Electric vehicle driving data management system

Abstract

The invention discloses an electric vehicle driving data management system, which comprises a multi-source data acquisition module, a processing module, a grading judgment module, a dynamic fusion module, an improvement module and an interaction feedback module, the multi-source data acquisition module is used for acquiring various driving data of the electric vehicle in real time; the processing module is used for receiving and processing various mass driving data; and the grading judgment module is used for dynamically judging the data grade and executing a grading storage strategy. The invention belongs to the technical field of electric vehicle driving data management, and aims at solving the problems that in the prior art, a data storage mode is single, distinguishing is not carried out according to the importance of driving data, and meanwhile a large deviation exists between a prediction result and the actual endurance mileage. The method has the technical effects that the data can be conveniently divided into the high importance level and the low importance level, the differentiated storage strategy is executed, and meanwhile, the actual energy consumption condition of the vehicle can be reflected more accurately.

Description

Technical Field

[0001] The present application belongs to the technical field of electric vehicle driving data management, and specifically relates to an electric vehicle driving data management system. Background Art

[0002] With the increasing popularity of electric vehicles, the management of their driving data has become increasingly important. Electric vehicle driving data covers a rich and diverse range of content. Among these, the vehicle's operating status is a crucial component of driving data. This includes various battery parameters such as charge, voltage, current, and temperature. As the core power source of electric vehicles, the battery's status directly affects the vehicle's range and power output.

[0003] In addition to battery status, driving habits are also an integral part of electric vehicle data. Drivers' acceleration, deceleration, braking frequency, and steering maneuvers are all recorded in detail. Different driving habits significantly impact vehicle energy consumption, wear, and safety.

[0004] Environmental information also plays a crucial role in electric vehicle driving data. This encompasses a wide range of factors, including the vehicle's geographic location, weather conditions, and road conditions. Geographical information can help determine the vehicle's specific location and trajectory, providing precise positioning for navigation systems. Weather conditions such as temperature, humidity, and wind speed can also impact vehicle performance.

[0005] In summary, electric vehicle driving data has a crucial impact on safe operation, performance optimization, and the user's driving experience. Accurate and comprehensive management of this driving data not only ensures safe electric vehicle operation but also provides rich data support for vehicle research and development, improvement, and after-sales service.

[0006] However, most current electric vehicle driving data management systems use a single data storage method and do not differentiate based on the importance of driving data, which increases storage costs and reduces data retrieval efficiency. At the same time, in terms of range prediction, existing systems mainly rely on the vehicle's own battery power and simple energy consumption models for estimation, without fully considering the impact of real-time driving behavior and environmental factors on range, resulting in a large deviation between the predicted results and the actual range, causing inconvenience to drivers. Summary of the Invention

[0007] The present application provides an electric vehicle driving data management system, which aims to solve the problems of the existing technology in that the data storage method is single, the driving data is not differentiated according to its importance, and there is a large deviation between the predicted results and the actual cruising range.

[0008] An electric vehicle driving data management system includes a multi-source data acquisition module, a processing module, a classification determination module, a dynamic fusion module, an improvement module and an interactive feedback module;

[0009] The multi-source data acquisition module is used to collect various types of driving data of electric vehicles in real time;

[0010] The processing module is used to receive and process massive amounts of various types of driving data;

[0011] The hierarchical determination module is used to dynamically determine the data level and implement a hierarchical storage strategy to ensure that key data of various types of driving data are processed and stored with priority;

[0012] The dynamic fusion module is used to integrate historical power consumption data, real-time driving behavior and environmental data;

[0013] The improved module can use the comprehensive data set to adjust the range prediction to compensate for the long-term decline in battery capacity;

[0014] The interactive feedback module is used to obtain and display various driving data of the vehicle in real time.

[0015] Furthermore, the multi-source data acquisition module includes a sensor unit, a data reading unit and a data transmission unit. The sensor unit can collect vehicle driving data such as battery power, motor speed and vehicle speed in real time to form vehicle status data;

[0016] The data reading unit is used to read historical power consumption values;

[0017] The data transmission unit is responsible for quickly and accurately transmitting various types of driving data collected in real time to the processing module.

