A trajectory tracking optimization method for automatic parking process

Abstract

The present application relates to the technical field of intelligent driving, and particularly relates to a trajectory tracking optimization method for an automatic parking process.The present application specifically comprises, on the basis of an established parking trajectory, splitting the parking trajectory into a multi-segment line combination based on a straight line trajectory and a circular arc trajectory, and establishing a multi-controller segmented control method for the multi-segment line combination, to maintain the uniformity of the driving speed of the vehicle in the path tracking process.The trajectory tracking optimization method for the automatic parking process according to the present application, on the basis of a planned automatic parking path, establishes a vehicle dynamics model based on a specified vehicle, thereby avoiding the problems of a large path tracking deviation and vehicle operation instability caused by the non-uniformity of the vehicle speed in the vehicle operation parameters during the vehicle path tracking process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent driving, and the IPC classification number is B60W30 / 06, and particularly relates to a trajectory tracking optimization method for an automatic parking process. BACKGROUND

[0002] With the development of intelligent manufacturing, automatic driving and auxiliary driving as a fast and efficient driving mode have been gradually applied to people's daily life. Among them, automatic parking as a function with a relatively wide application range in automatic driving better helps the driving personnel to relieve the parking difficulty problem in actual use. However, at present, after the path planning based on automatic parking is completed, due to the running error of the mechanical structure of the vehicle body itself, the observation error of the sensor and the environmental error and other problems, there are often large tracking path deviations and inaccurate tracking problems in the actual automatic parking path tracking. Especially in the mixed straight line and curve multi-end line path tracking, and in the position of the corner with large curvature, due to the poor uniformity of the tracking path transformation, the vehicle is prone to unstable driving, and even the problem of vehicle running deviation from the trajectory occurs, which affects the safety of vehicle driving.

[0003] The patent CN202011236830 provides an automatic parking method, device, equipment and storage medium, and an automatic parking condition evaluation system is established in the patent to determine whether it is suitable to perform automatic parking operation under the current state. The patent CN201710909016 provides a voice-controlled automatic parking triggering system and method, and in the patent, the display control sequence of the vehicle-mounted controller is adjusted, and when the automatic parking function is started, the automatic parking control program is preferentially played, and the quickness of the automatic parking control is improved.

[0004] However, the above-mentioned patent only makes corresponding improvements and designs on the environment and communication conditions involved in the automatic parking process, and does not involve the related research on the parking accuracy, parking efficiency and parking control method actually involved in the automatic parking process, and does not better solve the problems existing in the operation and execution of the vehicle itself in the automatic parking process, therefore, in view of the problem, the present application provides a trajectory tracking optimization method for an automatic parking process. SUMMARY

[0005] In view of the above-mentioned problems, the present application provides a trajectory tracking optimization method for an automatic parking process, which specifically comprises: on the basis of having established a parking trajectory, the parking trajectory is divided into a multi-segment line combination based on a straight line trajectory and a circular arc trajectory, and a multi-controller segmented control method is established for the multi-segment line combination to maintain the uniformity of the driving speed of the vehicle in the path tracking process.

[0006] Preferably, in the trajectory tracking optimization method, a vehicle kinematics model is first established, the front wheel steering angle during vehicle movement is taken as an input quantity, and the actual output quantity of vehicle operation is calculated through the vehicle kinematics model.

[0007] Specifically, the front wheel steering angle is used to control the turning radius of the steering wheel, and the turning radius of the steering wheel is used to control the turning radius of the wheel, wherein the greater the front wheel steering angle, the smaller the turning radius of the vehicle.

[0008] Preferably, in the multi-controller segmented control method, a parking trajectory auxiliary tracking mode based on visual recognition is established.

[0009] Preferably, in the multi-controller segmented control method, an improved model predictive control algorithm is established in the circular arc trajectory control.

[0010] Preferably, in the improved model predictive control algorithm, a vehicle prediction model is established, and the vehicle prediction model is fused with the parking trajectory data collected by visual recognition for control.

[0011] Preferably, the straight line trajectory and the circular arc trajectory are the turning points of vehicle speed change at the intersection position of the straight line and the circular arc.

[0012] Preferably, a steering control model based on speed mutation is separately established at the turning point.

[0013] Specifically, the turning point is divided into turning point one from straight line trajectory to circular arc trajectory and turning point two from circular arc trajectory to straight line trajectory.

[0014] Preferably, the straight line trajectory and the circular arc trajectory are distinguished by setting a curvature threshold value transformation range based on the center point of the vehicle front axle.

[0015] Specifically, in the actual parking trajectory tracking process, the generated parking trajectory based on the multi-segment line is not entirely composed of straight line trajectories and circular arc trajectories, wherein the straight line trajectory is set by the curvature threshold value transformation range based on the center point of the vehicle front axle, and the parking trajectory of the straight line trajectory also has a certain curvature range. Therefore, in order to eliminate the influence of the instability factor caused by the curvature range in the straight line trajectory on the vehicle running speed during path tracking, an error correction method based on a separate visual recognition mode is established.

[0016] Preferably, the trajectory tracking error in the straight line trajectory is corrected by a separate visual recognition mode.

[0017] Preferably, the trajectory tracking optimization method is applicable to any one of the vertical parking and the parallel parking environment.

[0018] Compared with the prior art, the present application has the beneficial effects that:

[0019] (1) The trajectory tracking optimization method for the automatic parking process, on the basis of the planned automatic parking path, establishes a vehicle dynamics model based on a specified vehicle, calculates the vehicle's own running parameters according to the vehicle dynamics model, the vehicle's own running parameters are input as input quantities by inputting the initial front wheel steering angle of the vehicle, to control the vehicle to follow the planned parking route, in the process of driving, the vehicle's own running parameters are obtained and adjusted in real time, and the front wheel steering angle of the next moment is obtained according to the vehicle's own running parameters, the front wheel steering angle data is fed back to the steering wheel to automatically control the steering radius of the steering wheel, to avoid the problem of large path tracking deviation and vehicle running instability caused by uneven vehicle speed in the process of vehicle path tracking.

