, the time-domain data is continuously clustered through the K-means algorithm to determine the driving style corresponding to the vehicle.

[0103] In the embodiments of the present application, the present application can determine the first frequency-domain feature parameters corresponding to the driving data of each data type among the driving data of multiple data types of the vehicle to characterize the frequency-domain features of the driving data of each data type. Then, inter-sequence correlation learning is performed on the frequency-domain feature parameters corresponding to the driving data of multiple data types, so that the frequency-domain features of the driving data of other data types can be fused in the frequency-domain feature parameters corresponding to the driving data of each data type, and the second frequency-domain feature parameters corresponding to the driving data of each data type are obtained. Based on this, the correlation between the driving data of each data type and the driving data of other data types can be characterized by the second frequency-domain feature parameters corresponding to the driving data of each data type. Furthermore, based on the second frequency-domain feature parameters corresponding to the driving data of each data type, the driving style corresponding to the vehicle can be determined. That is, by fusing the frequency-domain features of the driving data of multiple data types, and based on the frequency-domain feature parameters obtained after fusion, the characteristics of the driving data of the vehicle can be reflected from the overall data, so that the driving style of the entire vehicle can be determined more accurately and efficiently.

[0104] In some embodiments, as Figure 4 shown, in a driving style determination method provided by an embodiment of the present application, the above step S303 may specifically include S401-S404.

[0105] S401. Based on the first frequency-domain feature parameters corresponding to the driving data of each data type, construct spectral vectors corresponding to the driving data of multiple data types.

[0106] S402. Determine the query vector, key vector, and value vector corresponding to the spectral vector.

[0107] S403. Determine the attention weights between the query vector and the key vector.

[0108] S404. Based on the attention weights and the value vector, determine the second frequency domain characteristic parameters corresponding to the driving data of each data type.

[0109] It should be noted that in order to capture the dynamic correlation between the driving data of each data type among the driving data of multiple data types, based on the variable-oriented frequency-aware attention mechanism, the correlation between the driving data of multiple data types can be dynamically learned according to the spectral parameters of the driving data of different data types at a specific frequency.

[0110] In a possible implementation, for each preset frequency f, the spectral parameters corresponding to the driving data of each data type among the driving data of multiple data types form a spectral vector . Then, through three learnable matrices: the query matrix , the key matrix , and the value matrix , the spectral vector is linearly transformed into three different spaces to obtain the query vector , the key vector , and the value vector .

[0111] It should be noted that , where and are the dimensions of the key vector and the value vector respectively.

[0112] Furthermore, calculate the attention weight between the query vector and the key vector based on Formula Two.

[0113]

[0114] Among them, is used to calculate the similarity between the query vector and the key vector , divided by is to prevent the inner product from being too large, and the function converts the similarity into a probability distribution so that the sum of the attention weights is 1. , where represents the attention weight between the driving data of the i-th data type and the driving data of the j-th data type at the preset frequency f.

[0115] Finally, the attention weights Multiply with the value vector to obtain the second frequency domain feature parameter corresponding to the driving data of each data type finally . Among them, , represents the second frequency domain feature parameter corresponding to the driving data of the i-th data type at the preset frequency f. It integrates the features of the driving data of other data types, and the obtained second frequency domain feature parameter is dynamically learned by the attention mechanism

[0116] It can be understood that the variable-oriented frequency-aware attention mechanism interacts with the second frequency domain feature parameters corresponding to the driving data of each data type in the driving data of multiple data types. By simulating the dynamic correlation between the driving data through the frequency-aware attention mechanism, the analysis effect of interconnection between the driving data of each data type is achieved, which is beneficial to the reverse inference of the driving operation habits of different driving styles

[0117] In the embodiments of the present application, the present application can construct a spectrum vector corresponding to the driving data of multiple data types based on the first frequency domain feature parameter corresponding to the driving data of each data type, so as to represent the overall features corresponding to the driving data of multiple data types through the spectrum vector. Furthermore, sequence correlation learning is performed through the attention mechanism to determine the query vector, key vector, and value vector corresponding to the spectrum vector, and then the attention weight between the query vector and the key vector is determined. Thus, based on the attention weight and the value vector, the second frequency domain feature parameter corresponding to the driving data of each data type is determined. In this way, through the attention mechanism, sequence correlation learning can be efficiently and accurately performed on the frequency domain feature parameters corresponding to the driving data of multiple data types, so as to obtain the second frequency domain feature parameter that integrates the driving data of other data types, and more accurately represent the frequency domain features of the driving data of multiple data types

[0118] In some embodiments, as Figure 5 shown, in a driving style determination method provided by the embodiments of the present application, the above step S304 may specifically include S501-S504

[0119] S501. Process the second frequency domain feature parameter corresponding to the driving data of each data type through causal dilation convolution to obtain the processed second frequency domain feature parameter

[0120] In a possible implementation manner, in order to capture the dependence relationship between the second frequency domain feature parameters corresponding to the driving data of each data type, the causal dilation convolution of frequency components can be sampled to expand the receptive field of the convolution kernel without increasing the number of parameters, so as to capture a larger range of frequency dependence relationships

[0121] Exemplarily, the second frequency domain feature parameters corresponding to the driving data of each data type can be processed through Equation 3 to obtain the processed second frequency domain feature parameters .

