Vehicle turning performance testing method

Abstract

The invention relates to the technical field of vehicle aided driving, in particular to a vehicle turning performance test method, which comprises the following steps: S1, data acquisition: a sensing unit acquires lane line data and dynamic data of a vehicle in real time, and the dynamic data of the vehicle comprises a driving speed V and a yaw velocity omega of the vehicle; s2, data processing: performing preprocessing, feature extraction, fitting and calculation on the collected data to obtain a curve preview radius and a dynamic turning radius, and calculating a curve fusion radius through a weighted fusion algorithm; and S3, dynamic feedback: transmitting a calculation result to a driving assistance system through a communication unit, adjusting a steering control strategy of the vehicle by the driving assistance system according to the curve fusion radius, and displaying the dynamic change of the curve radius in real time through a monitoring system. According to the method, the curvature parameters of the lane line are collected in real time and combined with the dynamic data of the vehicle, so that high-precision curve radius estimation is realized, and the performance and the safety of a driving assistance system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle assisted driving, and in particular to a method for testing the cornering performance of a vehicle. Background Art

[0002] A patent with the publication number CN117584982A discloses a method for estimating the corner radius. By combining the corner preview distance calculated based on the vehicle speed and the perceived lane line equation, the corner preview radius is estimated, and it is fused with the dynamic turning radius to obtain a more accurate corner fusion radius, thereby improving the accuracy and reliability of the corner radius estimation. This method includes corner preview distance calculation, lane line equation derivation, dynamic turning radius calculation, and multiple filtering processes. These calculations may require relatively high computing resources and have relatively high requirements for the hardware performance of the vehicle's electronic control unit (ECU) or in-vehicle system. Although methods such as mean filtering and variable parameter first-order low-pass filtering are introduced to smooth the corner radius data, these filtering algorithms may not be able to effectively process rapidly changing corner radii (such as sharp turns or continuous curves) in some cases. At the same time, there are also the following defects: (1) It mainly relies on subjective test evaluation, which lacks objectivity and consistency and is difficult to provide accurate data support; (2) There is a lack of a tool that can calculate the vehicle corner radius in real time and cannot provide immediate corner parameter support for the driving assistance system; (3) Although the vehicle's perception system can obtain the curvature parameters of the lane line, these data are not fully utilized to calculate the corner radius; (4) There is a lack of effective data interaction between the vehicle's perception system and the driving assistance system, resulting in the inability to fully utilize the perception data to optimize the driving assistance function; (5) There is a lack of visualization of real-time corner parameters, making it difficult to intuitively evaluate the vehicle's cornering performance; (6) When the vehicle is driving under different road conditions (such as lane line clarity, road curvature change), it is difficult to provide stable corner parameter calculation. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: In order to solve the problems of the limitations of subjective evaluation, lack of real-time corner parameter calculation, insufficient utilization of lane line data, data compatibility and system integration deficiencies, lack of real-time data visualization, etc. existing in the prior art in the above background art, a method for testing the cornering performance of a vehicle is provided. By calculating the corner radius in real time and providing the dynamic change law, it provides objective and accurate data support for the optimization of the driving assistance system, thereby improving the safety and comfort of the vehicle when driving on a curve.

[0004] The technical solution adopted by the present invention to solve its technical problems is: a method for testing the cornering performance of a vehicle, comprising the following steps: S1. Data acquisition: The sensing unit collects lane line data and dynamic data of the vehicle in real time. The dynamic data of the vehicle includes the driving speed V and the yaw angular velocity ω of the vehicle; S2. Data processing: The collected data is preprocessed, feature extracted, fitted, and calculated to obtain the preview radius of the curve and the dynamic turning radius, and then the fusion radius of the curve is calculated through a weighted fusion algorithm; S3. Dynamic feedback: The calculation result is transmitted to the driving assistance system through the communication unit. The driving assistance system adjusts the steering control strategy of the vehicle according to the fusion radius of the curve, and the dynamic change of the curve radius is displayed in real time through the monitoring system.

