Driving style identification method, device and equipment

Abstract

The invention is suitable for the technical field of intelligent traffic driving behavior analysis, and provides a driving style identification method, device and equipment, and the method comprises the steps: obtaining the current driving data and current scene information of a current vehicle; according to the current scene information, a first driving style rule base in the driving style fuzzy recognition system is adjusted to obtain a second driving style rule base, and the first driving style rule base is obtained by simplifying a pre-constructed initial driving style rule base; the second driving style rule base comprises a plurality of second driving style rules, carrying out matching calculation on the current driving data and each second driving style rule to obtain the matching rule strength of each second driving style rule, and according to the matching rule strength of all the second driving style rules and the current vehicle, determining the current driving style of the vehicle. And determining the current driving style of the driver. According to the invention, the technical problems of low identification accuracy and insufficient dynamic adaptability of the driving style of the driver in the prior art can be solved.

Description

Technical Field

[0001] The present application belongs to the technical field of intelligent traffic driving behavior analysis, and in particular relates to a driving style recognition method, device and equipment. Background Art

[0002] Different drivers exhibit different driving styles, and drivers with different driving styles face different risks. For example, aggressive drivers may frequently accelerate and brake suddenly, and are more likely to exceed the speed limit; whereas cautious drivers tend to obey traffic rules and drive at a more stable speed. By identifying driving styles, the vehicle computer system can issue more frequent and stronger safety warnings to aggressive drivers, such as speeding warnings and warnings for following too close, effectively reducing the possibility of accidents.

[0003] Currently, the main method for identifying a driver's driving style is fuzzy reasoning. However, these methods rely primarily on fixed rule bases or empirical data analysis during the reasoning process, failing to consider external factors that influence the driver's actual driving experience. This results in low accuracy in identifying the driver's driving style and insufficient dynamic adaptability. Summary of the Invention

[0004] The embodiments of the present application provide a driving style recognition method, device, and apparatus, which can solve the technical problems in the prior art of low recognition accuracy and insufficient dynamic adaptability of a driver's driving style.

[0005] In a first aspect, an embodiment of the present application provides a method for identifying a driving style, comprising:

[0006] Get the current driving data and current scene information of the current vehicle;

[0007] adjusting a first driving style rule base in the driving style fuzzy recognition system based on the current scene information to obtain a second driving style rule base; wherein the first driving style rule base is obtained by simplifying a pre-built initial driving style rule base; and the second driving style rule base includes a plurality of second driving style rules;

[0008] performing matching calculations on the current driving data and each of the second driving style rules to obtain a matching rule strength between each of the second driving style rules and the current vehicle;

[0009] The current driving style of the driver driving the current vehicle is determined according to the matching rule strengths of all the second driving style rules with the current vehicle.

[0010] In a possible implementation of the first aspect, the driving style fuzzy identification system includes variable fuzzy set membership functions of a plurality of input variables;

[0011] The second driving style rule includes preconditions of each of the input variables;

[0012] Performing matching calculations on the current driving data and each of the second driving style rules to obtain a matching rule strength between each of the second driving style rules and the current vehicle includes:

[0013] Performing feature value extraction on the current driving data of the current vehicle to obtain parameter values ​​of a plurality of the input variables of the current vehicle;

[0014] calculating, according to the variable fuzzy set membership functions of the plurality of input variables in the driving style fuzzy recognition system and based on the parameter values ​​of the plurality of input variables, the variable fuzzy set membership degree of each input variable;

[0015] For each of the second driving style rules, the matching rule strength between the second driving style rule and the current vehicle is determined based on the preconditions of all the input variables of the second driving style rule and the variable fuzzy set membership of each of the input variables of the current vehicle.

[0016] In a possible implementation of the first aspect, determining, for each second driving style rule, the matching rule strength between the second driving style rule and the current vehicle based on preconditions of all the input variables of the second driving style rule and variable fuzzy set membership of each input variable of the current vehicle includes:

[0017] For each of the second driving style rules, determining a conditional membership of each of the input variables of the second driving style rule based on preconditions of all of the input variables of the second driving style rule and the variable fuzzy set membership of each of the input variables of the current vehicle;

[0018] The matching rule strength between the second driving style rule and the current vehicle is determined according to the conditional membership of all the input variables of the second driving style rule.

[0019] In a possible implementation of the first aspect, the second driving style rule includes a driving style conclusion;

[0020] The determining, based on the matching rule strengths of all the second driving style rules with the current vehicle, the current driving style of the driver of the current vehicle includes:

[0021] aggregating the driving style conclusions of all the second driving style rules according to the matching rule strength between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle;

[0022] performing a defuzzification process on the driving style fuzzy set of the current vehicle to obtain a driving style output value of the current vehicle;

[0023] Based on the driving style output value of the current vehicle, a current driving style of a driver driving the current vehicle is determined.

[0024] In a possible implementation of the first aspect, the driving style fuzzy recognition system includes a plurality of output fuzzy set membership functions of driving styles;

[0025] The driving style conclusions of all the second driving style rules are aggregated according to the matching rule strength between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle, including:

[0026] determining an activated driving style rule from the second driving style rule library according to a matching rule strength between each of the second driving style rules and the current vehicle;

[0027] cutting, based on the matching rule strength between each of the activated driving style rules and the current vehicle, the output fuzzy set membership function corresponding to the driving style conclusion of each of the activated driving style rules to obtain a valid output fuzzy set membership function corresponding to the driving style conclusion of each of the activated driving style rules;

[0028] All the activated driving style rules are combined according to the valid output fuzzy set membership functions corresponding to the driving style conclusions of all the activated driving style rules to obtain the driving style fuzzy set of the current vehicle.

[0029] In a possible implementation of the first aspect, before adjusting the first driving style rule base in the driving style fuzzy recognition system based on the current scene information to obtain the second driving style rule base, the method further includes:

[0030] Acquire a historical driving data set and a driving style corresponding to each piece of historical driving data in the historical driving data set;

[0031] establishing an initial driving style rule base based on the historical driving data set and the driving style corresponding to each piece of historical driving data in the historical driving data set; wherein the initial driving style rule base includes a plurality of initial driving style rules, each of the initial driving style rules including a precondition of an input variable;

[0032] All the initial driving style rules in the initial driving style rule base are clustered and simplified according to the preconditions of the input variables of all the initial driving style rules to obtain a first driving style rule base.

[0033] In a possible implementation of the first aspect, each of the initial driving style rules further includes a driving style conclusion;

[0034] The driving style fuzzy recognition system includes variable fuzzy set membership functions of a plurality of the input variables;

[0035] The method further comprises: clustering and simplifying all the initial driving style rules in the initial driving style rule base according to the preconditions of the input variables of all the initial driving style rules to obtain a first driving style rule base, including:

[0036] quantizing the premise conditions of the input variables of each of the initial driving style rules according to variable fuzzy set membership functions of the plurality of input variables in the driving style fuzzy recognition system to obtain premise quantization vectors of each of the initial driving style rules;

[0037] clustering all the initial driving style rules according to the premise quantization vectors of all the initial driving style rules to obtain a plurality of driving style rule clusters;

[0038] For each driving style rule cluster, determining a centroid of the driving style rule cluster according to the premise quantized vectors of all the initial driving style rules in the driving style rule cluster;

[0039] For each driving style rule cluster, determining a first driving style rule corresponding to the driving style rule cluster according to the centroid of the driving style rule cluster and the driving style conclusions of all the initial driving style rules in the driving style rule cluster;

[0040] The first driving style rule library is determined according to the first driving style rules corresponding to all the driving style rule clusters.

[0041] In a second aspect, an embodiment of the present application provides a driving style recognition device, comprising:

[0042] A data acquisition module is used to obtain the current driving data and current scene information of the current vehicle;

[0043] a rule base establishment module, configured to adjust a first driving style rule base in the driving style fuzzy recognition system based on the current scene information to obtain a second driving style rule base; wherein the first driving style rule base is obtained by simplifying a pre-established initial driving style rule base; and the second driving style rule base includes a plurality of second driving style rules;

[0044] a matching calculation module, configured to perform matching calculations on the current driving data and each of the second driving style rules to obtain a matching rule strength between each of the second driving style rules and the current vehicle;

[0045] The driving style recognition module is configured to determine a current driving style of a driver driving the current vehicle based on the matching rule strengths of all the second driving style rules with the current vehicle.

[0046] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the above-described methods when executing the computer program.

[0047] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements any of the methods described above.