[0122]

[0123] where K is the convolution kernel size, is the convolution kernel weight, and d is the dilation rate.

[0124] It should be noted that causality is reflected in the fact that the convolution operation only considers the current preset frequency f and the frequency components before it, and does not consider the frequency components after it. The dilation rate d controls the interval of the convolution kernel sampling points. For example, when d = 1, it is an ordinary convolution; when d = 2, the convolution kernel samples every other frequency point, thus expanding the receptive field. By using different dilation rates, frequency dependence relationships at different scales can be captured. The driving data of a specific driving style are often similar in the frequency component spectrum representation. For example, similar acceleration or deceleration habits usually have similar amplitudes in the frequency domain components and also have similar related harmonic frequencies. This strategy can effectively capture the important similarity associations and dependencies between similar frequencies or harmonically related frequencies, thereby helping to distinguish similar driving styles.

[0125] S502. Perform sampling processing on the processed second frequency domain feature parameters to obtain the sampled second frequency domain feature parameters.

[0126] In a possible implementation, to improve the calculation efficiency and focus on the second frequency domain feature parameters corresponding to the driving data of important data types, the processed second frequency domain feature parameters can be sampled through a frequency domain sampling mapping mechanism to obtain a set of discrete sampling frequency points . Then uniform sampling or non-uniform sampling can be adopted, such as sampling based on Mel frequency. The processed second frequency domain feature parameters are mapped to the sampling frequency points through Equation 4 to obtain the sampled second frequency domain feature parameters .

[0127]

[0128] where Interpolate represents an interpolation function, such as linear interpolation or spline interpolation.

[0129] S503. Perform time domain transformation on the sampled second frequency domain feature parameters to obtain the time domain feature parameters corresponding to the sampled second frequency domain feature parameters.

[0130] In a possible implementation, in order to convert the sampled second frequency-domain feature parameters back to the time domain, the inverse Fourier transform can be used to perform the inverse Fourier transform on the sampled second frequency-domain feature parameters to obtain the corresponding enhanced time-domain feature parameters .

[0131]

[0132] where is the frequency sampling interval.

[0133] S504. Determine the driving style corresponding to the vehicle based on the time-domain feature parameters.

[0134] In this way, by analyzing the frequency-domain data (i.e., the second frequency-domain feature parameters), different driving styles can be effectively distinguished. Finally, the enhanced high-order deep learning driving data features (i.e., the sampled second frequency-domain feature parameters ) are input into the clustering method for driving style recognition.

[0135] In the embodiments of the present application, the present application can process the second frequency-domain feature parameters corresponding to the driving data of each data type by means of causal dilated convolution, and perform sampling processing on the processed second frequency-domain feature parameters, and then perform time-domain transformation on the sampled second frequency-domain feature parameters to obtain the time-domain feature parameters corresponding to the sampled second frequency-domain feature parameters. Furthermore, based on the time-domain feature parameters, the driving style corresponding to the vehicle is determined. In this way, through causal dilated convolution, sampling and other processing on the frequency-domain data (the second frequency-domain feature parameters), efficient and accurate data processing can be performed on the data in the frequency domain. Then, the finally obtained frequency-domain data (the sampled second frequency-domain feature parameters) is transformed to the time domain to obtain the corresponding time-domain feature parameters. Thus, more accurate data analysis can be performed based on the time-domain feature parameters, and the driving style corresponding to the vehicle can be accurately determined.

[0136] In some embodiments, as Figure 6 shown, in a driving style determination method provided by the embodiments of the present application, the above step S504 may specifically include S601-S603.

[0137] S601. Perform data segmentation processing on the time-domain feature parameters based on a preset step length to obtain a plurality of data segments.

[0138] S602. Determine the feature data corresponding to each data segment among the plurality of data segments.

[0139] Among them, the characteristic data includes at least one of the following: mean value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter, where the driving fluctuation parameter is used to indicate the degree to which the data deviates from the reference value.

[0140] In a possible implementation, the time-domain characteristic parameters can be segmented according to a preset sliding window (i.e., a preset step size), and the statistical characteristics, frequency-domain characteristics, and time-series characteristics of the data included in each data segment are calculated.

[0141] Exemplarily, the time-domain characteristic parameters obtained under different scenarios can be segmented according to a time window with a time length of t seconds and a sliding step size of Δt seconds. Then, the statistical characteristics such as the mean value, maximum value, and standard deviation of the data included in each data segment (i.e., the time-domain characteristic parameters), as well as the high-order characteristics such as kurtosis and skewness, and the time-series characteristics such as time-varying driving fluctuations (i.e., the driving fluctuation parameter) are calculated.