[0005] Data acquisition provides basic data support for subsequent data processing, and data processing provides accurate radius data for evaluating the cornering performance of the vehicle; dynamic feedback realizes the dynamic monitoring and optimization of the vehicle cornering performance.

[0006] According to an embodiment of the present invention, the sensing unit in the step S1 includes a camera, a lidar, and a millimeter wave radar. The camera, the lidar, and the millimeter wave radar work together to collect lane line data and dynamic data of the vehicle.

[0007] By the camera, the lidar, and the millimeter wave radar working together to collect lane line data and dynamic data of the vehicle, the comprehensiveness and accuracy of data acquisition are improved. The use of multiple sensors can complement and verify each other, improving the robustness and reliability of the system.

[0008] According to an embodiment of the present invention, data processing is performed on the color image collected by the camera, specifically: Preprocessing: Convert the color image collected by the camera into a grayscale image, remove image noise through Gaussian filtering or median filtering, and use the Canny edge detection algorithm or the Sobel operator to extract the edge information in the image; Feature extraction: Use the Hough transform to transform the edge points in the image into the parameter space, and detect straight lines through a voting mechanism to extract the lane lines; Lane line fitting: Fit the extracted lane line points into a quadratic polynomial curve to obtain the mathematical expression of the lane line: , where a, b, and c are the coefficients of the lane line equation.

[0009] Convert the color image to a grayscale image, and remove noise through Gaussian filtering or median filtering to improve the image quality; use the Canny edge detection algorithm or Sobel operator to extract the edge information in the image, providing basic data for lane line detection; detect straight lines through the Hough transform to extract the lane line features; fit the lane line points to a quadratic polynomial curve to obtain the mathematical expression of the lane line, providing a mathematical model for subsequent curvature calculation.

[0010] According to an embodiment of the present invention, the lidar obtains the three-dimensional point cloud data of the surrounding environment by emitting laser beams and measuring the time difference of the reflected light, and performs data processing on the point cloud data, specifically: Preprocessing: Downsample to reduce the data volume and filter and denoise to remove noise points; Feature extraction: Use the DBSCAN clustering algorithm to cluster the point cloud and extract the lane line point cloud; Lane line fitting: Fit the geometric model of the lane line through the RANSAC algorithm.

[0011] By downsampling and filtering and denoising, the data volume is reduced and noise points are removed, improving the quality of the point cloud data; using the DBSCAN clustering algorithm to cluster the point cloud to extract the lane line point cloud, providing data support for lane line fitting; fitting the geometric model of the lane line through the RANSAC algorithm to improve the robustness and accuracy of lane line fitting.

[0012] According to an embodiment of the present invention, the millimeter-wave radar detects the distance, speed and angle of the target object by transmitting and receiving millimeter-wave signals, and performs data processing on the reflected signal, specifically: Feature extraction: Analyze the intensity difference of the reflected signal and extract the boundary features of the lane line; Lane line fitting: Combine the angle information detected by the radar and the vehicle driving direction, and infer the position of the lane line through geometric interpolation.

[0013] By analyzing the intensity difference of the reflected signal of the millimeter-wave radar, the boundary features of the lane line are extracted, providing data support for lane line detection; combining the angle information detected by the radar and the vehicle driving direction, and inferring the position of the lane line through geometric interpolation to improve the accuracy of lane line detection.

[0014] According to an embodiment of the present invention, according to the coefficients of the lane line equation, first calculate the curvature K of the lane line, and the calculation formula is: ; Then calculate the preview radius R of the curve 预瞄 , and the calculation formula is: ; Next, calculate the dynamic radius R 动态, the calculation formula is: ; where V is the driving speed of the vehicle and ω is the yaw angular velocity; Finally, calculate the curve fusion radius R 融合 , the calculation formula is: ; where α is the weight coefficient, and the value range is from 0 to 1.