[0048] In a fifth aspect, an embodiment of the present application provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute any of the methods described above.

[0049] Compared with the prior art, the embodiments of the present application have the following advantages: by obtaining the current driving data and current scene information of the current vehicle, adjusting the first driving style rule base in the driving style fuzzy recognition system according to the current scene information to obtain the second driving style rule base, matching and calculating the current driving data with each second driving style rule, obtaining the matching rule strength of each second driving style rule with the current vehicle, and determining the current driving style of the driver of the current vehicle based on the matching rule strength of all the second driving style rules with the current vehicle, since the first driving style rule base is obtained by simplifying the pre-built initial driving style rule base, and the second driving style rule base is based on the pre-built initial driving style rule base. The second driving style rule base is obtained by adjusting the first driving style rule base based on the current scene information, so that the second driving style rule base is also a simplified rule base. When identifying the driver's current driving style based on the second driving style rule base, time for matching and calculating with the second driving style rules in the second driving style rule base can be saved, thereby improving the efficiency of identifying the driver's current driving style. Moreover, the second driving style rule base is a rule adjusted based on the current scene information, which can adapt to the differences in driving behavior in the current scene, thereby improving the accuracy of driving style identification of the current vehicle in the current scene. Therefore, the technical problems of low driver driving style identification accuracy and insufficient dynamic adaptability in the prior art are effectively solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 This is a flowchart of a driving style recognition method provided by an embodiment of the present application;

[0052] Figure 2 This is a schematic structural diagram of a driving style fuzzy recognition system provided by an embodiment of the present application;

[0053] Figure 3 This is a schematic diagram of the structure of a variable fuzzy set membership function of speed provided by an embodiment of the present application;

[0054] Figure 4 This is a structural diagram of a driving style recognition device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0055] Driving styles can be categorized into multiple categories, reflecting the driver's habits and behavioral characteristics while driving. For example, driving styles can be divided into three types: conservative, moderate, and aggressive. Drivers with different driving styles face different accident risks. For example, aggressive drivers often accelerate, brake suddenly, or speed. This behavior significantly increases the probability of accidents such as rear-end collisions and rollovers. By identifying these driving styles, the in-vehicle terminal can remind the driver to pay attention to driving safety, thereby reducing the risk of accidents.

[0056] Currently, methods for identifying a driver's driving style include those based on statistical models, machine learning models, or fuzzy inference systems. However, precise mathematical models (statistical and machine learning models) are difficult to use to handle the ambiguity and uncertainty inherent in driving style conclusions. Fuzzy inference systems, on the other hand, can handle uncertainty and offer easy-to-understand rules, making them suitable for driving style identification. However, existing fuzzy inference systems often rely on empirical experience or preliminary data analysis when constructing their rule bases. This results in a large number of similar or redundant preconditions within the rule base, increasing the complexity of driving style identification.

[0057] Furthermore, existing driving style fuzzy inference systems rely primarily on a fixed rule base constructed from driving data under normal scenarios during the inference process, without considering the impact of different scenarios on driving behavior. However, in actual driving, different scenario factors such as weather changes, road type, and traffic conditions have a significant impact on the inference process of the driver's driving style. For example, a braking frequency that is considered a moderate driving style on clear days may be identified as an aggressive driving style in rainy or snowy weather. Existing driving style fuzzy inference systems often ignore these dynamic changes, resulting in insufficient dynamic adaptability in the existing driving style recognition process based on fuzzy inference systems and low recognition accuracy.

[0058] To address the above technical issues, the present application provides a driving style recognition method, device, and storage medium. The following, in conjunction with the accompanying drawings, provides a clear and complete description of the technical solutions in the present application. Obviously, the described embodiments represent only a portion of the present application, not all of the embodiments. All other embodiments derived by persons of ordinary skill in the art based on the present application without inventive effort are intended to fall within the scope of protection of this application.

[0059] The present application provides a method for identifying a driving style. Figure 1 , Figure 1: This is a flowchart of a driving style recognition method provided by an embodiment of the present application, including:

[0060] Step S11, obtaining the current driving data and current scene information of the current vehicle.

[0061] Step S12: Adjusting the first driving style rule base in the driving style fuzzy recognition system based on the current scene information to obtain a second driving style rule base. The first driving style rule base is obtained by simplifying a pre-built initial driving style rule base. The second driving style rule base includes a plurality of second driving style rules.

[0062] Step S13: performing matching calculations on the current driving data and each second driving style rule to obtain matching rule strengths between each second driving style rule and the current vehicle.

[0063] Step S14: determining the current driving style of the driver of the current vehicle according to the matching rule strengths of all the second driving style rules and the current vehicle.

[0064] It should be noted that the vehicle is equipped with an onboard terminal, which may be equipped with a visual interface that can display notification messages or vehicle-related information. The data acquisition module includes onboard sensors and other equipment, which can collect real-time vehicle driving data, vehicle operating status, vehicle fault information, vehicle location information, etc. Vehicle driving data may include vehicle speed, acceleration, steering angular velocity, accelerator pedal frequency, etc. Through the data acquisition module, the onboard terminal can obtain vehicle driving data, etc.

[0065] The on-board terminal includes an on-board OBU (On-Board Unit) module. The on-board OBU module establishes V2X communication with the roadside RSU (Road Side Unit) module, and can transmit the vehicle's driving data and other data to the road-side edge computing unit MEC (Multi-access Edge Computing) for data preprocessing to extract effective driving feature data. Among them, effective driving feature data can be set according to specific scenarios and specific needs. V2X communication is a kind of wireless communication. The road-side edge computing unit MEC can also perform corresponding processing and analysis on the vehicle's driving data, and feed back the results of the processing and analysis to the on-board terminal, which will be displayed in the on-board terminal to remind the driver.

[0066] Through the above-mentioned Internet of Vehicles technology, using on-board sensors, wireless communication equipment and other equipment, various data during the vehicle's driving process can be collected and transmitted in real time, providing a data basis for the driver's driving behavior analysis.

[0067] Specifically, to analyze the driver's driving style, the vehicle driven by the driver in question is determined to be the current vehicle. The vehicle's data acquisition module and onboard terminal collect the vehicle's current driving data and scene information and transmit it to the MEC. The MEC then acquires the vehicle's current driving data and scene information.

[0068] The current driving data may include the vehicle's speed, acceleration, deceleration, accelerator pedal opening, brake pedal opening, and pedaling frequency over a specific time period, which may be any one or more of the variables. The current scene information may include weather information, road section information, vehicle model, time information, etc. at the vehicle's current location.

[0069] It should be noted that the current location of the vehicle can be obtained through GPS and other devices, and weather information can be obtained by accessing the weather information system. This application does not make specific limitations on this.

[0070] The road-end edge computing unit MEC can also include a driving style fuzzy recognition system, which can analyze and identify the current driving data and current scene information of the driver driving the vehicle, and output the current driving style of the driver driving the current vehicle.

[0071] The driving style fuzzy recognition system is a multi-input, single-output fuzzy inference system based on fuzzy reasoning. Fuzzy reasoning is an effective tool for handling uncertainty and ambiguity, using fuzzy sets and fuzzy rules to describe complex inference systems. For example, input variables such as speed and acceleration may have different meanings and importance in different driving scenarios, making it difficult to accurately classify them using a single threshold or rule. Therefore, fuzzy reasoning is used to establish variable fuzzy sets for multiple input variables and output fuzzy sets for driving style, thus constructing a fuzzy inference system that adapts to actual driving environments.

[0072] In an optional example, when the input variable is speed, the corresponding variable fuzzy set includes very slow speed, slow speed, medium speed, fast speed, and very fast speed. When the input variable is acceleration, the corresponding variable fuzzy set includes small acceleration, medium acceleration, and large acceleration. When the input variable is negative acceleration, the corresponding fuzzy set includes small negative acceleration, medium negative acceleration, and large negative acceleration. When the input variable is accelerator pedal opening variance, the corresponding variable fuzzy set includes small accelerator pedal opening variance, medium accelerator pedal opening variance, and large accelerator pedal opening variance. For example, the variable fuzzy set of speed is represented by {very slow speed, slow speed, medium speed, fast speed, very fast speed}, and its domain is [0,120] km / h. The domain of a variable represents the range of values ​​of the variable.

[0073] Specifically, the driver's current driving style can be divided into conservative, moderate and aggressive types, which are represented by the output fuzzy set of driving style. That is, the output fuzzy set of driving style includes conservative, moderate and aggressive types.

