The Research On Autonomous Driving Motion Planning Considering The Uncertainty Of The Movement Of The Preceding Vehicle In Urban Environment

The development of autonomous driving is of great significance in improving road traffic safety and traffic efficiency.As one of the key modules in the automatic driving system,motion planning is responsible for connecting the system perception and control modules,and planning the local trajectory of the vehicle according to the current scene information of the vehicle and the vehicle’s own state,which is an intuitive embodiment of the wisdom of autonomous vehicles.The motion planning needs to ensure that,under the vehicle kinematics system,a safe and collision-free,comfortable ride and efficient driving trajectory are generated.Due to the high dimensionality and high nonlinearity of the trajectory,it is difficult to solve the trajectory directly within the planning period.Therefore,this paper solves by decoupling the trajectory into two sub-problems,path planning and speed planning.Path planning needs to avoid static obstacles in the scene,and the avoidance of moving obstacles is reflected in speed planning.The combination of the two results in the vehicle trajectory in space and time.The main research contents of this article are as follows:(1)For the path planning problem of urban structured roads,the planning space coordinate system is converted into the Frenet coordinate system to reduce the dimension and complexity of the path planning space.In order to meet the complex traffic scenes in cities,the double-circle model of vehicles is adopted,which can accurately describe the external contours of vehicles under the premise of ensuring a certain safety margin.In the limited planning period,combined with the static obstacles in the scene,through linearization to approximate the vehicle collision constraints,a local convex space is constructed,and the path is directly generated using quadratic planning in the convex space.(2)Speed planning needs to avoid moving obstacles in motion planning.Based on the reference path given by path planning,an S＿T motion space is constructed to describe the constraint relationship between the self-vehicle and the moving obstacle,reducing the complexity of the planning problem.By relaxing the kinematics nonlinear constraint,the method of constructing locally convex spaces is extended to velocity planning,and iteratively converges to the local optimal solution by quadratic programming.In addition,considering the uncertainty of motion of dynamic obstacles,in the process of speed planning including uncertainty,the conservative degree of speed planning is given,and the complexity of the original optimization problem is kept unchanged.(3)Build a real-time simulation platform,build a three-dimensional test scene,sensor model and a vehicle dynamics model based on Pre Scan,build an automatic driving software platform through ROS.For urban scenes,a trajectory planning strategy is proposed,and simulation experiments are completed on the simulation platform.