Research Of Autonomous Vehicle Local Path Planning And Control For Dynamic Obstacle Avoidance

In the research of local path planning with dynamic obstacles for autonomous vehicles,most of them only consider the safe distance constraint between vehicle and obstacles,but ignore the limits of vehicle mobility performance,so that the controller cannot accurately track the upper reference trajectory.Aiming at this limitation,this paper proposes a new local path planning layer and controller design idea.This article main content is:1.Aiming at the existence of dynamic obstacles in the driving environment,this thesis puts forward a local trajectory planning idea,based on the theory of linear time-varying model predictive control,and the monorail vehicle dynamic model,considers the constraints of vehicle front wheel steering angle,front wheel steering angular velocity,road adhesion limit and the tire side-slip angle,predicts the traveling track of vehicle and obstacle,finally design a optimization solution function to plan a trajectory,which avoids obstacles and close to the global reference path.2.To guarantee the accuracy of the vehicle tracking reference trajectory,this thesis establishs a more accurate predictive model of double rail vehicle dynamic model,based on vehicle state sensors(vehicle yaw velocity,yawing angle of car body,yawing angular velocity and position coordinates),through the control theory of model predictive control for front wheel steering angle.3.Finally,the local path planning layer and the controller are integrated,and the cosimulation platform is built by using CarSim and Simulink to realize the safe obstacle avoidance of self-driving vehicles in complex scenarios where curves and dynamic obstacles exist simultaneously,which verifies the feasibility of the local path planning layer and controller design theory proposed in this paper.