[0142] It should be noted that data such as the mean value, variance, maximum value, minimum value, median, and standard deviation are statistical characteristics, which can reflect the central tendency, degree of variation, and extreme situations of driving behavior. Kurtosis and skewness are high-order characteristics; the driving fluctuation parameter is a time-series characteristic.

[0143] Optionally, the multiple data segments obtained after data segmentation can be stored as a matrix, and the present application does not limit the specific values of the time length t seconds and the sliding step size Δt seconds. However, as a possible implementation, the time length can be set to t = 60 seconds and Δt = 10 seconds to calculate the mean value, maximum value, standard deviation, kurtosis, skewness, and driving fluctuation parameter of each time-domain characteristic parameter included in each data segment.

[0144] Exemplarily, the standard deviation S can be calculated by Formula Six.

[0145]

[0146] Among them, is the i-th data in a data segment (i.e., the time-domain characteristic parameter), is the mean value of all the data included in the data segment, and t is the number of all the data included in the data segment.

[0147] Exemplarily, the kurtosis K can be calculated by Formula Seven.

[0148]

[0149] Among them, is the mean value (i.e., ), is the standard deviation (i.e., S), and t is the number of all the data included in the data segment.

[0150] In a possible implementation, the driving fluctuation parameter captures the changes in instantaneous driving behavior by creating a time-series data stream at the driver level, and uses the driving fluctuation parameter as an alternative measure of the aggressiveness of the driving style. The driving fluctuation shows the degree to which the driving data deviates from the norm. A higher driving fluctuation means that the driver has larger fluctuations in speed, acceleration, and deceleration in the lateral and longitudinal directions, indicating that the driver has a higher degree of instability.

[0151] Exemplarily, the driving fluctuation parameter can be calculated by Equation (8).

[0152]

[0153] Wherein, is the i-th data in a data segment (i.e., the time-domain feature parameter), is the data before the i-th data in a data segment, is the parameter average value, and t is the number of all data included in the data segment.

[0154] S603. Determine the driving style corresponding to the vehicle based on the feature data corresponding to each data segment in multiple data segments.

[0155] In the embodiments of the present application, the present application can perform data segmentation processing on the determined time-domain feature parameters based on a preset step size to obtain multiple data segments, and then determine the feature data (average value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter) corresponding to each data segment in the multiple data segments. Thus, by analyzing the feature data corresponding to each data segment, the driving style corresponding to the vehicle can be determined. In this way, by determining the feature data corresponding to each data segment, the personal preference of the driver when driving the vehicle can be accurately characterized by the feature parameters, and thus the driving style corresponding to the vehicle can be accurately determined.

[0156] In some embodiments, as Figure 7 shown, in a driving style determination method provided by an embodiment of the present application, the above step S603 may specifically include S701-S702.

[0157] S701. Based on the feature data corresponding to each data segment in multiple data segments, determine the driving style corresponding to each data segment through the K-means clustering algorithm.

[0158] In a possible implementation, a standard K-means algorithm can be used to initialize the centroids of the embedded data points. The K-means algorithm divides the data points into k clusters and calculates a centroid for each cluster to obtain an initial centroid set composed of k centroids , learn the effective representation of data through layer-by-layer pre-training and fine-tuning. To more accurately assign samples (i.e., the feature data corresponding to each data segment in multiple data segments) to the corresponding clusters, it is necessary to first calculate the distance from the sample to the cluster center, and use the Student's distribution to calculate the probability of the sample to the clustering center (i.e., the cluster center). The greater the distance from the sample to the clustering center, the greater the probability that the sample belongs to the cluster.

[0159] Exemplarily, the probability of the sample to each clustering center can be calculated by Equation (9). .

[0160]

[0161] Among them, represents the sample data (the feature data corresponding to each data segment), represents the data probability, represents the calculation of similarity.

[0162] In a possible implementation, a clustering method based on the Kullback-Leibler Divergence (KL divergence) can be used to cluster the feature data corresponding to each data segment in multiple data segments to obtain multiple driving styles.

[0163] In a possible implementation, a deep learning clustering method based on KL divergence can also be used to cluster multiple data segments into multiple driving styles. The KL divergence is used to measure the difference between the probability distribution of a data point (the feature data corresponding to each data segment) belonging to a certain cluster and the central probability distribution of the cluster, and cluster the feature representations containing driving styles into multiple driving styles.

[0164] Exemplarily, the calculation of the KL divergence from distribution P to distribution Q is as shown in Equation (10).

[0165]

[0166] Among them, P represents the probability distribution of the data point belonging to the cluster, and Q represents the central probability distribution of the cluster.

[0167] In a possible implementation, the model parameters can be optimized by combining network loss and clustering loss, and gradually learn the latent representation of data that can fully represent driving styles through gradient descent.

