Electric vehicle control system module and method based on artificial intelligence
By using an AI-based electric vehicle control method, the future heat generation power and required heat dissipation power of the battery are predicted, solving the problem of lag response in the electric vehicle battery thermal management system and improving battery safety and range.
Patent Information
- Application Number
- CN202511839384.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-08
- Publication Date
- 2026-03-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing electric vehicle battery thermal management systems cannot predict future battery heating trends, resulting in delayed heat dissipation response. This may lead to battery overheating or excessive heat dissipation, affecting driving range and safety. Furthermore, they do not fully consider individual differences in driving behavior and dynamic changes in driving paths.
An artificial intelligence-based electric vehicle control method is adopted. By acquiring historical user driving habit data, electric vehicle driving path data, and battery status data, a user driving behavior analysis model is established to predict the battery's heat generation power and required heat dissipation power at various future moments, and select an appropriate heat dissipation mode.
It achieves forward-looking and proactive battery thermal management, which can activate or adjust the heat dissipation mode in advance, improve the safety margin of battery operation and driving range, and avoid the lagging control of traditional passive response.
Smart Images
Figure CN121608652A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of data analysis, specifically an electric vehicle control system module and method based on artificial intelligence. Background Technology
[0002] With the global energy transition and increasing environmental awareness, electric vehicles are gradually becoming the mainstream mode of transportation due to their advantages such as zero emissions, low noise, and high energy efficiency. However, the driving range, battery safety, and lifespan of electric vehicles remain key factors restricting their large-scale popularization and user acceptance. As the core component of electric vehicles, the performance of the power battery is closely related to temperature. In existing technologies, electric vehicle battery thermal management systems mostly employ passive or simple active control strategies. For example, most systems only activate fans or liquid cooling pumps for heat dissipation when the battery temperature reaches a preset threshold. This control strategy has significant drawbacks: First, it cannot predict the future heating trend of the battery, resulting in a delayed heat dissipation response. This may lead to insufficient heat dissipation under extreme driving or harsh operating conditions, causing battery overheating, accelerating battery aging, and even creating safety hazards. Second, when the actual heat generation of the battery is not high, a fixed heat dissipation mode may also cause excessive heat dissipation, resulting in energy waste and consuming valuable battery power, thereby shortening the vehicle's driving range. In addition, some improved solutions attempt to introduce monitoring of single parameters such as vehicle speed and ambient temperature to adjust the heat dissipation strategy, but these methods still do not fully consider individual differences in driving behavior and dynamic changes in driving paths. Different drivers have different driving habits, some gentle and some aggressive. Similarly, different driving path conditions will greatly affect the battery load and heat generation. To solve the problems raised in this background, this application designs an electric vehicle control system module and method based on artificial intelligence. Summary of the Invention
[0003] To address the aforementioned technical shortcomings, this application proposes an artificial intelligence-based electric vehicle control system module and method.
[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: This application provides an artificial intelligence-based electric vehicle control method, which includes the following specific steps: S1. Obtain historical user driving habit data, electric vehicle driving route data, and electric vehicle battery status data; S2. Establish a user driving behavior analysis model. Input historical user driving habit data into the user driving behavior analysis model to obtain the user driving aggression index. S3. Based on the electric vehicle's driving path data and the user's driving aggression index, predict and analyze the battery's heat generation power at various future moments. S4. Based on the electric vehicle battery status data and the predicted analysis results of battery heat generation power at future times, predict and analyze the required heat dissipation power of the battery at future times. S5. Based on the predicted and analyzed results of the heat dissipation power required by the battery at each future time, select the appropriate heat dissipation mode for the electric vehicle battery at each future time.
[0005] It should be noted that, as a preferred technical solution for the artificial intelligence-based electric vehicle control method, the specific steps of S1 are as follows: S11. Obtain historical user driving habit data from the database. The historical user driving habit data consists of historical user driving segment data. S12. Obtain the path data of the electric vehicle through navigation data and bench tests. The path data of the electric vehicle includes road slope data, speed data and transmission efficiency data at future times when the electric vehicle is driving. S13. Obtain electric vehicle battery status data from the database, including real-time electric vehicle battery temperature data, historical temperature change rate data under the same driving conditions, total mass data of the electric vehicle battery, and overall average specific heat capacity data of the battery pack.
[0006] It should be noted that, as a preferred technical solution for an artificial intelligence-based electric vehicle control method, step S2 includes the following specific steps: S21. Collect driving segment data from different drivers, road conditions, and times to build a driving behavior database; S22. Extract features from driving segments in the driving behavior database to obtain feature vectors for each feature group of the driving segment; S23. Construct a user driving behavior analysis model, input historical user driving habit data into the user driving behavior analysis model, and obtain the user driving aggression index.