[0015] Calculate the curvature K of the lane line according to the coefficients of the lane line equation, providing basic data for the calculation of the curve preview radius; calculate the curve preview radius R through the curvature K 预瞄 , providing reference data for vehicle cornering performance evaluation; calculate the dynamic radius R according to the driving speed V and yaw angular velocity ω of the vehicle 动态 , reflecting the actual turning performance of the vehicle; calculate the curve fusion radius R through the weighted fusion algorithm 融合 , comprehensively considering the preview radius and the dynamic radius, improving the accuracy and practicality of the radius calculation.

[0016] According to an embodiment of the present invention, perform mean filtering or low-pass filtering on the curve preview radius R 预瞄 and the dynamic radius R 动态 . Perform mean filtering or low-pass filtering on the curve preview radius R 预瞄 and the dynamic radius R 动态 , removing noise and outliers, and improving the stability and reliability of the radius data.

[0017] According to an embodiment of the present invention, the communication unit in step S3 is a CAN bus. In the CAN bus, the curvature parameters of the lane line are encapsulated in specific messages; in the driving assistance system, the curvature parameter signals on the CAN bus are mapped to system variables through CAPL programming.

[0018] Encapsulate the curvature parameters of the lane line in specific messages through the CAN bus to achieve efficient data transmission; map the curvature parameter signals on the CAN bus to system variables through CAPL programming to provide data support for the driving assistance system.

[0019] According to an embodiment of the present invention, the lane line curvature parameters include the left lane curvature parameter Lane1_Curvature and the right lane curvature parameter Lane2_Curvature. Through CAPL programming, the left lane curvature parameter Lane1_Curvature and the right lane curvature parameter Lane2_Curvature are respectively mapped to HostLane1_C2 and HostLane2_C2, and calculate the average value of the curve radius of the vehicle center line. The calculation formula is as follows: 。

[0020] Map the left lane curvature parameter Lane1_Curvature and the right lane curvature parameter Lane2_Curvature to HostLane1_C2 and HostLane2_C2 respectively, providing data support for the calculation of the curve radius of the vehicle center line; calculate the average value of the curve radius of the vehicle center line to provide comprehensive curve radius data for the driving assistance system.

[0021] According to an embodiment of the present invention, in step S3, the monitoring system generates a dynamic change curve of the curve radius and feeds it back to the driving assistance system.

[0022] The monitoring system generates a dynamic change curve of the curve radius and feeds it back to the driving assistance system in real time, providing intuitive curve radius change information for the driver or the automatic driving system; through the feedback of the dynamic curve, the driving assistance system can adjust the control strategy in real time to optimize the cornering performance of the vehicle.

