Variable-speed dynamic lane changing track planning method based on vehicle driving rule

Abstract

The invention relates to a variable-speed dynamic lane changing track planning method based on a vehicle driving rule. The driving safety and comfort of the vehicle are controlled through the acceleration and the jerk; and a speed change rule is fused into a trajectory planning method, a cycloid * and a circular curve * are adopted as motion primitive planning lane change trajectories, so that thelane change trajectory of the automatic driving vehicle is consistent with an automatic cruise trajectory, and the lane change trajectory is dynamically updated in real time in combination with the latest position and the safety distance of the vehicle until the automatic driving vehicle completes lane change. A speed change rule is originally introduced into lane change motion, and the vehicle lane change efficiency is improved; the acceleration and jerk required by comfort and safety are considered while the trajectory is generated, the process of searching for the optimal trajectory by setting an optimization target in the past is omitted, the optimal trajectory is directly calculated under constraint, and the efficiency is remarkably improved. After a certain time interval, dynamic iteration is carried out to recalculate the planning trajectory, thereby ensuring that the end point of the vehicle planning trajectory is beyond the safe distance in real time, and effectively correcting the error of path tracking.

Description

technical field [0001] The invention relates to a dynamic lane-changing track planning method for speed change based on vehicle driving rules. Background technique [0002] Autonomous driving is the trend of vehicle development. David González and others pointed out that real-time computing in dynamic environments is the biggest challenge for autonomous driving, which includes two aspects: 1. Efficient real-time computing of vehicle motion planning; response. [0003] Vehicle speed and trajectory curvature are closely related to the lateral acceleration of vehicle motion, which is directly related to vehicle safety, comfort, and stability (Chu et al., 2012). Vehicle motion planning refers to generating the geometric path of the starting point and the ending point, and giving the speed planning of the vehicle along the path. Commonly used are trajectory generation method, sampling-based algorithm and search-based algorithm. Trajectory generation methods include direct const...

AI Technical Summary

Problems solved by technology

[0006] 2. Improved path-velocity decomposition method. Path planning is a fundamental task for autonomous moving vehicles and has been extensively studied in the literature. However, most existing geometric path generation presupposes uniform motion and provides a series of waypoints. However, the actual running speed of the vehicle changes, so the planned path is not optimal, and sudden changes in acceleration will lead to control tracking errors and uncomfortable driving, which is often not feasible

The vehicle's own state uncertainty includes uncertainty in perception, positioning and control. For example, if the vehicle uses low-precision GPS positioning, and the planning algorithm does not take this into account, it will bring danger to the vehicle, making the self-driving car may become unsafe

Images