[0074] When the output driver's current driving style is conservative, it means that the driver generally avoids unnecessary speeding and pays more attention to traffic safety. When the output driver's current driving style is aggressive, it means that the driver tends to accelerate suddenly and drive at high speeds. When the output driver's current driving style is moderate, it means that the driver can flexibly adjust the speed according to road conditions, neither too fast nor too slow.

[0075] The driving style fuzzy recognition system pre-establishes an initial driving style rule base, which includes multiple initial driving style rules. Each initial driving style rule includes the preconditions of the input variables and the driving style conclusion. The input variables are the input variables of the driving style fuzzy recognition system and serve as the basis for judging driving style. They include speed, acceleration, negative acceleration, and accelerator pedal opening variance. When there are multiple input variables, many initial driving style rules are established. The driving style conclusion is one of conservative, moderate, and aggressive. The initial driving style rule base can be generated by the driver or domain experts, or it can be automatically constructed based on scenario information using artificial intelligence technology (such as a large language model).

[0076] In one optional example, an "IF-THEN" rule is generated based on domain expert knowledge and experience. This "IF-THEN" rule is an initial driving style rule that describes the logical relationship between input variables and driving style. For example, an initial driving style rule could be: "IF the speed is very fast, and the acceleration is large, and the negative acceleration is large, and the accelerator pedal opening variance is large, THEN the driving style is aggressive." Here, "IF the speed is very fast, and the acceleration is large, and the negative acceleration is large, and the accelerator pedal opening variance is large" are the prerequisites for this initial driving style rule, and "aggressive" is the driving style conclusion of this initial driving style rule.

[0077] In an optional example, it is assumed that the initial driving style rule base includes:

[0078] Initial driving style rule 1: The speed is very slow AND the acceleration is small AND the negative acceleration is small AND the accelerator pedal opening variance is small, the driving style is conservative;

[0079] Initial driving style rule 2: medium speed AND medium acceleration AND medium negative acceleration AND medium accelerator pedal opening variance, the driving style is normal;

[0080] Initial driving style rule 3: Faster speed AND greater acceleration AND greater negative acceleration AND greater variance in accelerator pedal opening, indicating an aggressive driving style;

[0081] Initial driving style rule 4: very fast speed AND large acceleration AND large negative acceleration AND large variance of accelerator pedal opening, the driving style is aggressive;

[0082] Initial driving style rule 5: slow speed AND small acceleration AND large negative acceleration AND small accelerator pedal opening variance, the driving style is conservative;

[0083] Initial driving style rule 6: medium speed AND large acceleration AND small negative acceleration AND large variance of accelerator pedal opening, the driving style is aggressive;

[0084] …….

[0085] Specifically, the number of initial driving style rules in the initial driving style rule base in the driving style fuzzy recognition system depends on the number of input variables and the number of variable fuzzy sets in the variable fuzzy set of each input variable.

[0086] Because each input variable influences the driving style conclusion, and similar initial driving style rules exist within the initial driving style rule base, each with different input variable preconditions but the same driving style conclusion. If a large number of similar or redundant initial driving style rules exist within the initial driving style rule base, without simplification, subsequent identification of the current vehicle's driving style conclusion requires comparing each initial driving style rule, which is time-consuming and inefficient.

[0087] To improve the efficiency of driving style recognition, all initial driving style rules can be clustered and simplified. For example, similar rules can be merged using a clustering algorithm. Multiple similar initial driving style rules can be merged to share a common driving style conclusion, forming a first driving style rule. All first driving style rules form a first driving style rule base. This simplified first driving style rule base can improve the efficiency of driving style recognition. Furthermore, when manual annotation of driving style conclusions of the first driving style rules is required, the annotation process can be more efficient, thereby improving the efficiency of establishing the first driving style rule base.

[0088] Current scene information represents the various road and weather conditions encountered during a vehicle's driving. It encompasses factors such as the geographic environment, weather conditions, and traffic congestion, and serves as a key basis for identifying driving styles. As an external factor, current scene information can also constrain or influence the driver's driving behavior, thereby impacting the driver's driving style. For example, a speed considered moderate in clear weather would be considered aggressive in rainy or snowy conditions.

[0089] like Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of a driving style fuzzy recognition system provided by one embodiment of the present application. Based on scene information, the system identifies the current scene. For example, if the current scene information is weather information, the system identifies the current scene as Scenario 1 - Rainy Weather, Scenario 2 - Snowy Weather, Scenario 3 - Windy Weather, and so on. Based on expert knowledge, a driving style rule base for each scenario is generated. For example, a fuzzy rule base for Scenario 1, a fuzzy rule base for Scenario 2, a fuzzy rule base for Scenario 3, and so on, for Scenario N. N is an integer. After the scene is identified, the system then identifies the driving style for that scene based on the fuzzy rule base corresponding to that scene.

[0090] In an optional example, the first driving style rule base in the driving style fuzzy recognition system is adjusted based on current scenario information. This adjustment can be made by adjusting the domain (threshold) or output fuzzy set membership function (coefficient) of the first driving style rule, or by adjusting the preconditions of the input variables or variable fuzzy set membership function of the first driving style rule. This adjustment results in a second driving style rule base. The second driving style rule base includes multiple second driving style rules. Dynamically adjusting or generating the second driving style rule base based on scenario information can adapt to driving behavior differences in different scenarios, thereby improving the accuracy, adaptability, and versatility of fuzzy reasoning.

[0091] In another specific example, if the second driving style rule base corresponding to a particular scenario differs significantly from the first driving style rule base, a new second driving style rule base can be generated. This new second driving style rule base can be generated manually based on domain experience, or by regenerating the second driving style rule base containing the scenario information based on a large language model. The large language model can be generated based on the input variables of the driving style fuzzy recognition system and the prompt words used to determine the driving style. However, if the second driving style rule base corresponding to a particular scenario differs slightly from the first driving style rule base, the first driving style rule base can be fine-tuned to generate a new second driving style rule base.

[0092] For example, when only the current scene information is weather change, the generation process of the second driving style rule base is as follows:

[0093] Assume the current weather is clear, which corresponds to the first driving style rule base. Quantize the driving style conclusions of all first driving style rules into integer ranges, with larger values ​​indicating an aggressive driving style. For example, a first driving style rule with a quantized driving style conclusion of [1, 3] corresponds to a conservative driving style; a quantized driving style conclusion of [4, 6] corresponds to a moderate driving style; and a quantized driving style conclusion of [7, 9] corresponds to an aggressive driving style. For example, a first driving style rule may state, "If the speed is low, the acceleration is medium, the negative acceleration is medium, and the accelerator pedal opening variance is small, then the driving style conclusion is 3." This means that the driving style conclusion of this first driving style rule corresponds to a "conservative" driving style.

[0094] When the weather changes, for example, if the current weather information changes from sunny to rainy or snowy, the range of the quantized driving style conclusion of the first driving style rule is modified or the quantized result is adjusted to form the second driving style rule. When the quantized driving style conclusion value of the second driving style rule is [1, 2], it corresponds to a conservative driving style; when the quantized driving style conclusion value is [3, 4], it corresponds to a moderate driving style; and when the quantized driving style conclusion value is [5, 6], it corresponds to an aggressive driving style. All second driving style rules constitute the second driving style rule base.

[0095] When the environment changes, the quantized range of the driving style conclusion of the first driving style rule can be modified or the quantized result can be adjusted, for example, adjusting the conservative quantization range from [1, 3] to [1, 2], or adjusting the quantization value from 3 to 4. These adjustments to the quantization range or quantization value can be achieved by adding or subtracting a coefficient. The size of the coefficient corresponds to the degree of environmental change. The larger the coefficient, the more likely the corresponding driving style fuzzy inference system is to identify the current driving data as an aggressive driving style.

[0096] Continuing with the above example, assume that the second driving style rule base includes:

[0097] Second driving style rule 1: The speed is very slow AND the acceleration is small AND the negative acceleration is small AND the accelerator pedal opening variance is small, the driving style is conservative;

[0098] Second driving style rule 2: medium speed AND medium acceleration AND medium negative acceleration AND medium accelerator pedal opening variance, the driving style is normal;

[0099] Second driving style rule 3: The speed is very fast AND the acceleration is large AND the negative acceleration is large AND the accelerator pedal opening variance is large, the driving style is aggressive;

[0100] Second driving style rule 4: slow speed AND small acceleration AND large negative acceleration AND small accelerator pedal opening variance, the driving style is conservative;

[0101] …….