[0023] Advantages of the present invention: (1) By calculating the curve radius in real time, the subjective driving experience is converted into objective numerical data, providing an accurate quantitative index for the cornering performance of the vehicle; it helps to more scientifically evaluate the lateral control performance of the vehicle and reduce errors caused by subjective evaluation; (2) It can calculate and feedback the change law of the curve radius in real time, providing dynamic data of the vehicle during cornering; it provides instant curve information for the driving assistance system (such as adaptive cruise control, lane keeping assistance), improving the response speed and control accuracy of the system; (3) Through accurate calculation of the curve radius, the driving assistance system can more accurately adjust the vehicle speed and steering angle, avoiding accident risks caused by inaccurate curve information; significantly improving the driving safety of the vehicle under complex curve conditions, especially when driving at high speed or with poor visibility; (4) It provides more accurate curve parameters for the driving assistance system, enabling it to adjust the control strategy according to the real-time curve radius; improving the stability and comfort of the vehicle during cornering, reducing unnecessary acceleration / deceleration or steering adjustments; (5) By generating a dynamic curve of the curve radius, the software can intuitively display the parameter changes during the vehicle cornering process; providing an intuitive tool for engineers and testers to facilitate the analysis and optimization of the vehicle's driving assistance functions; (6) By processing the lane line curvature parameters, it can stably calculate the curve radius under different road conditions (such as lane line clarity, road curvature change); enhancing the adaptability of the vehicle under complex road conditions and improving the robustness of the driving assistance system; (7) By real-time calculation and analysis of the curve parameters, the dependence on expensive high-precision testing equipment is reduced; the costs in the vehicle development and testing phases are lowered, and the R & D efficiency is improved. (8) The driving assistance system can provide more accurate control according to the real-time curve parameters, reduce driving fatigue, and enhance the driving experience; it enhances users' trust and satisfaction with the driving assistance system and promotes the popularization of autonomous driving technology. (9) Through real-time data feedback, data support is provided for the vehicle control system, facilitating further optimization of the driving assistance algorithm; it provides a data-driven optimization method for vehicle manufacturers, contributing to the continuous improvement of vehicle performance. (10) Based on the general programming language (CAPL) and the development tool (Vector CAPL Browser), it has good compatibility and scalability; it is convenient to integrate with other driving assistance systems and supports the adaptation of different vehicle models and sensor configurations. Through the provision of real-time and accurate curve parameter calculation, the present invention significantly improves the performance and safety of the driving assistance system, while reducing the development cost and enhancing the user experience; these technical effects not only optimize the driving assistance function of the vehicle, but also provide important technical support for the development of autonomous driving technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The present invention will be further described below in conjunction with the drawings and embodiments.

[0025] Figure 1 is the flowchart of the present invention.

[0026] Figure 2 is the schematic structural diagram of the lane lines of the present invention.

[0027] Figure 3 is the schematic diagram of a basic steering input model in a specific embodiment of the present invention.

[0028] Figure 4 is the schematic diagram of the driving trajectory of the vehicle in a curve in a specific embodiment of the present invention.

[0029] Figure 5 is the schematic diagram of different driving states and stages of the vehicle in a curve in a specific embodiment of the present invention.

[0030] Figure 6 is the schematic diagram of the continuous path tracking of the vehicle in a curve in a specific embodiment of the present invention.

[0031] Figure 7 is the flowchart of the operation of the lane tracking or path planning module in the driving assistance system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0033] As Figure 1 shown, a vehicle cornering performance test method includes the following steps: S1. Data acquisition: The sensing unit collects lane line data and vehicle dynamic data in real time. The vehicle dynamic data includes the vehicle's driving speed V and yaw angular velocity ω. S2. Data processing: The collected data is preprocessed, feature extracted, fitted, and calculated to obtain the corner preview radius and dynamic turning radius, and then the corner fusion radius is calculated through a weighted fusion algorithm. S3. Dynamic feedback: The calculation results are transmitted to the driving assistance system through the communication unit. The driving assistance system adjusts the vehicle's steering control strategy according to the corner fusion radius, and the dynamic change of the corner radius is displayed in real time through the monitoring system.

[0034] In step S1, the sensing unit includes a camera, a lidar, and a millimeter radar. The camera, lidar, and millimeter radar work together to collect lane line data and vehicle dynamic data.

[0035] The data processing of the color image collected by the camera is as follows: Preprocessing: The color image collected by the camera is converted into a grayscale image to reduce the amount of data and facilitate subsequent processing; Gaussian filtering or median filtering is used to remove image noise and improve image quality; the Canny edge detection algorithm or Sobel operator is used to extract the edge information in the image, providing a basis for the extraction of lane lines.

[0036] Feature extraction: The Hough transform is used to transform the edge points in the image into the parameter space, and lines are detected through a voting mechanism to extract lane lines. In addition, deep learning methods (such as convolutional neural networks CNN, including network structures such as U-Net and SegNet) can be used to perform semantic segmentation on the image to directly identify the lane line area. For the color features of the lane lines (such as white or yellow), the lane line area is extracted by setting color thresholds.