[0018] Furthermore, the classification determination module includes a threshold setting unit, a model building unit, a determination unit and a storage unit;

[0019] The threshold setting unit is used to preset thresholds according to key parameters during vehicle driving;

[0020] The model building unit is capable of building a machine learning model using a label set and a random forest classifier;

[0021] The determination unit can input key parameters in the real-time collected vehicle driving data into the rule engine. The rule engine matches the key parameters according to predefined rules and determines the level of the key parameters.

[0022] The storage unit can execute a differentiated storage strategy according to the data level finally determined by the determination unit.

[0023] Furthermore, the specific content of the model building unit is as follows:

[0024] a) Build a machine learning model based on the random forest classifier;

[0025] b) Optimization of machine learning models;

[0026] c) Introduction of rule engine.

[0027] Furthermore, the threshold setting unit further includes a feature labeling unit, which can label key parameters in the historical driving data that are greater than the threshold to form a labeling set.

[0028] Furthermore, the key parameter levels are divided into high importance levels and low importance levels. The high importance level data is stored in full, and the low importance level data is stored in compressed form.

[0029] Furthermore, the historical power consumption value data is extracted from the historical driving records of the vehicle, including 10-20 days of historical power consumption value data under different driving conditions.

[0030] Furthermore, the specific contents of the improvement module are as follows:

[0031] a) Extracting comprehensive data sets;

[0032] b) Learning dataset information;

[0033] c) Construction of ARIMA model;

[0034] d) Calculation of final prediction results.

[0035] Furthermore, the introduction of the rule engine is used to define hierarchical rule conditions based on key parameters and thresholds during vehicle driving.

[0036] Compared with the prior art, this application has at least the following beneficial effects:

[0037] This application is based on further analysis and research of existing technical problems. It adopts a hierarchical storage strategy through a hierarchical judgment module, divides data into high-importance levels and low-importance levels according to key parameters of the data, and implements a differentiated storage strategy to provide data support for the continuous operation and subsequent analysis of the system, avoiding low-importance data from occupying too much storage space. At the same time, by fusing multi-source data and using machine learning models to dynamically improve the cruising range, it can more accurately reflect the actual energy consumption of the vehicle and improve the accuracy of the cruising range prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A module diagram of an electric vehicle driving data management system provided in one embodiment of the present application. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of this application more clear, this application is further described in detail below with reference to the accompanying drawings and embodiments.

[0040] like Figure 1 As shown, an electric vehicle driving data management system is provided, which includes a multi-source data acquisition module, a processing module, a classification determination module, a dynamic fusion module, an improvement module and an interactive feedback module.

[0041] The multi-source data acquisition module is used to collect various types of driving data of electric vehicles in real time, including vehicle status data, driving behavior data, environmental data and historical power consumption data; after the vehicle is started, the data acquisition module starts to collect data in real time and sends the data to the processing module through the data transmission unit.

[0042] The multi-source data acquisition module includes a sensor unit, a data reading unit and a data transmission unit;

[0043] The sensor unit can collect vehicle driving data such as battery power, motor speed and vehicle speed in real time. These data reflect the basic operating status of the vehicle and form vehicle status data; and use on-board sensors to obtain information such as pedal depth, braking frequency, steering angle, etc., which are used to analyze the driver's operating habits and form driving behavior data; at the same time, the sensor module can use the environmental sensors installed on the vehicle or connect with external data sources (such as weather stations, map services) to obtain environmental information such as road slope, wind speed, temperature, etc. to form environmental data.

[0044] The data reading unit is used to read historical power consumption values, a key parameter. This historical power consumption value can then be combined with real-time driving behavior data and environmental data to create a more accurate range prediction model.

[0045] The data transmission unit plays a key role in connecting the multi-source data acquisition module with the processing module. It is responsible for quickly and accurately transmitting various types of driving data collected in real time to the processing module. Its core functions include ensuring real-time data transmission, guaranteeing data accuracy and integrity, enabling inter-system communication, and improving system reliability and stability, thereby supporting real-time vehicle monitoring, energy efficiency optimization, and intelligent decision-making.

[0046] The processing module is used to receive and process massive amounts of driving data to ensure its accuracy and consistency. The specific tasks are as follows:

[0047] After receiving various types of driving data, the processing module first cleans the various types of driving data to remove obvious outliers, and then normalizes the various types of driving data separately, mapping them to the [0, 1] interval to facilitate subsequent analysis and processing.

[0048] Remove obvious outliers: Use statistical methods (such as boxplots) to detect outliers, and then use clustering algorithms to identify outliers. If the outliers are caused by data entry errors or equipment failures, they can be directly deleted. For some explainable outliers, improvements can be made based on the actual situation.