[0102] For the current driving data of the current vehicle obtained above, that is, Figure 2 After the driving information is input into the driving style fuzzy recognition system, the current driving data can be preprocessed. This preprocessing can include data cleaning and denoising to ensure the accuracy of the data input into the driving style fuzzy recognition system, thereby improving the accuracy of the driving style recognition of the driver of the current vehicle. The preprocessed current driving data can then be calculated and analyzed to obtain parameter values ​​for multiple input variables of the current vehicle. These parameter values ​​can include parameter values ​​for speed, acceleration, negative acceleration, and accelerator pedal opening variance.

[0103] For any second driving style rule in the second driving style rule library, a matching calculation is performed between the parameter values ​​of multiple input variables and the second driving style rule, i.e., fuzzy reasoning is performed to obtain the matching rule strength between the second driving style rule and the current vehicle. The matching calculation can be performed by comparing the parameter values ​​of the input variables with the preconditions of the second driving style rule to calculate the matching rule strength between the second driving style rule and the current vehicle. For example, the matching rule strength between the parameter values ​​of the multiple input variables and the second driving style rule can be calculated using a variable fuzzy set membership function to obtain the matching rule strength between the second driving style rule and the current vehicle. The matching rule strength can be used to indicate the matching strength between the second driving style rule and the current vehicle. A greater matching rule strength indicates a greater matching strength between the second driving style rule and the current vehicle. Thus, the matching rule strengths of all second driving style rules and the current vehicle are calculated.

[0104] Based on the matching strengths of all second driving style rules with the current vehicle, the second driving style rule that best matches the current driving data of the current vehicle is selected. Based on the driving style conclusion of the second driving style rule that best matches the current driving data of the current vehicle, the current driving style of the driver of the current vehicle is determined. The current driving style is one of conservative, moderate, and aggressive.

[0105] The MEC (Mechanical Edge Computing) unit outputs the current driving style, which is then ultimately transmitted back to the vehicle terminal via V2X communication. The vehicle terminal's visual interface displays the current driving style. The vehicle terminal can also predict potential dangers based on the current driving style. When such dangers are predicted, it issues a warning to alert the driver to the impending danger, allowing them to take timely action to avoid it.

[0106] The identified current driving style or the degree of affiliation of the current driving style of the driver of the current vehicle is output to an intelligent driving assistance system, a vehicle insurance assessment agency, a driving behavior research agency, or a traffic safety warning system, etc., which can provide real-time or time-based data analysis support.

[0107] It is understood that the technical solution provided in this embodiment obtains the current driving data and current scene information of the current vehicle, and adjusts the first driving style rule base in the driving style fuzzy recognition system according to the current scene information to obtain a second driving style rule base. The current driving data is matched with each second driving style rule to obtain the matching rule strength of each second driving style rule with the current vehicle. Based on the matching rule strength of all the second driving style rules with the current vehicle, the current driving style of the driver of the current vehicle is determined. Since the first driving style rule base is obtained by simplifying the pre-built initial driving style rule base, and the second driving style rule base is obtained based on the current The second driving style rule base is obtained by adjusting the first driving style rule base based on the scene information, so that the second driving style rule base is also a simplified rule base. When identifying the driver's current driving style based on the second driving style rule base, time for matching and calculating with the second driving style rules in the second driving style rule base can be saved, thereby improving the efficiency of identifying the driver's current driving style. In addition, the second driving style rule base is a rule adjusted based on the current scene information, which can adapt to the differences in driving behavior in the current scene, thereby improving the accuracy of driving style identification of the current vehicle in the current scene. Therefore, the technical problems of low driver driving style identification accuracy and insufficient dynamic adaptability in the prior art are effectively solved.

[0108] In one possible implementation, the driving style fuzzy recognition system includes variable fuzzy set membership functions of multiple input variables, and the second driving style rules include preconditions for each input variable. In step S13, current driving data is matched with each second driving style rule to obtain a matching rule strength between each second driving style rule and the current vehicle, including:

[0109] Step S131 , extracting characteristic values ​​from the current driving data of the current vehicle to obtain parameter values ​​of multiple input variables of the current vehicle.

[0110] Step S132 : calculating the variable fuzzy set membership of each input variable based on the parameter values ​​of the multiple input variables according to the variable fuzzy set membership functions of the multiple input variables in the driving style fuzzy recognition system.

[0111] Step S133 : For each second driving style rule, the matching rule strength between the second driving style rule and the current vehicle is determined based on the preconditions of all input variables of the second driving style rule and the variable fuzzy set membership of each input variable of the current vehicle.

[0112] Specifically, the driving style fuzzy recognition system includes variable fuzzy set membership functions for multiple input variables. Each variable fuzzy set membership function represents the degree to which the input variable belongs to a fuzzy set. These variable fuzzy set membership functions are designed based on the data distribution characteristics of the input variables and the actual driving behavior. They reflect the degree to which the parameter value of the input variable belongs to a fuzzy set.

[0113] For example, when the input variable is speed, its variable fuzzy set membership function (speed membership function) is as follows: Figure 3 As shown, Figure 3 This is a schematic diagram of the structure of a variable fuzzy set membership function for speed provided by one embodiment of the present application, wherein the horizontal axis represents the domain of speed, and the vertical axis represents the degree to which the speed belongs to a certain fuzzy set. When the membership degree is 1, it means that it completely belongs to the fuzzy set; when the membership degree is 0, it means that it does not belong to the fuzzy set at all; when the membership degree is between (0, 1), it means that it partially belongs to the fuzzy set. The variable fuzzy set membership functions of other input variables are similar to the variable fuzzy set membership function of speed, indicating the degree to which the input variable belongs to a certain fuzzy set.

[0114] It should be noted that each second driving style rule includes preconditions of various input variables and a driving style conclusion.

[0115] Specifically, the current vehicle's driving data is obtained and, after preprocessing such as data cleaning and denoising, feature extraction is performed to extract the current vehicle's speed, acceleration, negative acceleration, and accelerator pedal position variance over a certain historical time period. Subsequently, parameter values ​​for multiple input variables of the current vehicle are calculated based on the current vehicle's speed, acceleration, negative acceleration, and accelerator pedal position variance over this time period. The speed parameter value can be the average speed over a period of time; an aggressive driving style may have a higher speed parameter value, while a conservative driving style may have a lower speed parameter value. The acceleration parameter value can be the average positive acceleration over a period of time, the negative acceleration parameter value can be the average negative acceleration over a period of time, and the accelerator pedal position variance parameter value can be the accelerator pedal position variance over a period of time. The accelerator pedal position variance indicates the degree to which the accelerator pedal position deviates from its average value over a period of time. Drivers with a large accelerator pedal position variance may have an aggressive driving style, frequently and drastically adjusting the accelerator pedal; whereas drivers with a small accelerator pedal position variance have a more stable driving style.

[0116] The driving style fuzzy recognition system includes variable fuzzy set membership functions of multiple input variables. The variable fuzzy set membership function can be a trapezoidal membership function or a triangular membership function. Figure 3As shown, the variable fuzzy set membership function for speed uses a triangular function to describe speed. The variable fuzzy set for speed is represented as {very slow, slow, medium, fast, very fast}, with a domain of [0, 120] km / h. Acceleration, negative acceleration, and accelerator pedal opening variance also each have corresponding variable fuzzy set membership functions.

[0117] According to the parameter values ​​of the input variables such as speed, acceleration, negative acceleration and accelerator pedal opening variance obtained above, fuzzification is performed through the variable fuzzy set membership function of the input variables such as speed, acceleration, negative acceleration and accelerator pedal opening variance, and the variable fuzzy set membership of the input variables such as speed, acceleration, negative acceleration and accelerator pedal opening variance of the current vehicle is calculated.

[0118] In an optional example, for the calculated input variable speed parameter value of 20km / h, according to Figure 3 The variable fuzzy set membership function for speed shown is fuzzified and its membership in each fuzzy set for speed is calculated: its membership in the very slow speed fuzzy set is μ1: μ1 = 0.2; its membership in the slow speed fuzzy set is μ2: μ2 = 0.8; its membership in the medium speed fuzzy set is μ3: μ3 = 0; its membership in the fast speed fuzzy set is μ4: μ4 = 0.2; and its membership in the very fast speed fuzzy set is μ5: μ5 = 0. Therefore, the variable fuzzy set membership for the current vehicle's input variable speed is [0.2, 0.8, 0, 0, 0].

[0119] In the same way, the variable fuzzy set membership of each input variable of the current vehicle (acceleration, negative acceleration and accelerator pedal opening variance) is calculated.