[0168] Based on this, the model parameters are optimized by combining the network loss and the clustering loss, enabling it to learn the latent representations of data features that can fully characterize driving styles. Finally, the feature representations of multiple driving styles are clustered into various driving style types (such as conservative, normal, and aggressive).

[0169] It should be noted that in the calculation of the loss function, the present application uses the joint optimization of two loss functions. The first loss function is the mean squared error (MSE) used to train the Fourier-enhanced feature modeling, which can ensure that the sequence after the encoder can well represent the original data and extract the part related to driving style from the data at a deeper level. The second loss function is provided by the KL divergence of the clustering algorithm. Minimizing the clustering loss can enable the feature sequence to be better assigned to k clusters by the clustering model. The clustering method based on KL divergence is a clustering technique based on the difference in probability distributions, which is an asymmetric index used to measure the difference between two probability distributions P and Q. In the clustering method, KL divergence can be used to measure the difference between the probability distribution of a data point belonging to a certain cluster and the central probability distribution of that cluster.

[0170] S702. Determine the driving style corresponding to the vehicle based on the driving style corresponding to each data segment.

[0171] In the embodiments of the present application, the present application can perform clustering processing on the feature data corresponding to each data segment in multiple data segments through the K-means clustering algorithm, so as to analyze and determine the driving style corresponding to each data segment. Then, based on the driving style corresponding to each data segment, the driving data of various data types of the vehicle is analyzed as a whole, so as to accurately determine the overall driving style corresponding to the vehicle.

[0172] In some embodiments, the driving styles include: conservative driving style, normal driving style, and aggressive driving style; as Figure 8 shown, in a driving style determination method provided by an embodiment of the present application, the above step S702 may specifically include S801 - S803.

[0173] S801. Determine the number of data segments corresponding to each driving style.

[0174] S802. Determine the driving style evaluation parameter based on the number of data segments corresponding to each driving style and the weight parameter of each driving style.

[0175] S803. Determine the driving style corresponding to the vehicle based on the driving style evaluation parameter.

[0176] In a possible implementation manner, through a style calculation method, the driving style categories in different scenarios are comprehensively calculated to obtain the driver's driving style score, so as to quantify the driving style.

[0177] It can be understood that by integrating the driving style clustering results obtained in multiple scenarios, the specific meanings represented by each driving style are judged according to the standard deviations of the lateral acceleration and the longitudinal acceleration, and then the scores of multiple scenarios are weighted and summed to obtain the total score, so as to shield the influence of the style in a single scenario on the overall style, thereby effectively improving the accuracy of driving style recognition.

[0178] In a possible implementation manner, the driving style evaluation parameter A can be determined by Formula XI.

[0179]

[0180] Among them, f represents the mapping relationship, i represents different scenarios, j represents different driving styles, represents the weight of different driving styles, represents the number of events of the j-th driving style in the i-th scenario (that is, the number of data segments corresponding to the j-th driving style).

[0181] It can be understood that the present application improves the construction of a weight matrix based on each driving style and its proportion in the total driving time. The calculation of this weight matrix not only considers the relative proportion of each driving style, but also adjusts according to the proportion of each driving scenario in the total driving time, so as to accurately quantify the importance of different driving styles and environments, and reflect this importance in the final comprehensive score. This calculation method provides a comprehensive and objective evaluation of the driver's driving style.

[0182] In the embodiments of the present application, the present application can determine the number of data segments corresponding to each driving style, and then determine the driving style evaluation parameter of the vehicle based on the number of data segments corresponding to each driving style and the weight parameter of each driving style. Thus, based on the driving style evaluation parameter, the driving style corresponding to the vehicle is determined. In this way, based on the driving style corresponding to each data segment, the number of data segments corresponding to each driving style included in the driving data of multiple data types of the vehicle can be determined. Thus, the driving style of each data segment is analyzed as a whole to obtain the driving style evaluation parameter of the vehicle as a whole, so as to accurately determine the overall driving style corresponding to the vehicle.

[0183] In some embodiments, in a driving style determination method provided by the embodiments of the present application, the above step S803 may specifically include S8031-S8033.

[0184] S8031. When the driving style evaluation parameter is greater than or equal to the first preset parameter and less than the second preset parameter, it is determined that the driving style corresponding to the vehicle is a conservative driving style, and the second preset parameter is greater than the first preset parameter.

[0185] S8032. When the driving style evaluation parameter is greater than or equal to the second preset parameter and less than the third preset parameter, it is determined that the driving style corresponding to the vehicle is a normal driving style, and the third preset parameter is greater than the second preset parameter.

[0186] S8033. When the driving style evaluation parameter is greater than or equal to the third preset parameter, it is determined that the driving style corresponding to the vehicle is an aggressive driving style.