[0049] For example, if an abnormally high speed value is caused by a brief sensor failure, the average of the preceding and following speed values ​​can be used as a replacement.

[0050] Normalization is the process of mapping individual data to a specific interval. The purpose is to eliminate the dimensional differences between different features, facilitate subsequent analysis and processing, and align timestamps.

[0051] The hierarchical determination module is used to dynamically determine the data level and implement a hierarchical storage strategy to ensure that key data of various types of driving data are processed and stored with priority, while reducing the storage overhead of non-critical data.

[0052] The classification determination module includes a threshold setting unit, a model building unit, a determination unit and a storage unit.

[0053] The threshold setting unit is used to preset thresholds based on key vehicle parameters during driving, facilitating subsequent triggering of corresponding level markings based on the thresholds. Key parameters include current vehicle speed, vehicle acceleration, battery charge, and motor temperature. Levels are categorized as high or low importance.

[0054] At the same time, the threshold setting unit also has a feature labeling unit, which can label key parameters in the historical driving data that are greater than the threshold to form a labeling set.

[0055] The model building unit can use the annotation set and random forest classifier to build a machine learning model. By training and optimizing the model, it can accurately predict and determine the levels of key parameters. The specific contents are as follows:

[0056] a) Building a machine learning model based on random forest classifier

[0057] The labeled dataset was divided into a training set and a test set. The training set and test set ratio was 7:3 to ensure that the model could fully learn the data features during training. The performance of the model was verified on the test set. The training set was used to train the random forest classifier. During training, the machine learning model learns how to predict the data level based on the input features.

[0058] b) Optimization of machine learning models

[0059] The model building unit can optimize model performance by adjusting the hyperparameters of the random forest classifier. Common hyperparameters include the number of decision trees, the maximum depth of each tree, the maximum number of features randomly selected for each node, etc. Grid search or random search can be used to find the optimal hyperparameter combination.

[0060] c) Introduction of rule engine

[0061] According to the key parameters and thresholds during vehicle driving, hierarchical rule conditions are defined, and rule files are written using Drools rule language (DRL) to express the rule conditions in the form of rules.

[0062] For example, the current vehicle speed > 100km / h is a high importance level, the vehicle speed < 100km / h is a low importance level; the vehicle acceleration > 3m / s 2 High importance level, vehicle acceleration <3m / s 2 It is a low importance level; the battery level <60℃ is a high importance level, and the battery level >60℃ is a low importance level; the motor temperature >90℃ is a high importance level, and the motor temperature <90℃ is a low importance level.

[0063] The judgment unit can input the key parameters of the vehicle driving data collected in real time into the rule engine. The rule engine matches the key parameters according to predefined rules and determines whether the key parameters meet the rule conditions. If the key parameters meet a certain rule condition, the rule engine outputs the preliminary judgment of the key parameter level. The key parameter level is divided into high importance level and low importance level.

[0064] The machine learning model extracts features for the key parameters determined by the rule engine. The extracted features are input into the trained random forest classifier, and the machine learning model outputs the predicted data level.

[0065] Compare the preliminary judgment results of the rule engine with the prediction results of the machine learning model. If the two results are consistent, the result is directly adopted as the final key parameter level. If the two results are inconsistent, the decision can be made based on the preset priority strategy. For example, the result of the rule engine can be prioritized because the rules of the rule engine are based on clear thresholds. Alternatively, the decision can be made based on the confidence level of the model. If the confidence level of the machine learning model is higher, the result of the machine learning model is adopted.

[0066] The storage unit implements differentiated storage strategies based on the data level ultimately determined by the determination unit, providing data support and historical records for the system's ongoing operation and subsequent analysis. High-importance data is fully stored. High-importance data is stored on high-performance storage devices and backed up regularly.

[0067] For low-importance data, compressed storage is used. The data is compressed using a data compression algorithm, and then stored on a low-cost storage device. When the data needs to be accessed, it is decompressed.