[0120] In an optional example, assume that the variable fuzzy set membership of the input variable speed of the current vehicle is [0, 0.5, 0.5, 0, 0], the variable fuzzy set membership of the acceleration of the current vehicle is: [0.3, 0.7, 0], the variable fuzzy set membership of the negative acceleration is: [0.8, 0.2, 0], and the variable fuzzy set membership of the accelerator pedal opening variance is: [0, 0.6, 0.4].

[0121] For each second driving style rule in the adjusted second driving style rule base, the preconditions of all input variables of the second driving style rule and the variable fuzzy set membership of all input variables of the current vehicle are compared and matched. The matching strength between the second driving style rule and the current vehicle is then determined. In this way, the matching strength between each second driving style rule and the current vehicle is calculated.

[0122] The matching rule strength between the second driving style rule and the current vehicle is a numerical value, either an activation strength or a maximum membership, indicating the degree of matching between the second driving style rule and the current vehicle. A higher matching rule strength indicates a higher degree of matching between the second driving style rule and the current vehicle, and a greater probability that the two vehicles share the same driving style conclusion.

[0123] In one possible implementation, in step S133, for each second driving style rule, the matching rule strength between the second driving style rule and the current vehicle is determined based on the preconditions of all input variables of the second driving style rule and the variable fuzzy set membership of each input variable of the current vehicle. This includes: for each second driving style rule, determining the conditional membership of each input variable of the second driving style rule based on the preconditions of all input variables of the second driving style rule and the variable fuzzy set membership of each input variable of the current vehicle; and determining the matching rule strength between the second driving style rule and the current vehicle based on the conditional membership of all input variables of the second driving style rule.

[0124] Specifically, for any second driving style rule, the conditional membership of each input variable of the second driving style rule is determined based on the preconditions of all input variables of the second driving style rule (including the preconditions of speed, acceleration, negative acceleration, and accelerator pedal opening variance), as well as the variable fuzzy set membership of each input variable of the current vehicle (including the variable fuzzy set membership of speed, acceleration, negative acceleration, and accelerator pedal opening variance). After comparing their sizes, the conditional membership of each input variable of the second driving style rule is determined.

[0125] Continuing with the above example, the prerequisites for the second driving style rule 1 are: very slow speed AND small acceleration AND small negative acceleration AND small accelerator pedal opening variance; after comparing with the variable fuzzy set membership of each input variable of the current vehicle, the conditional membership of each input variable of the second driving style rule 1 is calculated to be: [0, 0.3, 0.8, 0].

[0126] The prerequisites for the second driving style rule 2 are: medium speed AND medium acceleration AND medium negative acceleration AND medium accelerator pedal opening variance. After comparing with the variable fuzzy set membership of each input variable of the current vehicle, the conditional membership of each input variable of the second driving style rule 2 is calculated to be: [0.5, 0.7, 0.2, 0.6].

[0127] The prerequisites for the second driving style rule 3 are: very fast speed AND large acceleration AND large negative acceleration AND large variance of the accelerator pedal opening. After comparing the variable fuzzy set membership of each input variable of the current vehicle, the conditional membership of each input variable of the second driving style rule 3 is calculated to be: [0, 0, 0, 0.4];

[0128] The prerequisites for the second driving style rule 4 are: slower speed AND smaller acceleration AND larger negative acceleration AND smaller variance of the accelerator pedal opening. After comparing with the variable fuzzy set membership of each input variable of the current vehicle, the conditional membership of each input variable of the second driving style rule 4 is calculated to be: [0.5, 0.3, 0, 0].

[0129] In this way, the conditional membership of each input variable of all second driving style rules is calculated.

[0130] The minimum value method can be used to compare the conditional memberships of all input variables of any second driving style rule, and select the value with the minimum conditional membership as the matching rule strength between the second driving style rule and the current vehicle.

[0131] Continuing with the above example, the conditional membership of the input variables for the second driving style rule 1 is [0, 0.3, 0.8, 0]. The minimum value, 0, is selected as the matching rule strength between the second driving style rule 1 and the current vehicle. Following the same method, the matching rule strength between the second driving style rule 2 and the current vehicle is 0.2, the matching rule strength between the second driving style rule 3 and the current vehicle is 0, and the matching rule strength between the second driving style rule 4 and the current vehicle is 0.

[0132] In one possible implementation, the second driving style rule includes a driving style conclusion. In step S14, determining the current driving style of the driver of the current vehicle based on the matching rule strengths of all the second driving style rules with the current vehicle includes:

[0133] Step S141: Aggregating the driving style conclusions of all second driving style rules based on the matching rule strength between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle;

[0134] Step S142, performing defuzzification processing on the driving style fuzzy set of the current vehicle to obtain the driving style output value of the current vehicle;

[0135] Step S143 : determining the current driving style of the driver driving the current vehicle based on the driving style output value of the current vehicle.

[0136] Specifically, the driving style fuzzy recognition system includes multiple output fuzzy set membership functions of driving styles, such as an output fuzzy set membership function of a conservative driving style, an output fuzzy set membership function of a moderate driving style, and an output fuzzy set membership function of an aggressive driving style.

[0137] Using the Mamdani inference method, based on the matching strength of each second driving style rule with the current vehicle, the output fuzzy set membership functions of multiple driving styles, and the driving style conclusions of each second driving style, all the driving style conclusions of the second driving style rules are aggregated to obtain the driving style fuzzy set of the current vehicle. The driving style fuzzy set of the current vehicle represents the effective membership function of the current vehicle's driving behavior to each driving style. For example, the output fuzzy set membership function of each driving style is a triangular function similar to the variable fuzzy set membership function of speed, with the abscissa representing the driving style output value and the ordinate representing the membership degree of the driving style. The driving style fuzzy set of the current vehicle obtained after aggregation can be obtained by merging the output fuzzy set membership functions of each driving style in the triangular function, after cutting or scaling them according to the matching strength of the second driving style rule with the current vehicle. The driving style fuzzy set of the current vehicle can be a trapezoidal driving style fuzzy set obtained by removing the top of the triangular function.

[0138] Afterwards, the combined driving style fuzzy set of the current vehicle is defuzzified, i.e., defuzzified. The defuzzification (defuzzification) method may be a centroid method, a maximum membership method, etc., to obtain the driving style output value of the current vehicle.

[0139] For the current vehicle's driving style fuzzy set, using the centroid method as an example, the x-axis value corresponding to the centroid of the area covered by the current vehicle's driving style fuzzy set is calculated, which is the current vehicle's driving style output value. The specific calculation method is:

[0140] According to the following formula, the current vehicle's driving style output value y is calculated on the continuous domain X (for example, the domain after driving style quantization is 1-9): * :

[0141]

[0142] where y * Output value of the current vehicle’s driving style, x i is the i-th element in the domain X, μ i is x i The corresponding membership degree on the driving style fuzzy set of the current vehicle, N is the number of selected elements.

[0143] Based on the current vehicle's driving style output value, a threshold method or a direct mapping method is used to map it to a specific driving style category. The mapped specific driving style category is then used as the current driving style of the driver of the current vehicle. For example, suppose the calculated driving style output value of the current vehicle is 77.06, and the driving style thresholds are: conservative: [0, 30]; normal: [30, 60]; aggressive: [60, 90]. The clarity value of 77.06 falls within the "aggressive" range, so the current driving style of the driver of the current vehicle is identified as aggressive.

[0144] It is also possible to calculate the membership of the driving style output value of the current vehicle to the three driving styles (conservative, moderate, and aggressive) using the output fuzzy set membership functions of multiple driving styles pre-established in the driving style fuzzy recognition system, and then compare the sizes of the membership of these three driving styles. The driving style corresponding to the largest membership is selected as the current driving style of the driver of the current vehicle.

[0145] In one possible implementation, the driving style fuzzy recognition system includes multiple output fuzzy set membership functions of driving styles. Step S141 aggregates the driving style conclusions of all second driving style rules based on the matching rule strengths between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle. The system includes: determining an activated driving style rule from a second driving style rule library based on the matching rule strengths between each second driving style rule and the current vehicle; segmenting the output fuzzy set membership functions corresponding to the driving style conclusions of each activated driving style rule based on the matching rule strengths between each activated driving style rule and the current vehicle to obtain a valid output fuzzy set membership function corresponding to the driving style conclusions of each activated driving style rule; and merging all activated driving style rules based on the valid output fuzzy set membership functions corresponding to the driving style conclusions of all activated driving style rules to obtain a driving style fuzzy set of the current vehicle.