[0187] Exemplarily, the first preset parameter can be 0, the second preset parameter can be 35, and the third preset parameter can be 70. Thus, when the driving style evaluation parameter is between 0 and 35, the vehicle is defined as a conservative driving style; when the driving style evaluation parameter is between 35 and 70, the vehicle is classified as a normal driving style; and when the driving style evaluation parameter is above 70 (e.g., between 70 and 100), the vehicle is defined as an aggressive driving style. This classification standard not only helps to evaluate the driving style of the vehicle more precisely, but also provides a scientific basis for providing customized driving suggestions and intervention measures according to individual driving styles.

[0188] In the embodiments of the present application, the present application can set multiple parameter intervals and the driving styles corresponding to each parameter interval, so as to accurately determine the specific driving style corresponding to the vehicle by judging the parameter interval to which the driving style evaluation parameter belongs.

[0189] The present application realizes the construction and clustering of deep features of driving data in different scenarios by constructing a system including data preprocessing, data segmentation and feature extraction, Fourier enhanced feature analysis and modeling, KL divergence deep clustering technology, and style score calculation. Through this method, the robustness of such models to driving data with difficult-to-identify styles and abnormal noises can be improved, different driving scenarios can be adapted, and the generalization of driving style recognition can be improved. And by clustering first and then fusing the driving styles in each scenario, the influence of accidental styles in a single case on the overall style evaluation is shielded, effectively improving the accuracy of driving style recognition.

[0190] By converting natural driving data to the frequency domain perspective, the inherent limitations of traditional time domain methods, such as information discreteness and sparsity, are effectively overcome. Utilizing the compactness and overall perspective provided by the frequency domain representation, time data can be encoded more effectively and the underlying time patterns can be captured. Further, a variable-oriented frequency-aware attention mechanism is adopted to interact with various frequency domain sequence variables in driving data, such as longitudinal acceleration, speed, accelerator pedal position, and brake pedal state. By simulating the dynamic correlations between driving data variables through the frequency-aware attention mechanism, the interoperability analysis of various vehicle sensing data is achieved, thereby inversely inferring the operating habits of different driving styles of drivers. At the same time, causal dilated convolution is used to capture the dependencies between the frequency components of each frequency domain sequence. This strategy can effectively capture the important similarity associations between similar frequencies or harmonically related frequencies. The driving data of a specific driving style often shows similarities in the frequency component spectrum representation. For example, similar acceleration or deceleration habits usually have similar amplitudes in the frequency domain components and also have similar related harmonic frequencies. The analysis of the frequency components in the frequency domain variables can effectively distinguish different driving styles dominated by different drivers. After feature analysis and modeling, the KL divergence is used as a deep learning clustering scheme to model the distribution of data points within each cluster, thereby more accurately describing the similarities and differences between data points. By combining the network loss and the clustering loss, the optimization of model parameters is more comprehensive, considering not only the reconstruction quality of the data but also the clustering effect, improving the generalization ability of the model.

[0191] The above mainly introduces the solution provided by the embodiments of the present application from the perspective of the method. To implement the above functions, the driving style determination device or electronic device includes the corresponding hardware structure and / or software module for executing each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described function for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0192] The embodiments of the present application can, according to the above method, exemplarily divide the functional modules of the driving style determination device or electronic device. For example, the driving style determination device or electronic device can include each functional module corresponding to each functional division, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. It should be noted that the division of modules in the embodiments of the present application is illustrative, merely a logical functional division, and there can be other division methods in actual implementation.

[0193] Figure 9 is a block diagram of a driving style determination device shown according to an exemplary embodiment. Referring to Figure 9 , the driving style determination device 900 includes: an acquisition module 901 and a processing module 902.

[0194] The acquisition module 901 is configured to acquire driving data of multiple data types of a vehicle; the processing module 902 is configured to determine a first frequency domain feature parameter corresponding to the driving data of each data type among the driving data of multiple data types, where the first frequency domain feature parameter is used to characterize the frequency domain feature; the processing module 902 is further configured to determine a second frequency domain feature parameter corresponding to the driving data of each data type by performing inter-sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types, where the second frequency domain feature parameter integrates the frequency domain features of the driving data of other data types; the processing module 902 is further configured to determine the driving style corresponding to the vehicle based on the second frequency domain feature parameter corresponding to the driving data of each data type.

[0195] In a possible implementation manner, the first frequency domain feature parameter includes a spectrum parameter; the processing module 902 is specifically configured to perform frequency domain transformation on the driving data of each data type among the driving data of multiple data types based on a preset frequency to obtain a spectrum parameter corresponding to the driving data of each data type.

[0196] In a possible implementation manner, the processing module 902 is specifically configured to construct a spectrum vector corresponding to the driving data of multiple data types based on the first frequency domain feature parameter corresponding to the driving data of each data type; the processing module 902 is specifically configured to determine a query vector, a key vector, and a value vector corresponding to the spectrum vector; the processing module 902 is specifically configured to determine an attention weight between the query vector and the key vector; the processing module 902 is specifically configured to determine a second frequency domain feature parameter corresponding to the driving data of each data type based on the attention weight and the value vector.