[0068] The dynamic fusion module integrates historical power consumption data, real-time driving behavior, and environmental data to provide a comprehensive foundation for range analysis and enhance the accuracy of subsequent range predictions. Historical power consumption data is extracted from the vehicle's historical driving records. This includes 10-20 days of historical power consumption data under different driving conditions to ensure that long-term trends in battery capacity are accurately captured. This provides a historical benchmark for machine learning models, helping them understand the impact of different driving conditions on power consumption. The details are as follows:

[0069] a) Data alignment and time synchronization

[0070] The interpolation method is used to align the historical power consumption data, real-time driving behavior data and environmental data to ensure that the timestamps of the historical power consumption data, real-time driving behavior data and environmental data are consistent for subsequent fusion.

[0071] b) Feature Selection

[0072] Using machine learning models, we extract range-related features from historical power consumption values ​​and output a predicted data level to determine data importance, such as average power consumption and power consumption change rate. We also extract features from real-time driving behavior data, such as average speed and acceleration standard deviation. We also extract features from environmental data, such as ambient temperature and wind speed.

[0073] According to the importance of different data determined by the machine learning model, weights are assigned to each data and the weighted data are fused to form a comprehensive data set. The data from different data sources are then fused through the machine learning model. The machine learning model can learn the associations and interactions between different data sources, improve the fusion effect, and comprehensively consider the impact of multiple factors on the cruising range.

[0074] The improved module can use comprehensive data sets to adjust range predictions and compensate for the long-term decline in battery capacity. By integrating multi-source data and collaborating with recurrent neural network models, the accuracy and reliability of range predictions are improved. The specific contents are as follows:

[0075] a) Extracting comprehensive datasets

[0076] The window length is set to 5 minutes, and 30 time steps of data are captured each time as the model input. This helps preserve the temporal dependencies of the data and helps the model learn the temporal patterns in the data.

[0077] b) Learning dataset information

[0078] The intercepted comprehensive data set is input into the long short-term memory network (LSTM) model to capture their impact on the range and obtain the preliminary range prediction value y i At the same time, get the actual mileage value x at the corresponding time point i ; Calculate the residual z between the predicted result and the actual value. The residual z reflects the deviation between the LSTM model predicted value and the actual value. The calculation formula of the residual z is:

[0079] z=x i -y i

[0080] c) Construction of ARIMA model

[0081] The residual z is used as the input of the ARIMA model. The ARIMA model consists of an autoregressive (AR) part, an integral (I) part, and a moving average (MA) part. The autoregressive part captures the autocorrelation in the residual sequence, that is, the correlation between the current residual and the past residual. The integral part differs the residual sequence to eliminate non-stationarity and make the sequence stable. The moving average part captures random fluctuations and noise in the residual sequence. The ARIMA model fits the residual z through the parameter estimation method to obtain the residual correction value z i .

[0082] d) Calculation of final prediction results

[0083] The residual correction value z of the ARIMA model i Add it back to the residual z calculated by the LSTM model to get the final predicted value e of the cruising range. The calculation formula for the final predicted value e of the cruising range is:

[0084] e=z+z i

[0085] The interactive feedback module is used to capture and display various vehicle driving data in real time, including but not limited to key indicators such as current speed, remaining battery charge, mileage, and battery temperature. This data is presented through intuitive dashboards or charts, allowing users to understand the vehicle's operating status at a glance. Furthermore, based on the accurate predictions provided by the LSTM-ARIMA hybrid range prediction module, the interactive feedback module can display the results of range improvements. This includes a comparison of predicted range with actual range, as well as the range trend over time. This information allows users to clearly see the range improvement effect, thereby strengthening their confidence in vehicle performance.

[0086] In addition to displaying real-time data, the interactive feedback module also provides a historical data review function. Users can view driving data and mileage changes over a period of time, and analyze the impact of factors such as driving habits and road conditions on mileage. This helps users summarize experience, optimize driving behavior, and further improve mileage.

[0087] In the electric vehicle driving data management system, there is a close upload and control relationship between each module and the platform. These relationships jointly ensure the efficient and stable operation of the system, as well as the effective management and application of data.

[0088] The multi-source data acquisition module and platform are responsible for uploading all types of driving data collected to the platform in real time for subsequent processing and analysis. The platform can send configuration instructions to the multi-source data acquisition module to meet the data requirements of different scenarios;

[0089] The processing module uploads the processed data to the platform. At the same time, the platform can send processing rules or algorithm update instructions to the processing module to optimize the data processing process and improve data quality.

[0090] The tiered decision module can upload its results to the platform. These results help the platform understand the vehicle's current status and provide a basis for subsequent management. The platform can also send update instructions to the tiered decision module and, based on the results, issue control instructions to other modules.