[0037] Lane line fitting: The extracted lane line points are fitted into a quadratic polynomial curve to obtain the mathematical expression of the lane line: , where a, b, and c are the coefficients of the lane line equation, and the curvature of the lane line can be calculated through these coefficients.

[0038] The lidar emits laser beams and measures the time difference of the reflected light to obtain the three-dimensional point cloud data of the surrounding environment, and performs data processing on the point cloud data. Specifically: Preprocessing: Downsampling to reduce the data volume and filtering to remove noise points; Feature extraction: Using the DBSCAN clustering algorithm to cluster the point cloud and extract the lane line point cloud; Lane line fitting: Fitting the geometric model (such as a straight line or a curve) of the lane line through the RANSAC algorithm (Random Sample Consensus).

[0039] The millimeter-wave radar emits and receives millimeter-wave signals to detect the distance, speed, and angle of the target object, and performs data processing on the reflected signals. Specifically: Feature extraction: Analyzing the intensity difference of the reflected signals to extract the boundary features of the lane line; Lane line fitting: Combining the angle information detected by the radar and the vehicle driving direction, and inferring the position of the lane line through geometric interpolation.

[0040] To further improve the accuracy and robustness of lane line detection, the detection results of the camera, lidar, and millimeter-wave radar are fused. For example, using the image information of the camera and the three-dimensional point cloud information of the lidar, through data fusion algorithms such as Kalman filtering or particle filtering, more accurate lane line information is extracted; then, combining the visual information of the camera and the target detection information of the millimeter-wave radar to improve the reliability of lane line detection.

[0041] Among them, Kalman filtering is a recursive filtering method based on a linear system, which can effectively handle the problems of noise and outliers in lane line detection. In specific implementation, first establish a mathematical model of the lane line, usually representing the lane line as a polynomial curve (such as a quadratic or cubic curve), and initialize the state vector and covariance matrix; then, according to the state of the previous moment and the system model, predict the state of the current moment; then, obtain the lane line observation data of the current frame, calculate the Kalman gain, and update the state vector and covariance matrix using the observation data; finally, generate the lane line equation of the current moment according to the updated state vector.

[0042] According to the coefficients of the lane line equation, first calculate the curvature K of the lane line. The calculation formula is: ; Then calculate the preview radius R of the curve 预瞄 , and the calculation formula is: ; Next, calculate the dynamic radius R 动态 , and the calculation formula is: ; Where, V is the driving speed of the vehicle, and ω is the yaw angular velocity; Finally, calculate the curve fusion radius R 融合 , and the calculation formula is: ; Where, α is the weight coefficient, and its value range is from 0 to 1. Of course, it can also be adjusted according to the actual situation.

[0043] To improve the estimation accuracy, filter the preview radius R 预瞄 and the dynamic radius R 动态 , such as mean filtering, low-pass filtering, etc., to reduce the influence of noise and outliers; through optimizing algorithms and hardware acceleration (such as using a high-performance ECU or in-vehicle system), ensure that the calculation of the curve fusion radius R 融合 can be completed in real time to meet the requirements of the driving assistance system.

[0044] In addition, the calculation of the vehicle speed and the curve preview distance is to obtain the driving speed V of the vehicle in real time through the wheel speed sensor of the vehicle. According to the driving speed V of the vehicle and the preview time t (usually set according to the response time and driving strategy of the vehicle), calculate the distance that the vehicle can travel within a certain time, that is, the preview distance D 预瞄 The calculation formula is: .

[0045] The communication unit in step S3 is the CAN bus. In the CAN bus, the curvature parameters of the lane lines are encapsulated in specific messages, as shown in Table 1. The ZDC_Lane_1 and ZDC_Lane_2 messages respectively contain the curvature parameters Lane1_Curvature and Lane2_Curvature of the left and right lane lines. In the driving assistance system, map the curvature parameter signals on the CAN bus to system variables through CAPL programming, and map the left lane curvature parameter Lane1_Curvature and the right lane curvature parameter Lane2_Curvature to HostLane1_C2 and HostLane2_C2 respectively for subsequent calculation of the curve radius.