[0197] In a possible implementation manner, the processing module 902 is specifically configured to process the second frequency domain feature parameter corresponding to the driving data of each data type through causal dilated convolution to obtain a processed second frequency domain feature parameter; the processing module 902 is specifically configured to perform sampling processing on the processed second frequency domain feature parameter to obtain a sampled second frequency domain feature parameter; the processing module 902 is specifically configured to perform time domain transformation on the sampled second frequency domain feature parameter to obtain a time domain feature parameter corresponding to the sampled second frequency domain feature parameter; the processing module 902 is specifically configured to determine the driving style corresponding to the vehicle based on the time domain feature parameter.

[0198] In a possible implementation, the processing module 902 is specifically configured to perform data segmentation processing on the time-domain feature parameters based on a preset step size to obtain multiple data segments; the processing module 902 is specifically configured to determine the characteristic data corresponding to each of the multiple data segments, and the characteristic data includes at least one of the following: average value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter, and the driving fluctuation parameter is used to indicate the degree of deviation of the data from the reference value; the processing module 902 is specifically configured to determine the driving style corresponding to the vehicle based on the characteristic data corresponding to each of the multiple data segments.

[0199] In a possible implementation, the processing module 902 is specifically configured to determine the driving style corresponding to each data segment through the K-means clustering algorithm based on the characteristic data corresponding to each of the multiple data segments; the processing module 902 is specifically configured to determine the driving style corresponding to the vehicle based on the driving style corresponding to each data segment.

[0200] In a possible implementation, the driving styles include: conservative driving style, normal driving style, and aggressive driving style; the processing module 902 is specifically configured to determine the number of data segments corresponding to each driving style; the processing module 902 is specifically configured to determine the driving style evaluation parameter based on the number of data segments corresponding to each driving style and the weight parameter of each driving style; the processing module 902 is specifically configured to determine the driving style corresponding to the vehicle based on the driving style evaluation parameter.

[0201] In a possible implementation, the processing module 902 is specifically configured to determine that the driving style corresponding to the vehicle is a conservative driving style when the driving style evaluation parameter is greater than or equal to the first preset parameter and less than the second preset parameter, and the second preset parameter is greater than the first preset parameter; the processing module 902 is specifically configured to determine that the driving style corresponding to the vehicle is a normal driving style when the driving style evaluation parameter is greater than or equal to the second preset parameter and less than the third preset parameter, and the third preset parameter is greater than the second preset parameter; the processing module 902 is specifically configured to determine that the driving style corresponding to the vehicle is an aggressive driving style when the driving style evaluation parameter is greater than or equal to the third preset parameter.

[0202] In a possible implementation, an acquisition module 901 is specifically configured to acquire basic data of multiple data types collected by sensors of a vehicle. The multiple data types include: vehicle power data and driving operation data. The vehicle power data includes at least one of the following: driving speed, longitudinal acceleration, lateral acceleration, motor speed, and motor torque. The driving operation data includes at least one of the following: accelerator pedal angle, brake pedal angle, and steering angle. A processing module 902 is further configured to preprocess the basic data of the multiple data types to obtain preprocessed driving data of the multiple data types. The preprocessing includes at least one of the following: abnormal data error correction processing, missing data filling processing, and data noise reduction processing.

[0203] In a possible implementation, the basic data of the multiple data types includes data collected by the vehicle in multiple driving environments. The multiple driving environments include: highways, urban roads, commercial blocks, residential roads, and rural roads.

[0204] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0205] Figure 10 is a block diagram of an electronic device shown according to an exemplary embodiment. As Figure 10 shown, the electronic device 1000 includes, but is not limited to: a processor 1001 and a memory 1002.

[0206] Among them, the above-mentioned memory 1002 is used to store executable instructions of the above-mentioned processor 1001. It can be understood that the above-mentioned processor 1001 is configured to execute instructions to implement the driving style determination method in the above embodiments.

[0207] It should be noted that those skilled in the art can understand that Figure 10 the structure of the electronic device shown in Figure 10 does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than

[0208] The processor 1001 is the control center of the electronic device, connecting various parts of the entire electronic device through various interfaces and circuits. By running or executing software programs and / or modules stored in the memory 1002, and by invoking the data stored in the memory 1002, it performs various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole. The processor 1001 may include one or more processing units. Optionally, the processor 1001 may integrate an application processor and a modem processor. Among them, the application processor mainly processes the operating system, user interface, application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor may not be integrated into the processor 1001 either.

[0209] The memory 1002 can be used to store software programs and various data. The memory 1002 mainly includes a program storage area and a data storage area. Among them, the program storage area can store the operating system, application programs required by at least one functional module (such as a processing module, a storage module, etc.). In addition, the memory 1002 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0210] In an exemplary embodiment, there is also provided a computer-readable storage medium including instructions, such as the memory 1002 including instructions. The above instructions can be executed by the processor 1001 of the electronic device 1000 to implement the driving style determination method in the above embodiment.