[0007] It should be noted that, as a preferred technical solution for the artificial intelligence-based electric vehicle control method, the specific steps of S21 are as follows: collecting driving segment data of different drivers, road conditions, and times, performing algorithmic labeling on each driving segment data, and constructing a driving behavior database; wherein, different drivers refer to drivers of different ages, genders, and driving experience; different road conditions refer to road conditions when driving in urban congestion, on highways, in suburbs, and on mountain roads; different times refer to daytime, nighttime, peak hours, and off-peak hours; driving segment data refers to time-series segment data recorded at fixed time intervals; the process of performing algorithmic labeling on each driving segment data is as follows: calculating the standard deviation of longitudinal acceleration within each data segment, setting a driving segment threshold, assigning a label to each driving segment according to the set driving segment threshold, and combining the driving segment datasets after labeling to form a driving behavior database.
[0008] It should be noted that, as a preferred technical solution for the electric vehicle control method based on artificial intelligence, the specific steps of S22 are as follows: extracting features from driving segments in the driving behavior database to obtain feature vectors for each feature group of the driving segment. The feature groups of the driving segment include accelerator pedal feature group, brake pedal feature group, steering wheel feature group, and vehicle dynamic feature group. The accelerator pedal feature group includes the average opening degree, the standard deviation of the opening degree, the maximum opening degree, the average value of the pedal change rate, and the number of rapid accelerations. The brake pedal feature group includes the total number of brake pedal presses divided by the driving segment duration, the number of emergency brakings, and the total braking time divided by the segment duration. The steering wheel feature group includes the standard deviation of the steering wheel angle, the average value of the steering wheel rotation speed, and the number of sharp turns. The vehicle dynamic feature group includes the standard deviation of the vehicle speed, the standard deviation of the longitudinal acceleration, and the maximum value of the lateral acceleration.
[0009] It should be noted that, as a preferred technical solution for the artificial intelligence-based electric vehicle control method, the specific steps of S23 are as follows: A user driving behavior analysis model is constructed using the random forest algorithm and the feature vectors of each feature group of the driving segment. Historical user driving habit data is input to obtain the user driving aggression index. The process of constructing the user driving behavior analysis model involves: randomly shuffling the feature vectors of each feature group of the driving segment, dividing them into 70% as a training set and 30% as a test set. The training set is input into the random forest algorithm, which generates multiple training subsets for decision tree training until the number of decision trees reaches the required level. The CART algorithm is then used for individual decision tree training. The decision tree is constructed using the Gini coefficient minimization criterion for feature selection. The smaller the Gini coefficient, the purer the information contained in the dataset. After the decision tree is constructed, no pruning is performed; it grows according to its own rules. Then, some features are randomly selected from the feature set for node splitting selection. This process is repeated until the number of decision trees reaches the set target. By training 70% of the samples, the basic situation of the user driving behavior analysis model in analyzing the user driving aggression index is initially grasped. Then, the remaining 30% of the test set is analyzed to verify the user driving behavior analysis model and finally obtain the model with the highest accuracy in analyzing the user driving aggression index.
[0010] It should be noted that, as a preferred technical solution for the artificial intelligence-based electric vehicle control method, the specific steps of S3 are as follows: Based on the electric vehicle's driving path data and the user's driving aggression index, the battery's future heat generation power prediction analysis results are obtained. The process of predicting and analyzing the battery's future heat generation power is as follows: Obtain road slope data, speed data, and transmission efficiency data for each future moment of the electric vehicle's operation, and simultaneously obtain the user's driving aggression index; calculate the power required for the vehicle to maintain its driving state at each future moment: multiply the required traction force for each future moment of the driving state by the speed to obtain the required power for each future moment of the driving state. The required traction force is obtained by adding rolling resistance, slope resistance, air resistance, and acceleration resistance; divide the power required for the vehicle to maintain its driving state at each future moment by the transmission efficiency to obtain the predicted analysis results of the battery's future heat generation power under driving conditions; multiply the power required for the vehicle to maintain its driving state at each future moment by the transmission efficiency to obtain the predicted analysis results of the battery's future heat generation power under regenerative braking conditions; thus, the predicted analysis results of the battery's future heat generation power are obtained.