[0046] Table 1

[0047] Since the signals obtained from the perception end cannot be calculated, variables for the left and right lane lines are respectively set, and are assigned to the system variables through data transfer, and the required values are obtained through the calculation of the variable values. In the system lane change, variables for the left and right lane lines are added at the same time, and the average value R of the vehicle curve radius is obtained through calculation. The calculation formula is as follows: , The specific calculation formula after substituting the perceived curvature is as follows: 。

[0048] Among them, R1 represents the first lane line on the left side of the vehicle center, and R2 represents the first lane line on the right side of the vehicle center. When (R1 + R2) / 2, the average radius of the current self-lane curve can be obtained, which is convenient for perceiving the radius of the vehicle passing through the current curve and objectively judging whether the performance of the curve meets the requirements. For example Figure 2 As shown. In modern automotive technology, the performance optimization of the driving assistance system is crucial for improving driving safety and driving experience. The vehicle cornering performance test method of this embodiment provides accurate corner parameters for the driving assistance system by calculating the curve radius in real time, thereby optimizing the performance and safety of the vehicle when driving through curves; during the vehicle driving process, especially when driving through curves, the driving assistance system (such as lane keeping assistance, adaptive cruise control, etc.) requires accurate corner parameters to optimize the control strategy.

[0049] Traditional methods for evaluating corner parameters often rely on subjective evaluations and lack objectivity and real-time performance. To overcome this limitation, a method for calculating the curve radius based on data collected by a sensing unit provides objective, accurate, and real-time corner parameters for the driving assistance system by calculating the curve radius in real time. The calculation results are fed back to the driving assistance system (such as ADAS) via the CAN bus for optimizing its control strategy; a dynamic curve of the curve radius is generated through a monitoring system (such as CANOE) to visually display the change in the curve radius during the vehicle cornering process. Drivers and testers can observe the performance of the vehicle under different corner conditions in real time, thereby promptly discovering problems and making optimizations.

[0050] Specific embodiment: As Figure 3 shown, a basic steering input model, where fi in the figure represents the front wheel angle of the vehicle, d represents the lateral displacement of the vehicle or the steering target point, and the dashed line represents the ideal path or target path of the vehicle. As Figure 4 shown, the driving trajectory of the vehicle in a curve, where ci and fi in the figure represent the curvature of the driving path and the front wheel angle of the vehicle respectively, the arrow in the figure represents the driving direction of the vehicle, and the curve represents the expected path of the vehicle in the curve. As Figure 5 shown, different driving states or stages of the vehicle in a curve, where the red curve in the figure represents different positions or driving of the vehicle in the curve. As Figure 6 shown, the continuous path tracking of the vehicle in a curve, where the yellow arrow in the figure represents the steering input or control command of the vehicle, and the red dot represents the key position or state of the vehicle in the curve.

[0051] As Figure 7As shown, first, it is determined whether rodo enters this module normally, and the historical frame rodo is retained. This involves the processing and analysis of sensor data (such as cameras, radars, lidars, etc.) to determine whether the vehicle is on the expected lane; the state of the current frame is predicted based on the state of the previous frame, which usually involves using the vehicle's dynamic model and sensor data to predict the position and attitude of the vehicle at a future moment. Secondly, the main lane is searched for and the road boundary information is updated, which involves detecting the lane where the vehicle is currently located and tracking the lane boundaries, which is crucial for keeping the vehicle driving stably within the lane; then, the camera pitch angle is calculated through the vehicle position information, which involves the precise positioning of the vehicle, including using GPS, IMU (Inertial Measurement Unit) and other sensor data to determine the position of the vehicle in the global coordinate system; then, the stop line information is updated, corrected according to the observations, the lane line output module and the validity of the Roadmodel are updated, which involves the vehicle's understanding of the surrounding environment and the construction of the map, as well as the update of the map and model according to the real-time observation data. Although path planning is not explicitly mentioned in the flowchart, these steps are the basis for path planning because path planning relies on the understanding of the vehicle's current position, lane information and road environment.