[0020] (2) On the basis of (1), the present application establishes a multi-controller segmented control method by splitting the parking trajectory into a combination of straight line trajectory and circular arc trajectory in the process of path tracking, to avoid the problem of inconsistent vehicle tracking deviation caused by inconsistent vehicle tracking error in the traditional control process, and finally to better improve the stability of the vehicle in the process of path tracking and improve the use experience of automatic parking. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 To improve the model predictive control algorithm in the fusion control flowchart. DETAILED DESCRIPTION

[0022] Example 1:

[0023] The trajectory tracking optimization method for the automatic parking process in this embodiment specifically includes, on the basis of the established parking trajectory, splitting the parking trajectory into a combination of straight line trajectory and circular arc trajectory, establishing a multi-controller segmented control method for the combination of the straight line trajectory and the circular arc trajectory, to maintain the uniformity of the driving speed of the vehicle in the process of path tracking.

[0024] In the actual application scene of automatic parking, when the range of the application scene is small, the ratio of the steering control of the steering wheel to the front wheel steering angle will increase, and vice versa; when the vehicle speed suddenly changes and the vehicle performs a turn operation, the ratio of the steering control of the steering wheel to the front wheel steering angle will increase, and vice versa; therefore, according to the improvement of the control function of the current vehicle controller, in the trajectory tracking control based on the automatic parking process, the relationship between the parking following speed of the vehicle and the vehicle steering angle needs to be strictly controlled to avoid the problem of inaccurate automatic parking steering following caused by the limitation of the application scene and the sudden change of the vehicle speed.

[0025] The multi-controller segmented control method establishes a parking trajectory auxiliary tracking mode based on visual recognition.

[0026] Specifically, the visual sensor is used to assist the traditional laser ranging sensor in detecting the parking environment and calibrating the automatic parking position during the automatic parking process.

[0027] The laser ranging sensor is used to measure the parking distance in an environment with obstacles around the parking area. However, when the parking environment is relatively open and the obstacles are small, the laser ranging sensor alone cannot fully measure the specific distance of the vehicle from the parking space. At this time, the visual recognition device is added to assist in judgment without the need for obstacle-based reflection measurement. The parking space range can be directly photographed, and the point information required for automatic parking can be located, improving the efficiency and application range of automatic parking.

[0028] The improved model predictive control algorithm establishes a vehicle prediction model, and fuses the vehicle prediction model with the parking trajectory data collected by visual recognition.

[0029] Specifically, as shown in Figure 1 The specific process of the fusion control is as follows:

[0030] S1, calculate the error value between the output prediction sequence output by the vehicle prediction model and the visual recognition data;

[0031] S2, transmit the obtained error value to the target function to perform the first optimization of the output prediction sequence;

[0032] S3, on the basis of the first optimization of the output prediction sequence, calculate the error value between the output prediction sequence and the actual output, and transmit the obtained error value to the target function to perform the second optimization of the output prediction sequence. The output of the second optimization is output to adjust the vehicle speed.

[0033] The straight line trajectory and the circular arc trajectory are the turning points of the vehicle speed change at the intersection position of the straight line and the circular arc.

[0034] Wherein, since the tracking path is a known path, in the first turning point, the predicted curvature at the next moment is collected based on a separate visual recognition-based calculation method, and in the second turning point, the predicted curvature at the next moment is collected based on an improved model predictive control algorithm. A curvature interpolation section is established according to the output curvature of the current point and the predicted curvature at the next moment, the curvature interpolation section connects the curvature interpolation points into a smooth curve by establishing a plurality of uniform curvature interpolation points, so as to improve the uniformity of the vehicle speed when the vehicle tracks the path.

Claims

1. A trajectory tracking optimization method for an automatic parking process, characterized in that, Specifically, based on the established parking trajectory, the parking trajectory is divided into a combination of multiple straight lines and circular arcs, a multi-controller segmented control method is established for the combination to maintain the uniformity of the vehicle speed during path tracking; The straight lines and circular arcs are the turning points of the vehicle speed change at the intersection positions of the straight lines and circular arcs; A steering control model based on speed mutation is established at the turning points; The straight lines and circular arcs are distinguished by setting a curvature threshold value transformation range based on the center point of the vehicle front axle; The trajectory tracking error in the straight lines is corrected by a separate visual recognition method.

2. The trajectory tracking optimization method for automatic parking process according to claim 1, characterized in that, In the trajectory tracking optimization method, a vehicle kinematics model is first established, the front wheel steering angle during vehicle motion is taken as an input quantity, and the actual output quantity of the vehicle operation is calculated through the vehicle kinematics model.

3. The trajectory tracking optimization method for automatic parking process according to claim 1, characterized in that, In the multi-controller segmented control method, a visual recognition-based parking trajectory auxiliary tracking method is established.

4. The trajectory tracking optimization method for automatic parking process according to claim 3, characterized in that, In the multi-controller segmented control method, an improved model predictive control algorithm is established in the circular arc trajectory control.

5. The trajectory tracking optimization method for automatic parking process according to claim 4, characterized in that, The improved model predictive control algorithm establishes a vehicle prediction model, and simultaneously controls the vehicle prediction model and the parking trajectory data collected by visual recognition.

6. The trajectory tracking optimization method for automatic parking process according to claim 1, characterized in that, The trajectory tracking optimization method is applicable to any one of the vertical parking and parallel parking environments.

Images