Unmanned vehicle calibration method and system
By constructing a longitudinal control mathematical model and updating the on-board calibration table, the problem of unmanned vehicles not performing longitudinal control calibration during use is solved, automatic longitudinal control calibration is realized, and the stability and efficiency of vehicle performance are improved.
Patent Information
- Application Number
- CN202411523886.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-30
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing unmanned vehicles do not perform longitudinal control calibration during use, resulting in the degradation of vehicle performance over time, and automated calibration and optimization cannot be achieved.
By collecting vehicle historical driving data, building a longitudinal control mathematical model, generating new longitudinal control calibration data, and updating the on-board calibration table, it realizes automatic longitudinal control calibration of unmanned vehicles during use.
It realizes automatic longitudinal calibration of unmanned vehicles during use, improves the stability and efficiency of vehicle performance, and ensures that the vehicle always maintains the optimal operating state during long-term use.
Smart Images

Figure CN119024820A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of unmanned driving, and specifically relates to an unmanned driving vehicle calibration method and system. Background Art
[0002] Autonomous driving is a key area of artificial intelligence implementation supported by my country. Currently, autonomous driving is divided into L0-L5 levels, among which L4 autonomous driving vehicles mainly refer to vehicles that are completely independent of human driving in fixed scenes. In the existing technology, the longitudinal calibration of autonomous driving vehicles is only performed once after the vehicle is assembled, and the vehicle no longer actively performs longitudinal control calibration during use. Summary of the invention
[0003] The invention discloses an unmanned vehicle calibration method and system, which can realize the automatic completion of longitudinal control calibration of the unmanned vehicle during use.
[0004] In order to solve the above problems, the present invention provides a method for calibrating an unmanned vehicle, the method comprising: Collect historical vehicle driving data to build a longitudinal control mathematical model. The vehicle historical driving data includes: vehicle drive system data, vehicle brake system data, vehicle body system data, vehicle environment data, vehicle usage time data, and vehicle type data; the data needs to be checked for abnormal values, filled with missing values, and normalized. The longitudinal control mathematical model includes: a motor torque prediction model and a brake oil pressure prediction model; different longitudinal control mathematical models are constructed for different types of vehicles.
[0005] Based on the longitudinal control mathematical model, when the use of the unmanned vehicle exceeds a certain cycle threshold, new longitudinal control calibration data is generated from the driving data of the unmanned vehicle within a certain period of time.
[0006] The unmanned vehicle onboard calibration table is updated. The new longitudinal control calibration data generated includes: a speed, acceleration and motor torque calibration table; a speed, acceleration and brake oil pressure calibration table.
[0007] On the other hand, the present invention provides an unmanned vehicle calibration system, which includes: a data acquisition module: configured to obtain data during vehicle driving and mark it as historical data; a data preprocessing module: configured to receive, preprocess and forward the vehicle historical driving data; a data storage module: configured to classify and store the forwarded vehicle historical driving data; a data analysis module: configured to use the data from the data storage module to construct a longitudinal control mathematical model and predict vehicle longitudinal control calibration data; a data application module: configured to monitor the vehicle, store the predicted vehicle longitudinal control calibration data and update the on-board calibration table. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] In order to more clearly illustrate the embodiments and technical solutions of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0009] Figure 1 This is a flow chart of the unmanned vehicle calibration method of the present invention.
[0010] Figure 2 This is a diagram of the architecture of the unmanned vehicle calibration system of the present invention.
[0011] Figure 3 This is a flow chart of data collection and processing for the unmanned vehicle calibration system of the present invention.
[0012] Figure 4 This is a flow chart of the automatic vehicle calibration of the unmanned vehicle calibration system of the present invention. DETAILED DESCRIPTION
[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0014] See also Figure 1 , an embodiment of the present invention discloses a method for calibrating an unmanned vehicle, comprising the following steps.
[0015] S101: Collecting historical vehicle driving data to build a longitudinal control mathematical model. The historical vehicle driving data includes: vehicle driving system data, vehicle braking system data, vehicle body system data, vehicle environment data, vehicle use time data, and vehicle type data.
