Research On Key Technologies Of Unmanned Driving Trajectory Planning And Tracking Control

As the number of vehicles in China continues to increase,problems such as traffic congestion and accidents have become increasingly prominent.Autonomous driving technology is expected to become a major development direction in the future transportation industry.Compared with other driving conditions,urban roads have a more complex environment with high uncertainty and complex traffic participants.A complete and mature technological solution that can meet the needs of urban autonomous driving is still being explored.This article is based on the national key research and development program "Research and Practice of Group Intelligent Baggage Transport Vehicles for Airport Baggage Transfer"(2021YFE0203600).It focuses on the key technologies of trajectory planning and tracking control for autonomous vehicles in urban environments.The specific contents are as follows:The problem that the traditional pathfinding A~* algorithm planning path nodes is too large to meet the steering characteristics of unmanned vehicles and the distance to obstacles is too close.Two improvement methods are proposed to improve the A~*algorithm: expanding to the 24 neighboring positions during the search and incorporating the advantages of artificial potential field into the heuristic function.The path turning angle planned by the A~* algorithm after neighborhood expansion and heuristic function improvement is smaller,and it can Keeping a certain distance from obstacles is more in line with the needs of unmanned driving path planning.After studying the problem of unsmooth front-end planning path and insufficient consideration of safety and comfort,this paper proposes a back-end trajectory optimization method to transform trajectory planning into a numerical optimization problem.Aiming at the problem that the trajectory optimization algorithm is not conducive to obstacle avoidance,it is proposed to use the properties of Bezier curves for hard constraint optimization,and the trajectory planned by the numerical optimization method is more in line with the needs of unmanned vehicle trajectory tracking.Aiming at the problem that the trajectory optimization with hard constraints may deviate from the initial path,an improved motion corridor generation method is proposed to ensure the solution space and keep the control area within a certain range as much as possible,so as to achieve the purpose of improving the optimized trajectory.The study investigated a model predictive tracking control strategy for autonomous driving.Based on the vehicle dynamics model,a multi-constraint linear time-varying model predictive controller was designed.The effectiveness and robustness of the controller were verified through simulations at different road surface conditions with varying levels of adhesion and different vehicle speeds.To validate the effectiveness of the proposed algorithm in this paper,trajectory optimization,on-road data validation,and analysis of tracking control were performed.The results show that the trajectory optimization algorithm proposed in this paper can generate collision-free trajectories that meet practical requirements.The Model Predictive Control(MPC)controller can accurately guide the vehicle to track the planned trajectory,meeting the requirements for control accuracy and smoothness.