[0211] In actual implementation, Figure 9 the functions of the processing module 902 in Figure 10 can all be implemented by the processor 1001 in

[0212] invoking the computer program stored in the memory 1002. The specific execution process can refer to the description of the driving style determination method part in the above embodiment, which will not be elaborated here.

[0213] In an exemplary embodiment, an embodiment of the present application further provides a computer program product including one or more instructions, which can be executed by the processor 1001 of the electronic device 1000 to complete the driving style determination method in the above embodiment.

[0214] It should be noted that when the instructions in the above computer-readable storage medium or the one or more instructions in the computer program product are executed by the processor of the electronic device, each process of the above embodiment of the driving style determination method is implemented, and the same technical effects as those of the above driving style determination method can be achieved. To avoid repetition, details are not described here again.

[0215] Figure 11 FIG. is a schematic structural diagram of a computer system shown according to an exemplary embodiment. It should be noted that Figure 11 the computer system of the electronic device shown is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present application.

[0216] As Figure 11 shown, the computer system includes a central processing unit (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) or the program loaded from the storage section into the random access memory (RAM), such as executing the method in the above embodiment. In the RAM, various programs and data required for system operation are also stored. The CPU, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus. The I / O interface is used to implement functions such as input, output, communication, and storage of data, and the storage function can be specifically implemented through a removable medium.

[0217] The following components are connected to the I / O interface: an input section including a keyboard, a mouse, etc.; an output section including a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section including a hard disk, etc.; and a communication section including a network interface card such as a local area network (LAN) card, a modem, etc. The communication section performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media, such as magnetic disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that the computer programs read from them can be installed into the storage section as needed.

[0218] In particular, according to an embodiment of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product that includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), various functions defined in the system of the present application are executed.

[0219] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a CD-ROM, an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The computer program included on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0220] Through the description of the above embodiments, those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above division of each functional module is used as an example. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.

[0221] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules or units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical or other forms.

[0222] The units described as separate components may or may not be physically separated. The components displayed as units can be one physical unit or multiple physical units, that is, they can be located in one place, or they can be distributed to multiple different places. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0223] In addition, in each embodiment of this application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0224] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to enable a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the methods of the embodiments of this application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0225] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. A driving style determination method, characterized in that, The driving style determination method includes: Obtaining driving data of multiple data types of the vehicle; Determining a first frequency domain feature parameter corresponding to the driving data of each data type among the driving data of the multiple data types, where the first frequency domain feature parameter is used to characterize the frequency domain feature; By performing inter-sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types, determining a second frequency domain feature parameter corresponding to the driving data of each data type, where the second frequency domain feature parameter integrates the frequency domain features of the driving data of other data types; Based on the second frequency domain feature parameter corresponding to the driving data of each data type, determining the driving style corresponding to the vehicle.

2. The driving style determination method according to claim 1, wherein The first frequency domain feature parameter includes a spectrum parameter; The determining of the first frequency domain feature parameter corresponding to the driving data of each data type among the driving data of the multiple data types includes: For the driving data of each data type among the driving data of the multiple data types, performing a frequency domain transformation on the driving data of each data type based on a preset frequency to obtain a spectrum parameter corresponding to the driving data of each data type.

3. The driving style determination method according to claim 1, characterized in that The determining of the second frequency domain feature parameter corresponding to the driving data of each data type by performing inter-sequence correlation learning on the first frequency domain feature parameters corresponding to the driving data of multiple data types includes: Based on the first frequency domain feature parameter corresponding to the driving data of each data type, constructing a spectrum vector corresponding to the driving data of the multiple data types; Determining a query vector, a key vector, and a value vector corresponding to the spectrum vector; Determining the attention weight between the query vector and the key vector; Based on the attention weight and the value vector, determining the second frequency domain feature parameter corresponding to the driving data of each data type.

4. The driving style determination method according to claim 1, wherein The determining of the driving style corresponding to the vehicle based on the second frequency domain feature parameter corresponding to the driving data of each data type includes: Processing the second frequency domain feature parameter corresponding to the driving data of each data type through causal dilated convolution to obtain a processed second frequency domain feature parameter; Performing a sampling process on the processed second frequency domain feature parameter to obtain a sampled second frequency domain feature parameter; Performing a time domain transformation on the sampled second frequency domain feature parameter to obtain a time domain feature parameter corresponding to the sampled second frequency domain feature parameter; Based on the time domain feature parameter, determining the driving style corresponding to the vehicle.