[0016] Optionally, the vehicle drive system data includes: vehicle speed, acceleration, battery voltage, battery current, motor torque, motor speed, etc.
[0017] Optionally, the vehicle braking system data includes: brake oil pump pressure and oil pump cross-sectional area.
[0018] Optionally, the vehicle body system data includes: vehicle outline dimensions, wheelbase, track width, total mass, and curb weight.
[0019] Optionally, vehicle environment data include: road slope angle, road friction coefficient, and ambient temperature.
[0020] Optionally, the vehicle usage time data includes: factory time, operating time, and standby time.
[0021] Optional vehicle type data: vehicle model, motor specifications, brake system specifications.
[0022] In S101, data collection also includes data preprocessing to extract complete and reliable feature data, specifically including: outlier check: select obviously inaccurate data from the data and process them accordingly. Optional checking methods include: size check, median check; optional processing methods include discarding or filling in missing values.
[0023] Missing value filling: Process the missing data collected. Optional processing methods include: median filling, mean filling, and discarding. Data normalization: Standardize the data in the same dimension. Optional processing methods include: standard normalization and benchmark normalization.
[0024] In S101, the mathematical methods for constructing the longitudinal control mathematical model optionally include: least square method and gradient descent method based on fitting; regression algorithm based on linear regression, logistic regression and support vector machine; integrated learning algorithm based on GBboost and Xgboost.
[0025] In S101, the longitudinal control mathematical model first needs to be constructed for different types of vehicles and usage time. When a certain type of vehicle has been driving for a certain time, calling the model of the corresponding time and type can improve the prediction accuracy. The optional longitudinal control data model can predict the motor torque and brake oil pressure.
[0026] S102: Based on the longitudinal control mathematical model, when the use of the unmanned vehicle exceeds a certain cycle threshold, the driving data of the unmanned vehicle in a certain period of time in the recent period is used to generate new longitudinal control calibration data. Optionally, the calibration period is based on a natural time period or a vehicle use cycle. Optionally, the driving data of the unmanned vehicle in a certain period of time in the recent period of time can be based on a natural time period or a vehicle use cycle.
[0027] S103: Update the onboard calibration table of the unmanned vehicle. The generation of new longitudinal control calibration data includes: a calibration table of speed, acceleration and motor torque; a calibration table of speed, acceleration and brake oil pressure. The calibration table format can also be an acceleration and throttle pedal opening and closing degree calibration table; an acceleration and brake pedal opening and closing degree calibration table. The method of optionally updating the onboard calibration table includes remote upgrade and local upgrade.
[0028] On the other hand, see Figure 2This example discloses an unmanned vehicle calibration system, specifically including: a data acquisition module: configured to obtain data during vehicle driving and mark it as historical data; optionally, the data acquisition module can be a vehicle-mounted module or an independent module outside the vehicle, and obtain vehicle driving data through wired or wireless methods. It should be noted that due to the different internal data networks of the vehicle, the data acquisition module can obtain data through the vehicle CAN bus or through the vehicle-side data gateway.
[0029] Data preprocessing module: configured to receive, preprocess and forward the vehicle historical driving data.
[0030] Data storage module: configured to classify and store the forwarded vehicle historical driving data; optionally, the classified storage can be stored according to vehicle type and usage time; the storage method can adopt a relational database or a transactional database.
[0031] Data analysis module: configured to use the data stored in the data storage module to construct a longitudinal control mathematical model and predict the longitudinal control calibration data of the vehicle; one function of this module is to construct a longitudinal data model. In this embodiment, the longitudinal control module is iterated regularly through the data acquisition module and the preprocessing module, and the model is optionally constructed according to conditions such as vehicle usage status and manual intervention. Another function is to predict the longitudinal control calibration data of the vehicle. In this embodiment, the prediction model is triggered by monitoring the usage time of the vehicle exceeding a certain period threshold, and the longitudinal calibration data of the vehicle is optionally regularly predicted through natural time.