[0011] It should be noted that, as a preferred technical solution for the electric vehicle control method based on artificial intelligence, the specific steps of S4 are as follows: Predictive analysis of the required heat dissipation power of the battery at each future moment is performed using the electric vehicle battery status data and the predicted analysis results of the battery heat generation power at each future moment. The process of predicting the required heat dissipation power of the battery at each future moment involves: acquiring real-time temperature data of the electric vehicle battery, historical temperature change rate data under the same driving conditions, total mass data of the electric vehicle battery, overall average specific heat capacity data of the battery pack, and the predicted analysis results of the battery heat generation power at each future moment; using Euler's method to numerically integrate the real-time temperature data of the electric vehicle battery and the historical temperature change rate data under the same driving conditions to obtain the predicted battery temperature value at each future moment; obtaining the battery temperature change rate at each future moment based on the predicted battery temperature value; multiplying the total mass data of the electric vehicle battery, the overall average specific heat capacity data of the battery pack, and the battery temperature change rate at each future moment to obtain the battery storage power at each future moment; and subtracting the battery storage power at each future moment from the predicted analysis results of the battery heat generation power at each future moment to obtain the predicted analysis results of the required heat dissipation power of the battery at each future moment.
[0012] It should be noted that, as a preferred technical solution for the electric vehicle control method based on artificial intelligence, the specific steps of S5 are as follows: select a suitable heat dissipation mode for the electric vehicle battery based on the predicted analysis results of the heat dissipation power required by the battery at each future moment. The specific process is as follows: obtain the heat dissipation power of each heat dissipation mode of the electric vehicle battery at each future moment, subtract the heat dissipation power of each heat dissipation mode from the predicted analysis results of the heat dissipation power required by the battery at each future moment, add the results together and perform absolute value processing to obtain the heat dissipation mode with the smallest absolute value result, and take it as the most suitable heat dissipation mode for the electric vehicle battery at each future moment.
[0013] The artificial intelligence-based electric vehicle control system module is implemented based on the aforementioned artificial intelligence-based electric vehicle control method. Specifically, it includes a battery thermal management data acquisition module, a user driving behavior habit analysis module, a battery heat generation power prediction and analysis module, a battery heat dissipation power prediction and analysis module, and a battery heat dissipation mode selection module. The battery thermal management data acquisition module is used to acquire historical user driving habit data, electric vehicle driving path data, and electric vehicle battery status data. The user driving behavior habit analysis module is used to establish a user driving behavior analysis model. Historical user driving habit data is input into the user driving behavior analysis model to obtain the user driving aggression index. The battery heat generation power prediction and analysis module is used to predict and analyze the battery heat generation power at various future moments based on the electric vehicle's driving path data and the user's driving aggression index. The battery heat dissipation power prediction and analysis module is used to predict and analyze the required heat dissipation power of the battery at each future moment based on the battery status data of the electric vehicle and the battery heat generation power prediction and analysis results at each future moment. The battery heat dissipation mode selection module is used to select a suitable heat dissipation mode for the electric vehicle battery at each future time based on the predicted analysis results of the heat dissipation power required by the battery at each future time.
[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention acquires historical user driving habit data, electric vehicle driving path data, and electric vehicle battery status data; establishes a user driving behavior analysis model, inputs historical user driving habit data into the user driving behavior analysis model to obtain a user driving aggression index; performs battery heat generation power prediction analysis based on electric vehicle driving path data and user driving aggression index; performs battery heat generation power prediction analysis based on electric vehicle battery status data and the battery heat generation power prediction analysis results for future times; selects appropriate heat dissipation modes for electric vehicle batteries for future times based on the battery heat dissipation power prediction analysis results for future times; and achieves forward-looking and proactive thermal management by integrating historical driving habits, real-time path information, and battery status data to accurately predict battery heat generation power and required heat dissipation power for future times, thus initiating or adjusting appropriate heat dissipation modes in advance before the battery temperature rises significantly, changing the passive response and delayed control situation in traditional technologies, and greatly improving the safety margin of battery operation. Attached Figure Description
[0015] Figure 1 This is a schematic diagram of the overall process of the artificial intelligence-based electric vehicle control method of this application.
[0016] Figure 2 This is a flowchart illustrating step S2 of the artificial intelligence-based electric vehicle control method of this application.
[0017] Figure 3 This is a schematic diagram of the overall framework of the artificial intelligence-based electric vehicle control system module of this application.
[0018] Figure 4 This is a flowchart illustrating step S3 of the artificial intelligence-based electric vehicle control method of this application.
[0019] Figure 5 This is a flowchart illustrating step S4 of the artificial intelligence-based electric vehicle control method of this application. Detailed Implementation
[0020] To better understand this application, various aspects of this application will be described in more detail with reference to the accompanying drawings.