5. The driving style determination method according to claim 4, wherein The determining of the driving style corresponding to the vehicle based on the time domain feature parameter includes: Based on a preset step size, performing data segmentation processing on the time domain feature parameter to obtain multiple data segments; Determining the characteristic data corresponding to each data segment among the multiple data segments, where the characteristic data includes at least one of the following: average value, variance, maximum value, minimum value, median, standard deviation, kurtosis, skewness, and driving fluctuation parameter, and the driving fluctuation parameter is used to indicate the degree of deviation of the data from a reference value; Based on the characteristic data corresponding to each data segment among the multiple data segments, determining the driving style corresponding to the vehicle.

6. The driving style determination method according to claim 5, characterized in that The determining of the driving style corresponding to the vehicle based on the characteristic data corresponding to each data segment among the multiple data segments includes: Based on the characteristic data corresponding to each of the multiple data segments, determine the driving style corresponding to each data segment through the K-means clustering algorithm; Based on the driving style corresponding to each data segment, determine the driving style corresponding to the vehicle.

7. The driving style determination method according to claim 6, characterized in that, The driving styles include: a conservative driving style, a normal driving style, and an aggressive driving style; The determining the driving style corresponding to the vehicle based on the driving style corresponding to each data segment includes: Determine the number of data segments corresponding to each driving style; Based on the number of data segments corresponding to each driving style and the weight parameters of each driving style, determine the driving style evaluation parameter; Based on the driving style evaluation parameter, determine the driving style corresponding to the vehicle.

8. The driving style determination method according to claim 7, wherein The determining the driving style corresponding to the vehicle based on the driving style evaluation parameter includes: In the case where the driving style evaluation parameter is greater than or equal to the first preset parameter and less than the second preset parameter, determine that the driving style corresponding to the vehicle is the conservative driving style, and the second preset parameter is greater than the first preset parameter; In the case where the driving style evaluation parameter is greater than or equal to the second preset parameter and less than the third preset parameter, determine that the driving style corresponding to the vehicle is the normal driving style, and the third preset parameter is greater than the second preset parameter; In the case where the driving style evaluation parameter is greater than or equal to the third preset parameter, determine that the driving style corresponding to the vehicle is the aggressive driving style.

9. The driving style determination method according to claim 1, wherein The obtaining the driving data of multiple data types of the vehicle includes: Obtain the basic data of multiple data types collected by the vehicle sensors, and the multiple data types include: vehicle power data and driving operation data. The vehicle power data includes at least one of the following: driving speed, longitudinal acceleration, lateral acceleration, motor speed, motor torque. The driving operation data includes at least one of the following: accelerator pedal angle, brake pedal angle, steering angle; Preprocess the basic data of the multiple data types to obtain the preprocessed driving data of the multiple data types. The preprocessing includes at least one of the following: abnormal data error correction processing, missing data filling processing, data noise reduction processing.

10. The driving style determination method according to claim 1, wherein The basic data of the multiple data types includes the data collected by the vehicle in multiple driving environments, and the multiple driving environments include: highways, urban roads, commercial blocks, residential roads, rural roads.

11. A driving style determination device, characterized in that, The driving style determination device includes: an acquisition module and a processing module; The acquisition module is used to acquire the driving data of multiple data types of the vehicle; The processing module is used to determine the first frequency domain feature parameter corresponding to the driving data of each data type in the driving data of the multiple data types, and the first frequency domain feature parameter is used to characterize the frequency domain feature; The processing module is further used to determine the second frequency domain feature parameter corresponding to the driving data of each data type by performing sequence correlation learning on the frequency domain feature parameters corresponding to the driving data of multiple data types. The second frequency domain feature parameter integrates the frequency domain features of the driving data of other data types; The processing module is further configured to determine the driving style corresponding to the vehicle based on the second frequency domain characteristic parameters corresponding to the driving data of each data type.

12. An electronic device, characterized in that, Comprising: a processor and a memory; wherein, the memory is configured to store one or more programs, the one or more programs include computer execution instructions, when the electronic device runs, the processor executes the computer execution instructions stored in the memory, and the electronic device executes the driving style determination method according to any one of claims 1-10.

13. A vehicle, characterized in that, The vehicle includes the driving style determination device according to claim 11, and the vehicle is configured to implement the driving style determination method according to any one of claims 1-10.

14. A driving style determination system, characterized in that, The driving style determination system includes: a sensor and a controller; The sensor is configured to obtain driving data of multiple data types of the vehicle; The controller is configured to determine first frequency domain characteristic parameters corresponding to the driving data of each data type in the driving data of multiple data types, and the first frequency domain characteristic parameters are used to characterize frequency domain characteristics; The controller is further configured to determine second frequency domain characteristic parameters corresponding to the driving data of each data type by performing sequence correlation learning on the frequency domain characteristic parameters corresponding to the driving data of multiple data types, and the second frequency domain characteristic parameters fuse the frequency domain characteristics of the driving data of other data types; The controller is further configured to determine the driving style corresponding to the vehicle based on the second frequency domain characteristic parameters corresponding to the driving data of each data type.

Images