[0032] Data application module: configured to monitor the vehicle, store the predicted vehicle longitudinal control calibration data and update the vehicle calibration table. The data application module monitors whether the vehicle meets the calibration conditions, calls the corresponding longitudinal control model to predict the longitudinal control calibration of the vehicle according to the vehicle type and usage time, and then generates the vehicle calibration table based on the prediction results, and finally completes the vehicle automatic calibration by remotely or locally updating the vehicle calibration table.
[0033] exist Figure 2 The data acquisition module includes a data acquisition unit and a protocol conversion unit.
[0034] Data collection unit: collects historical vehicle driving data. CAN, 5G and other communication methods can be used optionally.
[0035] Protocol conversion unit: convert and send vehicle historical driving data, receive predicted calibration data and update the vehicle calibration table. Since the data format of the vehicle historical driving data is different from the network transmission format, protocol conversion is required. Optionally, the vehicle-side CAN protocol can be converted to the wireless transmission MQTT and TCP protocols. At the same time, when updating the vehicle calibration table, the network data is converted to the vehicle-side data format.
[0036] exist Figure 2 The data preprocessing module includes a data preprocessing unit and a data forwarding unit.
[0037] Data preprocessing unit: configured to receive vehicle historical driving data and perform abnormal value checking, missing value filling, and normalization processing. Optionally, the data preprocessing unit completes Figure 1 The data preprocessing work in step S101 can also complete other preprocessing work for model building.
[0038] Data forwarding unit: sends pre-processed vehicle driving data. The pre-processed vehicle data is forwarded to the data storage module for model building and prediction, and also sent to the data analysis module for monitoring vehicles.
[0039] exist Figure 2 The data storage module includes: data warehouse layer and data application layer.
[0040] Data warehouse layer: configured to format and classify the storage of vehicle drive system data, vehicle brake system data, vehicle body system data, vehicle environment data, vehicle usage time data, and vehicle type data.
[0041] Data application layer: configured as specific vehicle driving data of different vehicle types. The data is divided into different types in the data application layer to support the construction of different types of longitudinal control models.
[0042] exist Figure 2 The data analysis module includes: data model unit and data analysis unit.
[0043] Data model unit: constructs a longitudinal control mathematical model based on the historical driving data of the vehicle. Constructs different longitudinal control mathematical models based on vehicle data of different types and time periods.
[0044] Data analysis unit: generates new longitudinal control calibration data based on the recent driving data of the vehicle within a certain period of time. The data analysis unit is based on the longitudinal control model of the data model unit, inputs the driving data of the calibrated vehicle within a certain period of time, and outputs new longitudinal calibration data.
[0045] exist Figure 2 The data application modules include: vehicle monitoring unit, calibration unit and vehicle update unit.
[0046] Vehicle monitoring unit: monitors in real time whether the vehicle has reached the calibration cycle. Through the real-time data of the data preprocessing module, monitors in real time whether the vehicle has reached the calibration threshold.
[0047] Calibration table unit: stores predicted vehicle longitudinal control calibration data.
[0048] Vehicle update unit: configured to remotely update the on-board calibration table. The predicted vehicle longitudinal control calibration data of the calibration table unit is updated to the on-board calibration table through the protocol conversion unit of the data acquisition unit.
[0049] See also Figure 3 , is an embodiment of an unmanned vehicle calibration method and system provided by the present invention, specifically including data collection and model construction.
[0050] S301: The data acquisition unit collects historical driving data of the unmanned vehicle bus and controller, specifically including vehicle driving data, vehicle braking data, vehicle body data, vehicle environment data, and vehicle type data.
[0051] S302: The protocol conversion unit converts the data in the vehicle-side format into data in the network format and sends it.
[0052] S303: The preprocessing module performs outlier check, missing data processing, normalization processing on the vehicle data, and forwards it.
[0053] S304: The data storage module stores the pre-processed data according to different vehicle types.
[0054] S305: The data analysis module constructs corresponding longitudinal control data models according to the data of different vehicle types.
[0055] See also Figure 4 , is an embodiment of an unmanned vehicle calibration method and system provided by the present invention, specifically including new calibration data generation and update.
[0056] S401: The data application module monitors the vehicle status in real time and identifies vehicles that exceed the calibration cycle threshold.
[0057] S402: The data application module retrieves the recent driving data of the vehicle that exceeds the calibration cycle threshold from the data storage module.