[0021] To address the technical problems raised in the background art, this application provides a preferred embodiment: The specific content of this embodiment is as follows: like Figure 1 As shown, the artificial intelligence-based electric vehicle control method includes the following specific steps: S1. Obtain historical user driving habit data, electric vehicle driving route data, and electric vehicle battery status data; In this embodiment, the specific steps of S1 are as follows: S11. Historical user driving habit data consists of historical user driving segment data, obtained from the database. This includes the average accelerator pedal opening, standard deviation of opening, maximum opening, average of pedal change rate and number of rapid accelerations, total number of brake pedal presses divided by driving segment duration, number of emergency brakings, total braking time divided by segment duration, standard deviation of steering wheel angle, average of steering wheel rotation speed and number of sharp turns, standard deviation of vehicle speed, standard deviation of longitudinal acceleration and maximum value of lateral acceleration. S12. The data on the path of the electric vehicle includes road slope data, speed data and transmission efficiency data at future times when the electric vehicle is in motion. The road slope data and speed data at future times when the electric vehicle is in motion are obtained through navigation data, and the transmission efficiency data is obtained through bench tests. S13. Electric vehicle battery status data includes real-time battery temperature data, historical temperature change rate data under the same driving conditions, total battery mass data, and overall average specific heat capacity data of the battery pack. This data is retrieved from the database. S2, such as Figure 2 As shown, a user driving behavior analysis model is established. Historical user driving habit data is input into the user driving behavior analysis model to obtain the user driving aggression index. S21. Collect driving segment data from different drivers, road conditions, and times to build a driving behavior database; In this embodiment, S21 includes the following specific steps: collecting driving segment data of different drivers, road conditions, and times, performing algorithmic labeling on each driving segment data, and constructing a driving behavior database; wherein, different drivers are drivers of different ages, genders, and driving experience; different road conditions are road conditions when driving in urban congestion, highways, suburbs, and mountain roads; different times are daytime, nighttime, peak hours, and off-peak hours; driving segment data is time-series segment data recorded at fixed time intervals; the process of performing algorithmic labeling on each driving segment data is as follows: calculating the standard deviation of longitudinal acceleration within each data segment, setting a driving segment threshold, assigning a label to each driving segment according to the set driving segment threshold, and combining the driving segment data after labeling into a driving behavior database; it should be noted that the driving segment threshold can be set in the following way: sorting the standard deviation of each driving segment from smallest to largest, determining two quantiles for the sorting result, when the 30% and 70% quantiles are selected as quantiles, then threshold 1 is the 30% quantile and threshold 2 is the 70% quantile.
[0022] S22. Extract features from driving segments in the driving behavior database to obtain feature vectors for each feature group of the driving segment; In this embodiment, S22 includes the following specific steps: extracting features from driving segments in the driving behavior database to obtain feature vectors for each feature group of the driving segment. Each feature group includes an accelerator pedal feature group, a brake pedal feature group, a steering wheel feature group, and a vehicle dynamics feature group. The accelerator pedal feature group includes average pedal opening, standard deviation of opening, maximum opening, average pedal change rate, and number of rapid accelerations. The brake pedal feature group includes the total number of brake pedal presses divided by the driving segment duration, the number of emergency brakings, and the total braking time divided by the segment duration. The steering wheel feature group includes the standard deviation of steering wheel angle, average steering wheel rotation speed, and number of sharp turns. The vehicle dynamics feature group includes the standard deviation of vehicle speed, the standard deviation of longitudinal acceleration, and the maximum value of lateral acceleration. It should be noted that in the accelerator pedal feature group, the average pedal opening reflects average power demand, the standard deviation of opening reflects stability, the maximum opening reflects maximum power request, the average pedal change rate reflects emergency acceleration intent, and the number of rapid accelerations is calculated as follows: whenever the pedal change rate exceeds 80... The number of rapid accelerations is incremented by 1 for events with a rate of % / s; in the brake pedal feature group, the total number of brake pedal presses divided by the driving segment duration reflects the complexity of the driving environment. The number of emergency braking events is calculated as follows: the braking pressure instantaneously exceeds a certain set threshold. The total braking time divided by the segment duration reflects the relative proportion of the driving process in a deceleration state; in the steering wheel feature group, the standard deviation of the steering wheel angle reflects the steering amplitude, the average steering wheel rotation speed reflects sharp steering, and the number of sharp turns is calculated as the rotation speed exceeds 150 degrees per second; in the vehicle dynamics feature group, the standard deviation of vehicle speed reflects speed fluctuations, the standard deviation of longitudinal acceleration reflects the severity and fluctuation of acceleration and deceleration during driving, and the maximum value of lateral acceleration reflects the intensity of cornering.