[0052] By calculating the curve preview radius and the dynamic turning radius and fusing the two to obtain the curve fusion radius, the driving trajectory of the vehicle in a curve can be predicted more accurately. This enables the driving assistance system to make adaptive adjustments in advance, such as decelerating and adjusting the steering angle, thereby reducing the accident risk caused by inaccurate curve information. The curve radius data provides an important basis for optimizing the control strategy of the driving assistance system. For example, the lane keeping assistance system can adjust the lateral control of the vehicle according to the real-time curve radius, and the adaptive cruise control system can adjust the vehicle speed according to the curve radius, so as to achieve more intelligent and safer driving assistance functions. Through real-time feedback and precise control, the driving assistance system can better handle complex curve conditions, reduce the operation burden of the driver when driving in a curve, and improve driving comfort and convenience. Based on the general CAPL programming language and development tools (such as Vector CAPL Browser), it has good compatibility and can adapt to different vehicle models and sensor configurations; this means that it can be widely applied to various vehicle platforms without a large amount of customized development for each vehicle model. System variables can be customized according to the needs of the simulation project, and users can add, edit, and delete system variables according to actual needs to meet the test and optimization needs in different scenarios. The dynamic curve display facilitates engineers and testers to quickly analyze the cornering performance of the vehicle, discover potential problems and optimize them, record the curve radius data during the vehicle cornering process, and provide data support for subsequent detailed analysis and performance evaluation. These data can be used to develop more advanced driving assistance algorithms and further improve the intelligent level of the vehicle. By optimizing the control strategy of the driving assistance system, the instability and driving fatigue of the vehicle when driving in a curve can be reduced, and the driving experience of users can be improved. For example, the vehicle can decelerate and steer more smoothly in a curve, reducing unnecessary acceleration and deceleration operations, making the driving process more comfortable. The provided accurate curve radius data and optimized driving assistance functions can enhance users' trust in the driving assistance system. Users are more willing to use these functions, thus promoting the popularization and development of autonomous driving technology. By real-time sensing the curvature parameters of the lane lines, the calculation method can be dynamically adjusted to ensure accurate curve radius estimation under various complex road conditions, and it can adapt to different types of curves and road conditions, including urban roads, highways, rural roads, etc. By using a filtering algorithm to reduce the influence of noise and outliers, the stability and reliability of the system under harsh environmental conditions are improved.

[0053] The vehicle cornering performance test method of this embodiment realizes high-precision estimation of the corner radius by collecting lane curvature parameters in real time and combining vehicle dynamic data. It not only provides objective and real-time corner parameters, significantly improving the performance and safety of the driving assistance system, but also has good compatibility, scalability and cost-effectiveness. These advantages not only optimize the driving experience, but also provide important support for the development of intelligent driving technology. Compared with traditional methods, it does not rely on high-cost and poor anti-interference high-precision perception devices, but realizes accurate estimation on low-cost devices through optimized algorithms and data processing, reducing the overall cost of the system. At the same time, it also provides a complete development environment and tool support, simplifies the development process, enables developers to quickly implement and verify new algorithms, reduces development time and resource investment, and reduces R & D costs.

[0054] Inspired by the above ideal embodiments according to the present invention, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A vehicle cornering performance test method, characterized in that, It includes the following steps: S1. Data acquisition: The sensing unit collects lane line data and vehicle dynamic data in real time. The vehicle dynamic data includes the driving speed V and yaw angular velocity ω of the vehicle. S2. Data processing: The collected data is preprocessed, feature extracted, fitted, and calculated to obtain the curve preview radius and dynamic turning radius, and then the curve fusion radius is calculated through a weighted fusion algorithm. S3. Dynamic feedback: The calculation result is transmitted to the driving assistance system through the communication unit. The driving assistance system adjusts the vehicle steering control strategy according to the curve fusion radius and displays the dynamic change of the curve radius in real time through the monitoring system.