[0146] Specifically, the driving style fuzzy recognition system includes multiple driving style output fuzzy set membership functions: conservative driving style output fuzzy set membership functions, moderate driving style output fuzzy set membership functions, and aggressive driving style output fuzzy set membership functions. When the driving style conclusions of all second driving style rules are aggregated to obtain the current vehicle's driving style fuzzy set:

[0147] Each second driving style rule is associated with a matching rule strength with the current vehicle. For any second driving style rule, its matching rule strength with the current vehicle is determined to be greater than 0. If the matching rule strength is greater than 0, the corresponding output fuzzy set is valid and is determined as the active driving style rule. In this way, the active driving style rule can be determined from all second driving style rules.

[0148] For any activated driving style rule, its corresponding driving style conclusion is determined. Then, based on the matching strength between the activated driving style rule and the current vehicle, the output fuzzy set membership function corresponding to the corresponding driving style conclusion is clipped or scaled (cropped or scaled) to obtain the valid output fuzzy set membership function corresponding to the driving style conclusion of the activated driving style rule. In this way, the valid output fuzzy set membership function corresponding to the driving style conclusion of each activated driving style rule can be calculated.

[0149] In an alternative example, for a driving style conclusion that activates a driving style rule (e.g., "Driving Style = Aggressive"), its membership function is typically a fuzzy set (e.g., a triangular or trapezoidal function). The output fuzzy set membership function corresponding to this driving conclusion is truncated by the matching rule strength. For each point in the output fuzzy set corresponding to this driving conclusion, its new membership is min(original membership, matching rule strength). For example, the output fuzzy set membership function for the "Aggressive" driving style conclusion is a triangular function with a domain range of [70, 80, 90], assuming a matching rule strength of 0.6. For each point in the output fuzzy set corresponding to "Aggressive," its new membership is min(original membership, 0.6). For example, a point with an original membership of 0.8 becomes min(0.8, 0.6) = 0.6 after truncation, meaning its effective output fuzzy set membership function is the truncated output fuzzy set membership function.

[0150] According to the effective output fuzzy set membership functions corresponding to the driving style conclusions of all activated driving style rules, the maximum value of the membership degrees contributed by different activated driving style rules to the same output fuzzy set can be taken by taking the maximum value method, and all activated driving style rules can be merged to obtain the driving style fuzzy set of the current vehicle.

[0151] For example, if the driving style conclusion for activating driving style rule 1 is "aggressive," the membership function at a certain point is 0.6. If the driving style conclusion for activating driving style rule 2 is also "aggressive," the membership function at the same point is 0.3. Therefore, for the combined driving style fuzzy set for the current vehicle, the membership function at that point in the driving style fuzzy set with the "aggressive" conclusion is max(0.6,0.3)=0.6.

[0152] In one possible implementation, in step S12, the first driving style rule base in the driving style fuzzy recognition system is adjusted according to the current scene information to obtain a second driving style rule base. Prior to this, the method further includes:

[0153] Step S101: Acquire a historical driving data set and a driving style corresponding to each piece of historical driving data in the historical driving data set.

[0154] Step S102 : establishing an initial driving style rule base based on the historical driving data set and the driving style corresponding to each piece of historical driving data in the historical driving data set; wherein the initial driving style rule base includes a plurality of initial driving style rules, each of which includes a prerequisite for an input variable.

[0155] Step S103 : clustering and simplifying all the initial driving style rules in the initial driving style rule base according to the preconditions of the input variables of all the initial driving style rules to obtain a first driving style rule base.

[0156] Specifically, before adjusting the first driving style rule base in the driving style fuzzy recognition system according to the current scene information to obtain the second driving style rule base, the method further includes:

[0157] A historical driving data set is obtained. The historical driving data set includes multiple historical driving data items, each of which includes at least one or more of speed, acceleration, deceleration, and accelerator pedal opening variance. Furthermore, each historical driving data item corresponds to a driving style. Driving styles can be conservative, moderate, or aggressive. The historical driving data can be acquired through preprocessing, analysis, and calculation after collection using sensors or other equipment. Driving style annotation can be achieved using a big data model or through manual annotation by experts.

[0158] Based on the historical driving data set, data features are extracted and calculated for pre-set input variables to obtain the historical values ​​of the input variables corresponding to each historical driving data item. Initial driving style rules are established based on the historical values ​​of the input variables corresponding to each historical driving data item and the driving style associated with each historical driving data item. The driving style conclusions of the initial driving rules are determined by the driving style associated with each historical driving data item. All initial driving rules constitute the initial driving rule library.

[0159] The number of initial driving style rules in the initial driving style rule base depends on the number of input variables and the number of variable fuzzy sets in the variable fuzzy set of each input variable.

[0160] Assume that the driving style fuzzy recognition system has n input variables, and the i-th input variable x i Divided into mi variable fuzzy sets, then the number p of initial driving style rules in the initial driving style rule base is the product of the number of variable fuzzy sets of each input variable, that is, p = m1×m2…×m n .

[0161] Continuing with the above example, assume there are four input variables: speed, acceleration, deceleration, and accelerator pedal opening variance. The number of variable fuzzy sets for speed is five, and the number of variable fuzzy sets for acceleration, deceleration, and accelerator pedal opening variance is three, three, and three, respectively. This allows us to construct an initial driving style rule base containing 135 initial driving style rules (calculated using the formula 5 × 3 × 3 × 3).

[0162] Because each input variable influences the driving style conclusion, and some of the 135 initial driving style rules are similar, their input variables differ but their driving style conclusions are the same. If the initial driving style rule base contains a large number of similar or redundant initial driving style rules, without simplification, subsequent identification of the current vehicle's driving style conclusion requires comparing all 135 initial driving style rules, which is time-consuming and inefficient. To improve the efficiency of driving style identification, all initial driving style rules can be clustered and simplified, and similar rules can be merged. Multiple similar initial driving style rules can share a common driving style conclusion, forming a first driving style rule. All of these first driving style rules form the first driving style rule base.

[0163] It can be understood that, based on the simplified first driving style rule base, the recognition efficiency can be improved when the driver's driving style is identified, and when the driving style conclusion of the first driving style rule needs to be manually labeled, the labeling efficiency can be improved, thereby improving the efficiency of establishing the first driving style rule base.

[0164] For example, each initial driving style rule contains premise items for four input variables (speed, acceleration, negative acceleration, and accelerator pedal opening variance), where the premise conditions can be a variable fuzzy set in a variable fuzzy set of the four input variables, or the premise conditions can be quantified as the numerical values ​​of the four input variables. These four numerical values ​​can be used to represent the premise conditions of the four input variables of the initial driving style rule, and to represent the four dimensions of the initial driving style rule.

[0165] Based on the four input variables of the initial driving style rules, all initial driving style rules are clustered and simplified. The 135 initial driving style rules in the initial driving style rule base are formed into driving style rule clusters. Assuming 95 driving style rule clusters are formed, the initial driving style rules in the same cluster share the same driving style conclusion. All initial driving style rules are simplified into 95 first driving style rules, while the remaining 40 initial driving style rules use the driving style conclusion corresponding to their respective driving style rule clusters. By reducing the number of initial driving style rules to 70% to 80% of the original number, the rule base structure is simplified, driving style recognition is improved, and the efficiency of establishing the first driving style rule base is increased.

[0166] In one possible implementation, each initial driving style rule further includes a driving style conclusion; the driving style fuzzy recognition system includes variable fuzzy set membership functions of multiple input variables. In step S103, all initial driving style rules in the initial driving style rule base are clustered and simplified based on the preconditions of the input variables of all initial driving style rules to obtain a first driving style rule base, including:

[0167] Based on variable fuzzy set membership functions of multiple input variables in a driving style fuzzy recognition system, the premise conditions of the input variables of each initial driving style rule are quantified to obtain a premise quantization vector for each initial driving style rule; based on the premise quantization vectors of all initial driving style rules, all initial driving style rules are clustered to obtain multiple driving style rule clusters; for each driving style rule cluster, the centroid of the driving style rule cluster is determined based on the premise quantization vectors of all initial driving style rules in the driving style rule cluster; for each driving style rule cluster, the first driving style rule corresponding to the driving style rule cluster is determined based on the centroid of the driving style rule cluster and the driving style conclusions of all initial driving style rules in the driving style rule cluster; and based on the first driving style rules corresponding to all driving style rule clusters, a first driving style rule base is determined.