[0023] S23. Construct a user driving behavior analysis model, input historical user driving habit data into the user driving behavior analysis model, and obtain the user driving aggression index; In this embodiment, S23 includes the following specific steps: A user driving behavior analysis model is constructed using the random forest algorithm and feature vectors of each feature group of the driving segment. Historical user driving habit data is input to obtain the user driving aggression index. The process of constructing the user driving behavior analysis model is as follows: The feature vectors of each feature group of the driving segment are randomly shuffled and divided into 70% as the training set and 30% as the test set. The training set is input into the random forest algorithm, which generates multiple training subsets for decision tree training until the number of decision trees reaches the required level. A single decision tree is constructed using the CART algorithm, and feature selection is performed using the Gini coefficient minimization criterion. The smaller the Gini coefficient value, the purer the information contained in the dataset. After the decision tree is constructed, no pruning is performed; it grows according to its own rules. Then, a portion is randomly selected from the feature set. Features are used to calculate node splitting selection. This process is repeated until the number of decision trees reaches the target. By training on 70% of the samples, the basic situation of the user driving behavior analysis model in analyzing the user driving aggression index is initially grasped. Then, the remaining 30% of the test set is analyzed to verify the user driving behavior analysis model and finally obtain the model with the highest accuracy in analyzing the user driving aggression index. It should be noted that the random forest algorithm can handle numerical features without complex feature scaling. It can also automatically evaluate feature importance, determine the contribution of each behavior to aggression, is not sensitive to outliers, and is not prone to overfitting. The training and prediction speeds are fast, meeting the real-time requirements of in-vehicle systems. Randomly selecting some features from the feature set for calculating node splitting selection can avoid overfitting in the classification model.
[0024] S3, such as Figure 4 As shown, the battery heat generation power prediction analysis is performed at various future moments based on the electric vehicle's driving path data and the user's driving aggression index. In this embodiment, the specific steps of S3 are as follows: Based on the electric vehicle's driving path data and the user's driving aggression index, the battery's future heat generation power prediction analysis results are obtained. The process of predicting and analyzing the battery's future heat generation power involves: acquiring road slope data, speed data, and transmission efficiency data for each future moment of the electric vehicle's operation, and simultaneously acquiring the user's driving aggression index; calculating the power required for the vehicle to maintain its driving state at each future moment: multiplying the required traction force for each future moment by the speed to obtain the required power for the driving state at each future moment. The required traction force is obtained by considering rolling resistance, slope resistance, air resistance, and acceleration resistance. Add them together; divide the power required for the vehicle to maintain its driving state at each future moment by the transmission efficiency to obtain the predicted analysis results of the battery heat generation power of the electric vehicle at each future moment under driving conditions; multiply the power required for the vehicle to maintain its driving state at each future moment by the transmission efficiency to obtain the predicted analysis results of the battery heat generation power of the electric vehicle at each future moment under regenerative braking conditions; obtain the predicted analysis results of the battery heat generation power of the electric vehicle at each future moment in this way; it should be noted here that rolling resistance: vehicle mass, gravitational acceleration and rolling resistance coefficient (depending on tire type and road conditions, generally, for cars on good roads, the rolling resistance coefficient is about 0.01 to 0).Multiply by (between 0 and 15); Gradient drag: the sine of vehicle mass, gravitational acceleration, and slope angle (in navigation data, slope is usually expressed as a percentage and needs to be converted to angle); Air resistance: half, air density (varying with temperature and altitude), air resistance coefficient, vehicle frontal area, and vehicle speed (speed relative to air, usually ignoring wind speed and taking vehicle speed); Acceleration drag: the product of vehicle mass and vehicle acceleration; When the vehicle is traveling at a constant speed, acceleration drag is 0; when the vehicle is going uphill, gradient drag is positive; when going downhill, gradient drag is negative (helping the vehicle move and reducing the required traction force); The predicted analysis result of battery heat generation power at future times is the power required for the vehicle to maintain its driving state at future times divided by the transmission efficiency (in drive mode) or multiplied by the transmission efficiency (in regenerative braking mode, recovering energy). The difference between drive and regenerative braking: in drive mode, and the refueling category of fuel vehicles. Similarly, when the accelerator pedal is pressed, the battery pack supplies direct current (DC) to the drive motor, which is then converted to alternating current (AC) by an inverter. This drives the motor to rotate, and the motor's torque is transmitted to the wheels through a reducer, propelling the vehicle forward. In this process, the motor is an energy-consuming device, converting electrical energy into mechanical energy. Regenerative braking mode is a key technology for improving energy efficiency. When the accelerator pedal is released or the brake pedal is lightly applied, the vehicle's control system cuts off the battery's power supply to the motor. Due to the vehicle's inertia, the wheels, in turn, drive the motor to rotate. At this time, the motor no longer consumes electricity but is forced to rotate by the wheels. According to the principle of electromagnetic induction, a motor driven to rotate by an external force becomes a generator, producing current. The generated AC is rectified into DC by an inverter and then supplied to the battery pack to charge the battery. In this process, the motor is a power generation device, converting the vehicle's kinetic energy (i.e., inertia) back into electrical energy.