2. The vehicle cornering performance test method according to claim 1, characterized in that: In step S1, the sensing unit includes a camera, a lidar, and a millimeter-wave radar. The camera, lidar, and millimeter-wave radar work together to collect lane line data and vehicle dynamic data.

3. The vehicle cornering performance test method according to claim 2, characterized in that: Data processing is performed on the color image collected by the camera. Specifically: Preprocessing: The color image collected by the camera is converted into a grayscale image. Gaussian filtering or median filtering is used to remove image noise, and the Canny edge detection algorithm or Sobel operator is used to extract the edge information in the image. Feature extraction: The Hough transform is used to convert the edge points in the image to the parameter space, and a straight line is detected through a voting mechanism to extract the lane line. Lane line fitting: The extracted lane line points are fitted into a quadratic polynomial curve to obtain the mathematical expression of the lane line: ， where a, b, and c are the coefficients of the lane line equation.

4. The vehicle cornering performance test method according to claim 2, wherein: The lidar obtains the three-dimensional point cloud data of the surrounding environment by emitting laser beams and measuring the time difference of the reflected light. Data processing is performed on the point cloud data. Specifically: Preprocessing: Downsampling is performed to reduce the data volume, and filtering and denoising are performed to remove noise points. Feature extraction: The DBSCAN clustering algorithm is used to cluster the point cloud to extract the lane line point cloud. Lane line fitting: The geometric model of the lane line is fitted through the RANSAC algorithm.

5. The vehicle cornering performance test method according to claim 4, characterized in that: The millimeter-wave radar detects the distance, speed, and angle of the target object by emitting and receiving millimeter-wave signals. Data processing is performed on the reflected signal. Specifically: Feature extraction: The intensity difference of the reflected signal is analyzed to extract the boundary features of the lane line. Lane line fitting: Combining the angle information detected by the radar and the vehicle driving direction, the position of the lane line is inferred through geometric interpolation.

6. The vehicle cornering performance test method according to claim 3, wherein: According to the coefficients of the lane line equation, first calculate the curvature K of the lane line. The calculation formula is: ； Recalculate the curve preview radius R 预瞄 , and the calculation formula is as follows: ； Next, calculate the dynamic radius R 动态 , and the calculation formula is as follows: ； where V is the driving speed of the vehicle and ω is the yaw angular velocity. Finally, calculate the curve fusion radius R 融合 , and the calculation formula is as follows: ； where α is the weight coefficient, and its value range is from 0 to 1.

7. The vehicle cornering performance test method according to claim 6, characterized in that: For the curve preview radius R 预瞄 and the dynamic radius R 动态 Perform mean filtering or low-pass filtering.

8. The vehicle cornering performance test method according to claim 1, characterized in that: The communication unit in step S3 is the CAN bus. In the CAN bus, the curvature parameter of the lane line is encapsulated in a specific message; in the driving assistance system, the curvature parameter signal on the CAN bus is mapped to the system variable through CAPL programming.

9. The vehicle cornering performance test method according to claim 8, characterized in that: The lane line curvature parameters include the left lane curvature parameter Lane1_Curvature and the right lane curvature parameter Lane2_Curvature. Through CAPL programming, the left lane curvature parameter Lane1_Curvature and the right lane curvature parameter Lane2_Curvature are respectively mapped to HostLane1_C2 and HostLane2_C2, and the average value of the curve radius of the vehicle center line is calculated. The calculation formula is as follows: 。 10. The vehicle cornering performance test method according to claim 9, wherein: In step S3, the monitoring system generates a dynamic change curve of the curve radius and feeds it back to the driving assistance system.

Images