[0168] Specifically, the driving style fuzzy recognition system can pre-establish variable fuzzy set membership functions for multiple input variables. Each initial driving style rule includes preconditions for the input variables and a driving style conclusion. The preconditions for the input variables can include preconditions for each of the four input variables. For example, the preconditions for the input variables of an initial driving style rule may include preconditions for speed, acceleration, negative acceleration, and throttle pedal opening variance.

[0169] Clustering and simplifying all the initial driving style rules in the initial driving style rule base can be achieved through the following methods:

[0170] First, for any initial driving style rule, the preconditions for each input variable of the initial driving style rule can be quantified by calculating the centroid of the variable fuzzy set membership function of each input variable, thereby obtaining the precondition quantized value of the input variable of the initial driving style rule. After sequentially calculating the precondition quantized values ​​for each input variable of the initial driving style rule, all of these precondition quantized values ​​form the precondition quantized vector of the initial driving style rule.

[0171] For example, by calculating the centroid of the fuzzy set membership function of the speed and acceleration variables, the fuzzy descriptions of "faster speed" and "greater acceleration" in the premise conditions of a certain initial driving style rule can be converted into specific numerical values ​​as the premise quantification value.

[0172] Following the same method, the premise quantization vectors of all initial driving style rules are calculated.

[0173] Based on the premise quantization vectors of all initial driving style rules, a clustering algorithm (such as K-means clustering or hierarchical clustering) is used to cluster all initial driving style rules into multiple driving style rule clusters. The initial driving style rules in each driving style rule cluster are considered similar rules, and two initial driving style rules that do not belong to the same driving style rule cluster are considered to have significant differences.

[0174] It should be noted that the number of driving style rule clusters can be set as needed, for example, according to the number of driving style conclusions, or according to actual scenarios, which is not specifically limited in this embodiment.

[0175] For any driving style rule cluster, a first driving style rule can be generated based on all the initial driving style rules it contains. The specific method is as follows:

[0176] For this driving style rule cluster, the centroid of the driving style rule cluster is calculated using the following formula based on the premise quantization vectors of all the initial driving style rules in the driving style rule cluster. The dimension of the centroid vector of the driving style rule cluster is related to the number of input variables. Each dimension of the centroid vector corresponds to an input variable. That is, the coordinates of each dimension in the centroid vector are related to the same input variable for all the initial driving style rules in the driving style rule cluster. For the coordinate v of any dimension in the centroid vector:

[0177]

[0178] Where M is the number of initial driving style rules in the driving style rule cluster; hj represents the premise quantized value of the input variable corresponding to the j-th initial driving style rule; w j Represents the membership degree corresponding to the premise quantized value of the corresponding input variable in the j-th initial driving style rule.

[0179] In this way, the coordinates of each dimension of the centroid vector of the driving style rule cluster can be calculated, and the coordinates of each dimension constitute the centroid vector of the driving style rule cluster.

[0180] In the same way, the centroid vector of each driving style rule cluster is calculated.

[0181] For each driving style rule cluster, based on the driving style conclusions of all initial driving style rules in the driving style rule cluster, the initial driving style rules with the same driving style conclusion can be grouped into a conclusion group. The number of initial driving style rules in each conclusion group is counted, and the conclusion group with the largest number of initial driving style rules is selected. The driving style conclusion corresponding to the conclusion group with the largest number of initial driving style rules is used as the driving style conclusion of the driving style rule cluster.

[0182] A first driving style rule is generated for each driving style rule cluster. The driving style conclusion of each driving style rule cluster is used as the driving style conclusion of the corresponding first driving style rule. A first driving style rule is formed based on the centroid vector of the driving style rule cluster and the driving style conclusion.

[0183] In this way, the first driving style rule corresponding to each driving style rule cluster is obtained, and all the first driving style rules constitute a first driving style rule library.

[0184] It is understood that the technical solution provided in this embodiment obtains the current driving data and current scene information of the current vehicle, and adjusts the first driving style rule base in the driving style fuzzy recognition system according to the current scene information to obtain a second driving style rule base. The current driving data is matched with each second driving style rule to obtain the matching rule strength of each second driving style rule with the current vehicle. Based on the matching rule strength of all the second driving style rules with the current vehicle, the current driving style of the driver of the current vehicle is determined. Since the first driving style rule base is obtained by simplifying the pre-built initial driving style rule base, and the second driving style rule base is obtained based on the current The second driving style rule base is obtained by adjusting the first driving style rule base based on the scene information, so that the second driving style rule base is also a simplified rule base. When identifying the driver's current driving style based on the second driving style rule base, time for matching and calculating with the second driving style rules in the second driving style rule base can be saved, thereby improving the efficiency of identifying the driver's current driving style. In addition, the second driving style rule base is a rule adjusted based on the current scene information, which can adapt to the differences in driving behavior in the current scene, thereby improving the accuracy of driving style identification of the current vehicle in the current scene. Therefore, the technical problems of low driver driving style identification accuracy and insufficient dynamic adaptability in the prior art are effectively solved.

[0185] The present application also provides a driving style recognition device, referring to Figure 4 , Figure 4 FIG. 6 is a schematic diagram of a driving style recognition device provided in an embodiment of the present application. The driving style recognition device 6 includes:

[0186] The data acquisition module 61 is used to obtain the current driving data and current scene information of the current vehicle;

[0187] a rule base establishing module 62 for adjusting a first driving style rule base in the driving style fuzzy recognition system based on current scene information to obtain a second driving style rule base; wherein the first driving style rule base is obtained by simplifying a pre-established initial driving style rule base; and the second driving style rule base includes a plurality of second driving style rules;

[0188] a matching calculation module 63 for performing matching calculations on the current driving data and each second driving style rule to obtain a matching rule strength between each second driving style rule and the current vehicle;

[0189] The driving style recognition module 64 is configured to determine a current driving style of a driver driving the current vehicle based on the matching rule strengths of all second driving style rules with the current vehicle.

[0190] In one possible implementation, the driving style fuzzy recognition system includes variable fuzzy set membership functions of a plurality of input variables, and the second driving style rule includes preconditions for each input variable;

[0191] The matching calculation module 63 includes:

[0192] The feature value extraction unit 631 is used to extract feature values ​​from the current driving data of the current vehicle to obtain parameter values ​​of multiple input variables of the current vehicle;

[0193] a membership calculation unit 632 for calculating the variable fuzzy set membership of each input variable based on the parameter values ​​of the multiple input variables according to the variable fuzzy set membership functions of the multiple input variables in the driving style fuzzy recognition system;

[0194] The rule matching unit 633 is configured to determine, for each second driving style rule, the strength of the matching between the second driving style rule and the current vehicle based on the preconditions of all input variables of the second driving style rule and the variable fuzzy set membership of each input variable of the current vehicle.

[0195] In one possible implementation, the rule matching unit 633 is specifically configured to determine, for each second driving style rule, the conditional membership of each input variable of the second driving style rule based on the preconditions of all input variables of the second driving style rule and the variable fuzzy set membership of each input variable of the current vehicle; and determine the strength of the matching rule between the second driving style rule and the current vehicle based on the conditional membership of all input variables of the second driving style rule.

[0196] In one possible implementation, the second driving style rule includes a driving style conclusion;

[0197] The driving style recognition module 64 includes:

[0198] a driving style rule aggregation unit 641 for aggregating the driving style conclusions of all second driving style rules based on the matching rule strength between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle;

[0199] a defuzzification unit 642 for performing defuzzification processing on the driving style fuzzy set of the current vehicle to obtain an output value of the driving style of the current vehicle;

[0200] The driving style determining unit 643 is configured to determine the current driving style of the driver driving the current vehicle based on the driving style output value of the current vehicle.

[0201] In one possible implementation, the driving style fuzzy recognition system includes a plurality of output fuzzy set membership functions of driving styles;

[0202] The driving style rule aggregation unit 641 is specifically configured to determine an activated driving style rule from the second driving style rule library based on the matching rule strength between each second driving style rule and the current vehicle; segment the output fuzzy set membership functions corresponding to the driving style conclusions of each activated driving style rule based on the matching rule strength between each activated driving style rule and the current vehicle to obtain a valid output fuzzy set membership function corresponding to the driving style conclusions of each activated driving style rule; and merge all activated driving style rules based on the valid output fuzzy set membership functions corresponding to the driving style conclusions of all activated driving style rules to obtain a driving style fuzzy set for the current vehicle.