[0025] S4, such as Figure 5 As shown, based on the electric vehicle battery status data and the predicted analysis results of battery heat generation power at various future times, the predicted analysis of the required heat dissipation power of the battery at various future times is performed. In this embodiment, the specific steps of S4 are as follows: The electric vehicle battery status data and the predicted analysis results of battery heat generation power at future times are used to predict the required heat dissipation power of the battery at future times. The process of predicting the required heat dissipation power of the battery at future times involves: acquiring real-time temperature data of the electric vehicle battery, historical temperature change rate data under the same driving conditions, total mass data of the electric vehicle battery, overall average specific heat capacity data of the battery pack, and the predicted analysis results of battery heat generation power at future times; applying Euler's method to the real-time temperature data of the electric vehicle battery and the historical temperature change rate data under the same driving conditions to obtain the predicted battery temperature value at future times; obtaining the battery temperature change rate at future times based on the predicted battery temperature value; multiplying the total mass data of the electric vehicle battery, the overall average specific heat capacity data of the battery pack, and the battery temperature change rate at future times to obtain the battery storage power at future times; and subtracting the battery storage power at future times from the predicted analysis results of battery heat generation power to obtain the predicted analysis results of the required heat dissipation power of the battery at future times. It should be noted that Euler's method is used for numerical integration. Integration aims to predict future battery temperature changes. It estimates the temperature at the next moment using the current temperature and rate of temperature change. Each step uses the current state (temperature and rate of change) to predict the next temperature, and uses the predicted temperature as the current temperature for the next step. This process is repeated until the entire time range to be predicted is covered. Based on the law of conservation of energy: the heat generated by the battery equals the heat absorbed and stored by the battery itself plus the heat that needs to be removed by the cooling system. Therefore, the required cooling power equals the predicted heat generation power minus the stored power. Calculating the future starts from the real-time temperature, eliminating the accumulation of historical prediction errors and ensuring the prediction closely tracks the battery's actual state. Historical temperature change rate data under the same driving conditions serves as a dynamic behavioral reference for the prediction results, providing empirical evidence and enhancing the rationality of the prediction. The total mass data of the electric vehicle battery and the overall average specific heat capacity data of the battery pack are core physical parameters for calculating battery thermal inertia, quantifying the battery's heat capacity. As a whole, the battery pack absorbs how much heat for every 1°C increase in temperature, measuring the battery pack's thermal buffering capacity. S5. Based on the predicted and analyzed results of the heat dissipation power required by the battery at each future time, select the appropriate heat dissipation mode for the electric vehicle battery at each future time.
[0026] In this embodiment, the specific steps of S5 are as follows: Based on the predicted analysis results of the heat dissipation power required by the battery at various future times, a suitable heat dissipation mode is selected for the electric vehicle battery. Specifically, this involves obtaining the heat dissipation power of each heat dissipation mode for the electric vehicle battery at various future times, subtracting the heat dissipation power of each heat dissipation mode from the predicted analysis results of the heat dissipation power required by the battery at various future times, summing the results, and performing absolute value processing to obtain the heat dissipation mode with the smallest absolute value. This is then taken as the most suitable heat dissipation mode for the electric vehicle battery at various future times. It should be noted that the heat dissipation power of each heat dissipation mode at various future times represents the corresponding heat dissipation capacity. This ensures accurate matching of the most suitable heat dissipation mode for future times, thereby avoiding insufficient or wasted heat dissipation capacity. Since the selection is based on prediction, the system can switch heat dissipation modes smoothly and in advance, avoiding the drastic switching of heat dissipation modes caused by sudden temperature changes in traditional control, making the system work more stably.
[0027] Based on the above implementation, this embodiment has the following advantages over the prior art: This embodiment acquires historical user driving habit data, electric vehicle driving path data, and electric vehicle battery status data; establishes a user driving behavior analysis model, inputs historical user driving habit data into the user driving behavior analysis model to obtain the user driving aggression index; performs battery heat generation power prediction analysis at future times based on electric vehicle driving path data and user driving aggression index; performs battery heat generation power prediction analysis at future times based on electric vehicle battery status data and battery heat generation power prediction analysis results; selects appropriate heat dissipation modes for electric vehicle batteries at future times based on battery heat dissipation power prediction analysis results; by integrating historical driving habits, real-time path information, and battery status data, it accurately predicts battery heat generation power and required heat dissipation power at future times, achieving forward-looking and proactive thermal management, activating or adjusting appropriate heat dissipation modes in advance before the battery temperature rises significantly, changing the passive response and delayed control situation in traditional technology, and greatly improving the safety margin of battery operation.