[0058] S403: The data application module calls the longitudinal control data model corresponding to the calibration cycle threshold vehicle for prediction.
[0059] S404: The data application module generates a vehicle calibration table using the predicted calibration data of the vehicle.
[0060] S405: The data application module updates the newly generated vehicle calibration table to the vehicle that exceeds the calibration period threshold, thereby completing automatic vehicle calibration.
[0061] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0062] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for calibrating an unmanned vehicle, It is characterized in that The method comprises: Collect historical vehicle driving data to build a longitudinal control mathematical model; Based on the longitudinal control mathematical model, when the use of the unmanned vehicle exceeds a certain cycle threshold, the driving data of the unmanned vehicle within a certain period of time in the recent period is used to generate new longitudinal control calibration data; Update the on-board calibration table for unmanned vehicles.
2. According to the unmanned vehicle calibration method of claim 1, It is characterized in that The collecting of vehicle historical driving data includes: Vehicle drive system data, vehicle brake system data, vehicle body system data, vehicle environment data, vehicle usage time data, and vehicle type data; The data collection process includes outlier checking, missing value filling, and data normalization.
3. According to the unmanned vehicle calibration method of claim 1, It is characterized in that The longitudinal control mathematical model includes: Motor torque prediction model, brake oil pressure prediction model; Different longitudinal control mathematical models are constructed for vehicles of different types and usage times.
4. According to the unmanned vehicle calibration method of claim 1, It is characterized in that The generating of new longitudinal control calibration data comprises: Speed, acceleration and motor torque calibration table; Speed, acceleration and brake oil pressure calibration table.
5. An unmanned vehicle calibration system using the unmanned vehicle calibration method according to any one of claims 1 to 4, It is characterized in that The system comprises: Data acquisition module: configured to obtain data during vehicle driving and mark it as historical data; Data preprocessing module: configured to receive, preprocess and forward the vehicle historical driving data; Data storage module: configured to classify and store the forwarded vehicle historical driving data; Data analysis module: configured to construct a longitudinal control mathematical model and predict vehicle longitudinal control calibration data using the data of the data storage module; Data application module: configured to monitor the vehicle, store the predicted vehicle longitudinal control calibration data and update the on-board calibration table.
6. According to claim 5, an unmanned vehicle calibration system, It is characterized in that The data acquisition module comprises: Data collection unit: collects historical vehicle driving data; Protocol conversion unit: converts and sends vehicle historical driving data, receives predicted calibration data and updates the on-board calibration table.
7. The unmanned vehicle calibration system according to claim 5, It is characterized in that The data preprocessing module comprises: Data preprocessing unit: configured to receive vehicle historical driving data and perform outlier checking, missing value filling, and normalization processing; Data forwarding unit: sends pre-processed vehicle driving data.
8. The unmanned vehicle calibration system according to claim 5, It is characterized in that The data storage module comprises: Data warehouse layer: configured to format and classify and store vehicle drive system data, vehicle brake system data, vehicle body system data, vehicle environment data, vehicle usage time data, and vehicle type data; Data application layer: configured as specific vehicle driving data for different vehicle types.
9. The unmanned vehicle calibration system according to claim 5, It is characterized in that The data analysis module includes: Data model unit: constructs a longitudinal control mathematical model based on the vehicle's historical driving data; Data analysis unit: Generates new longitudinal control calibration data based on the vehicle's recent driving data within a certain period of time.
10. The unmanned vehicle calibration system according to claim 5, It is characterized in that The data application module includes: Vehicle monitoring unit: monitors in real time whether the vehicle has reached the calibration cycle; Calibration table unit: stores predicted vehicle longitudinal control calibration data; Vehicle Update Unit: Configured to remotely update the on-board calibration table.
Citation Information
Patent Citations
Method and system for calibrating endurance mileage of electric automobile
CN109606119A
Automatic calibration method for road axle load spectrum detecting system
CN109696232A
Vehicle control parameter determination method and device, vehicle-mounted controller and unmanned vehicle
CN110967991A
Method for predictive maintenance of electric vehicle battery based on big data machine learning
CN114529067A