[0203] In one possible implementation, the first driving style rule base in the driving style fuzzy recognition system is adjusted according to the current scene information to obtain a second driving style rule base. Previously, the data acquisition module 61 is further configured to acquire a historical driving data set and a driving style corresponding to each piece of historical driving data in the historical driving data set.

[0204] The rule base establishing module 62 is further configured to establish an initial driving style rule base based on the historical driving data set and the driving style corresponding to each piece of historical driving data in the historical driving data set; wherein the initial driving style rule base includes a plurality of initial driving style rules, each of which includes a precondition for an input variable;

[0205] The rule base establishing module 62 is further configured to cluster and simplify all the initial driving style rules in the initial driving style rule base according to the preconditions of the input variables of all the initial driving style rules to obtain a first driving style rule base.

[0206] In a possible implementation, each initial driving style rule further includes a driving style conclusion;

[0207] The driving style fuzzy recognition system includes variable fuzzy set membership functions of multiple input variables;

[0208] The rule base establishing module 62 is further configured to quantify the preconditions of the input variables of each initial driving style rule based on the variable fuzzy set membership functions of multiple input variables in the driving style fuzzy recognition system to obtain a precondition quantization vector for each initial driving style rule; cluster all initial driving style rules based on the precondition quantization vectors of all initial driving style rules to obtain multiple driving style rule clusters; determine the centroid of each driving style rule cluster based on the precondition quantization vectors of all initial driving style rules in the driving style rule cluster; determine a first driving style rule corresponding to each driving style rule cluster based on the centroid of the driving style rule cluster and the driving style conclusions of all initial driving style rules in the driving style rule cluster; and determine a first driving style rule base based on the first driving style rules corresponding to all driving style rule clusters.

[0209] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of this application. Their specific functions and technical effects can be found in the method embodiment section and will not be repeated here.

[0210] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0211] An embodiment of the present application also provides an electronic device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps of any of the above-mentioned method embodiments when executing the computer program.

[0212] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned various method embodiments can be implemented.

[0213] An embodiment of the present application provides a computer program product. When the computer program product is run on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned various method embodiments when executing the computer program product.

[0214] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the process of the above-mentioned method embodiment by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can at least include: any entity or device capable of carrying computer program code to the camera / terminal device, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. For example, a USB flash drive, mobile hard drive, magnetic disk, or optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals or telecommunication signals.

[0215] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0216] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0217] In the embodiments provided in this application, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely illustrative. For example, the division of the modules or units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0218] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0219] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for identifying a driving style, characterized in that: include: Get the current driving data and current scene information of the current vehicle; adjusting a first driving style rule base in the driving style fuzzy recognition system based on the current scene information to obtain a second driving style rule base; wherein the first driving style rule base is obtained by simplifying a pre-built initial driving style rule base; and the second driving style rule base includes a plurality of second driving style rules; performing matching calculations on the current driving data and each of the second driving style rules to obtain a matching rule strength between each of the second driving style rules and the current vehicle; The current driving style of the driver driving the current vehicle is determined according to the matching rule strengths of all the second driving style rules with the current vehicle.

2. The driving style recognition method according to claim 1, wherein: The driving style fuzzy identification system includes variable fuzzy set membership functions of multiple input variables; The second driving style rule includes preconditions of each of the input variables; Performing matching calculations on the current driving data and each of the second driving style rules to obtain a matching rule strength between each of the second driving style rules and the current vehicle includes: Performing feature value extraction on the current driving data of the current vehicle to obtain parameter values ​​of a plurality of the input variables of the current vehicle; calculating, according to the variable fuzzy set membership functions of the plurality of input variables in the driving style fuzzy recognition system and based on the parameter values ​​of the plurality of input variables, the variable fuzzy set membership degree of each input variable; For each of the second driving style rules, the matching rule strength between the second driving style rule and the current vehicle is determined based on the preconditions of all the input variables of the second driving style rule and the variable fuzzy set membership of each of the input variables of the current vehicle.

3. The driving style recognition method according to claim 2, wherein: Determining, for each second driving style rule, the matching rule strength between the second driving style rule and the current vehicle based on preconditions of all the input variables of the second driving style rule and the variable fuzzy set membership of each input variable of the current vehicle, includes: For each of the second driving style rules, determining a conditional membership of each of the input variables of the second driving style rule based on preconditions of all of the input variables of the second driving style rule and the variable fuzzy set membership of each of the input variables of the current vehicle; The matching rule strength between the second driving style rule and the current vehicle is determined according to the conditional membership of all the input variables of the second driving style rule.

4. The driving style recognition method according to claim 1, wherein: The second driving style rule includes a driving style conclusion; The determining, based on the matching rule strengths of all the second driving style rules with the current vehicle, the current driving style of the driver of the current vehicle includes: aggregating the driving style conclusions of all the second driving style rules according to the matching rule strength between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle; performing a defuzzification process on the driving style fuzzy set of the current vehicle to obtain a driving style output value of the current vehicle; Based on the driving style output value of the current vehicle, a current driving style of a driver driving the current vehicle is determined.

5. The driving style recognition method according to claim 4, wherein: The driving style fuzzy recognition system includes a plurality of output fuzzy set membership functions of driving styles; The driving style conclusions of all the second driving style rules are aggregated according to the matching rule strength between each second driving style rule and the current vehicle to obtain a driving style fuzzy set of the current vehicle, including: determining an activated driving style rule from the second driving style rule library according to a matching rule strength between each of the second driving style rules and the current vehicle; cutting, based on the matching rule strength between each of the activated driving style rules and the current vehicle, the output fuzzy set membership function corresponding to the driving style conclusion of each of the activated driving style rules to obtain a valid output fuzzy set membership function corresponding to the driving style conclusion of each of the activated driving style rules; All the activated driving style rules are combined according to the valid output fuzzy set membership functions corresponding to the driving style conclusions of all the activated driving style rules to obtain the driving style fuzzy set of the current vehicle.

6. The driving style recognition method according to claim 1, wherein: Before the step of adjusting the first driving style rule base in the driving style fuzzy recognition system according to the current scene information to obtain the second driving style rule base, the method further includes: Acquire a historical driving data set and a driving style corresponding to each piece of historical driving data in the historical driving data set; establishing an initial driving style rule base based on the historical driving data set and the driving style corresponding to each piece of historical driving data in the historical driving data set; wherein the initial driving style rule base includes a plurality of initial driving style rules, each of the initial driving style rules including a precondition of an input variable; All the initial driving style rules in the initial driving style rule base are clustered and simplified according to the preconditions of the input variables of all the initial driving style rules to obtain a first driving style rule base.

7. The driving style recognition method according to claim 6, wherein: Each of the initial driving style rules further includes a driving style conclusion; The driving style fuzzy recognition system includes variable fuzzy set membership functions of a plurality of the input variables; The method further comprises: clustering and simplifying all the initial driving style rules in the initial driving style rule base according to the preconditions of the input variables of all the initial driving style rules to obtain a first driving style rule base, including: quantizing the premise conditions of the input variables of each of the initial driving style rules according to variable fuzzy set membership functions of the plurality of input variables in the driving style fuzzy recognition system to obtain premise quantization vectors of each of the initial driving style rules; clustering all the initial driving style rules according to the premise quantization vectors of all the initial driving style rules to obtain a plurality of driving style rule clusters; For each driving style rule cluster, determining a centroid of the driving style rule cluster according to the premise quantized vectors of all the initial driving style rules in the driving style rule cluster; For each driving style rule cluster, determining a first driving style rule corresponding to the driving style rule cluster according to the centroid of the driving style rule cluster and the driving style conclusions of all the initial driving style rules in the driving style rule cluster; The first driving style rule library is determined according to the first driving style rules corresponding to all the driving style rule clusters.

8. A driving style recognition device, characterized in that: include: A data acquisition module is used to obtain the current driving data and current scene information of the current vehicle; a rule base establishment module, configured to adjust a first driving style rule base in the driving style fuzzy recognition system based on the current scene information to obtain a second driving style rule base; wherein the first driving style rule base is obtained by simplifying a pre-established initial driving style rule base; and the second driving style rule base includes a plurality of second driving style rules; a matching calculation module, configured to perform matching calculations on the current driving data and each of the second driving style rules to obtain a matching rule strength between each of the second driving style rules and the current vehicle; The driving style recognition module is configured to determine a current driving style of a driver driving the current vehicle based on the matching rule strengths of all the second driving style rules with the current vehicle.

9. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer program product, characterized in that When the computer program product is run on an electronic device, the electronic device is enabled to execute the method according to any one of claims 1 to 7.

Images