[0028] like Figure 3As shown, this embodiment also provides an artificial intelligence-based electric vehicle control system module, which is implemented based on the aforementioned artificial intelligence-based electric vehicle control method. Specifically, it includes a battery thermal management data acquisition module, a user driving behavior habit analysis module, a battery heat generation power prediction and analysis module, a battery heat dissipation power prediction and analysis module, and a battery heat dissipation mode selection module. The battery thermal management data acquisition module is used to acquire historical user driving habit data, electric vehicle driving path data, and electric vehicle battery status data. The user driving behavior habit analysis module is used to establish a user driving behavior analysis model, inputting historical user driving habit data into the model to obtain a user driving aggression index. The battery heat generation power prediction and analysis module is used to predict and analyze the battery heat generation power at future times based on the electric vehicle driving path data and the user driving aggression index. The battery heat dissipation power prediction and analysis module is used to predict and analyze the required heat dissipation power of the battery at future times based on the electric vehicle battery status data and the predicted heat generation power at future times. The battery heat dissipation mode selection module is used to select a suitable heat dissipation mode for the electric vehicle battery at future times based on the predicted heat dissipation power required at future times.
[0029] The specific steps for each unit module in the AI-based electric vehicle control system module of this application to implement the corresponding functions can be found in the steps in the embodiments of the AI-based electric vehicle control method described above, and will not be repeated here.
[0030] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this application.
Claims
1. An electric vehicle control method based on artificial intelligence, characterized by, The application relates to a method for predicting battery heat dissipation power of an electric vehicle. The method comprises the following steps: S1, acquiring historical user driving habit data, path condition data of electric vehicle driving and battery condition data of the electric vehicle; S2, establishing a user driving behavior analysis model, inputting the historical user driving habit data into the user driving behavior analysis model to obtain a user driving aggressiveness index; S3, performing battery future time point battery heat production power prediction analysis based on the path condition data of electric vehicle driving and the user driving aggressiveness index; S4, performing future time point battery required heat dissipation power prediction analysis based on the electric vehicle battery condition data and the future time point battery heat production power prediction analysis result; 2.The AI-based electric vehicle control method of claim 1, wherein, S5, selecting a suitable heat dissipation mode for the electric vehicle battery at each future time point according to the future time point battery required heat dissipation power prediction analysis result. The S2 comprises the following specific steps: S21, collecting driving segment data of different drivers, road conditions and times to construct a driving behavior database; S22, extracting features from the driving segments in the driving behavior database to obtain feature vectors of each feature group of the driving segments; 3.The AI-based electric vehicle control method of claim 2, wherein, S23, constructing a user driving behavior analysis model, inputting the historical user driving habit data into the user driving behavior analysis model to obtain a user driving aggressiveness index.
4. The artificial intelligence-based electric vehicle control method of claim 3, wherein, The specific steps of the S21 are as follows: collecting driving segment data of different drivers, road conditions and times, algorithmically labeling each driving segment data to construct a driving behavior database; wherein the different drivers are drivers of different ages, genders and driving experiences; the different road conditions are road conditions during city congestion, highway driving, suburban driving and mountain road driving; the different times are day, night, peak and flat peak; the driving segment data are time sequence segment data recorded in a fixed time segment; the process of algorithmically labeling each driving segment data is as follows: calculating the standard deviation of the longitudinal acceleration in each data segment, setting a driving segment threshold, assigning a label to each driving segment according to the set driving segment threshold, and collecting the driving segment data after the label assignment as the driving behavior database. The specific steps of the S22 are as follows: extracting features from the driving segments in the driving behavior database to obtain feature vectors of each feature group of the driving segments, wherein the feature groups of the driving segments include an accelerator pedal feature group, a brake pedal feature group, a steering wheel feature group and a vehicle dynamic feature group, the accelerator pedal feature group includes an average opening degree, a standard deviation of the opening degree, a maximum opening degree, an average value of a pedal change rate and a number of sudden accelerations, the brake pedal feature group includes a total number of brake pedal depressions divided by a driving segment time length, a number of sudden brakes and a total brake pedal depression time divided by a segment time length, the steering wheel feature group includes a standard deviation of a steering wheel rotation angle, an average value of a steering wheel rotation speed and a number of sudden turns, and the vehicle dynamic feature group includes a standard deviation of a vehicle speed, a standard deviation of a longitudinal acceleration and a maximum value of a lateral acceleration. 