[0091] The dynamic fusion module is responsible for uploading the fused data to the platform, providing richer decision support. The platform can send fusion strategy and algorithm update instructions to the dynamic fusion module to optimize the data fusion effect. Furthermore, the platform can adjust the operating parameters of other modules based on the fused data.

[0092] The Improvement Module continuously improves system performance, algorithm accuracy, and other aspects based on platform data analysis results and user feedback. The platform can send improvement tasks or target instructions to the Improvement Module to clarify the direction and focus of improvement. Furthermore, the platform can evaluate and adjust the Improvement Module based on the effectiveness of the improvements.

[0093] The interactive feedback module is responsible for interacting with users and providing them with information such as platform processing results and notifications. The platform can send instructions such as feedback content and feedback method to the interactive feedback module to optimize the user interaction experience.

[0094] In the above-mentioned electric vehicle driving data management system, a hierarchical storage strategy is adopted through the hierarchical judgment module. According to the key parameters of the data, the data is divided into high-importance levels and low-importance levels, and a differentiated storage strategy is implemented to provide data support for the continuous operation and subsequent analysis of the system, thereby avoiding low-importance data occupying too much storage space. At the same time, by fusing multi-source data and using machine learning models to dynamically improve the cruising range, it can more accurately reflect the actual energy consumption of the vehicle and improve the accuracy of the cruising range prediction.

[0095] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. An electric vehicle driving data management system, characterized in that: It includes multi-source data acquisition module, processing module, classification determination module, dynamic fusion module, improvement module and interactive feedback module; The multi-source data acquisition module is used to collect various types of driving data of electric vehicles in real time; The processing module is used to receive and process massive amounts of various types of driving data; The hierarchical determination module is used to dynamically determine the data level and implement a hierarchical storage strategy to ensure that key data of various types of driving data are processed and stored with priority; The dynamic fusion module is used to integrate historical power consumption data, real-time driving behavior and environmental data; The improved module can use the comprehensive data set to adjust the range prediction to compensate for the long-term decline in battery capacity; The interactive feedback module is used to obtain and display various driving data of the vehicle in real time.

2. The electric vehicle driving data management system according to claim 1, characterized in that: The multi-source data acquisition module includes a sensor unit, a data reading unit and a data transmission unit. The sensor unit can collect vehicle driving data such as battery power, motor speed and vehicle speed in real time to form vehicle status data; The data reading unit is used to read historical power consumption values; The data transmission unit is responsible for quickly and accurately transmitting various types of driving data collected in real time to the processing module.

3. The electric vehicle driving data management system according to claim 1, characterized in that: The classification determination module includes a threshold setting unit, a model building unit, a determination unit and a storage unit; The threshold setting unit is used to preset thresholds according to key parameters during vehicle driving; The model building unit is capable of building a machine learning model using a label set and a random forest classifier; The determination unit can input key parameters in the real-time collected vehicle driving data into the rule engine. The rule engine matches the key parameters according to predefined rules and determines the level of the key parameters. The storage unit can execute a differentiated storage strategy according to the data level finally determined by the determination unit.

4. The electric vehicle driving data management system according to claim 3, characterized in that: The specific contents of the model building unit are as follows: a) Build a machine learning model based on the random forest classifier; b) Optimization of machine learning models; c) Introduction of rule engine.

5. The electric vehicle driving data management system according to claim 3, characterized in that: The threshold setting unit further includes a feature labeling unit, which can label key parameters in the historical driving data that are greater than a threshold value to form a labeling set.

6. The electric vehicle driving data management system according to claim 3, characterized in that: The key parameter levels are divided into high importance levels and low importance levels. The high importance level data is stored in full, and the low importance level data is stored in compressed form.

7. The electric vehicle driving data management system according to claim 1, characterized in that: The historical power consumption value data is extracted from the historical driving records of the vehicle, including 10-20 days of historical power consumption value data under different driving conditions.

8. The electric vehicle driving data management system according to claim 1, characterized in that: The specific contents of the improved module are as follows: a) Extracting comprehensive data sets; b) Learning dataset information; c) Construction of ARIMA model; d) Calculation of final prediction results.

9. The electric vehicle driving data management system according to claim 4, characterized in that: The introduction of the rule engine is used to define hierarchical rule conditions according to key parameters and thresholds during vehicle driving.

Images