5.The AI-based electric vehicle control method of claim 4, wherein, The specific steps of S23 are: using a random forest algorithm and feature vectors of each feature group of the driving segment to construct a user driving behavior analysis model, inputting historical user driving habit data to obtain a user driving aggressiveness index, wherein the process of constructing the user driving behavior analysis model is: randomly shuffling the feature vectors of each feature group of the driving segment, dividing 70% as a training set and 30% as a test set, inputting the training set into the random forest algorithm, the algorithm obtains multiple training subsets for decision tree training until the number of obtained decision trees reaches the requirement, using the CART algorithm to construct a single decision tree, using the Gini index minimization criterion for feature selection, the smaller the Gini coefficient value, the purer the information contained in the data set, after the decision tree is constructed, no pruning is performed, and the decision tree grows according to its own rules, then some features are randomly selected from the feature set for node splitting selection, the above process is repeated until the number of decision trees reaches the set requirement, through training and learning 70% of the samples, the user driving behavior analysis model preliminarily masters the basic situation of the user driving aggressiveness index analysis, then the remaining 30% of the test set is analyzed, thereby verifying the user driving behavior analysis model and finally obtaining the model with the highest accuracy of user driving aggressiveness index analysis. 6.The AI-based electric vehicle control method of claim 5, wherein, The specific steps of S3 are: obtaining battery future time battery heat power prediction analysis results according to the path condition data of the electric vehicle and the user driving aggressiveness index, wherein the future time battery heat power prediction analysis process is: obtaining road slope data, speed data and transmission efficiency data of the electric vehicle at each future time, and obtaining the user driving aggressiveness index; calculating the power required for the vehicle to maintain the driving state at each future time: multiplying the required traction at each future time by the speed to obtain the power required for the vehicle to maintain the driving state at each future time, the required traction is obtained by adding the rolling resistance, slope resistance, air resistance and acceleration resistance; dividing the power required for the vehicle to maintain the driving state at each future time by the transmission efficiency to obtain the battery heat power prediction analysis result of the electric vehicle under the driving condition at each future time; multiplying the power required for the vehicle to maintain the driving state at each future time by the transmission efficiency to obtain the battery heat power prediction analysis result of the electric vehicle under the regenerative braking condition at each future time; in this way, the battery heat power prediction analysis result of the electric vehicle maintaining at each future time is obtained.
7. The artificial intelligence-based electric vehicle control method of claim 6, wherein, The specific steps of S4 are: the battery case data of the electric vehicle and the future battery heat generation power prediction analysis result are used to predict the future battery heat dissipation power required; the future battery heat dissipation power prediction analysis process is: the real-time temperature data of the electric vehicle battery, the historical temperature change rate data under the driving condition, the total mass data of the electric vehicle battery, the overall average specific heat capacity data of the battery pack and the future battery heat generation power prediction analysis result are obtained, the Euler method numerical integration is used for the real-time temperature data of the electric vehicle battery and the historical temperature change rate data under the driving condition, and the future battery temperature prediction value is obtained; the future battery temperature change rate is obtained according to the future battery temperature prediction value, the total mass data of the electric vehicle battery, the overall average specific heat capacity data of the battery pack and the future battery temperature change rate are multiplied, the future battery storage power is obtained, and the future battery heat generation power prediction analysis result is subtracted from the future battery storage power, and the future battery heat dissipation power prediction analysis result is obtained.
8. The artificial intelligence-based electric vehicle control method of claim 7, wherein, The specific steps of S5 are: according to the future battery heat dissipation power prediction analysis result, a suitable heat dissipation mode is selected for the electric vehicle battery, and the specific process is: the heat dissipation power of each heat dissipation mode of the electric vehicle battery at each future time is obtained, the future battery heat dissipation power prediction analysis result is subtracted from the heat dissipation power of each heat dissipation mode, the results are added and then subjected to absolute value processing, the heat dissipation mode with the smallest absolute value result is obtained, and the heat dissipation mode is used as the most suitable heat dissipation mode of the electric vehicle battery at each future time.
9. An electric vehicle control system based on artificial intelligence, implemented based on the electric vehicle control method based on artificial intelligence of any one of claims 1-8, characterized in that, It specifically includes a battery thermal management data acquisition module, a user driving behavior habit analysis module, a battery heat generation power prediction analysis module, a battery heat dissipation power prediction analysis module and a battery heat dissipation mode selection module, the battery thermal management data acquisition module is used to acquire historical user driving habit data, path condition data of the electric vehicle and electric vehicle battery condition data; The user driving behavior habit analysis module is used to establish a user driving behavior analysis model, input the historical user driving habit data into the user driving behavior analysis model, and obtain a user driving aggressiveness index; The battery heat generation power prediction analysis module is used to perform battery future battery heat generation power prediction analysis based on the path condition data of the electric vehicle and the user driving aggressiveness index; The battery heat dissipation power prediction analysis module is used to perform future battery heat dissipation power prediction analysis based on the electric vehicle battery condition data and the future battery heat generation power prediction analysis result; The battery heat dissipation mode selection module is used to select a suitable heat dissipation mode for the future electric vehicle battery at each future time according to the future battery heat dissipation power prediction analysis result.