Research On Motion Planning Algorithm For Unmanned Vehicles In Dilemmas

As the key to realize autonomous driving of intelligent driving vehicles,motion planning technology to plan a safe,comfortable,efficient,and feasible trajectory for the vehicle has always been the focus of intelligent driving research.However,it is unrealistic for intelligent driving vehicles to completely avoid obstacles of 100%,due to the constraints of technology,economy,and other external uncontrollable factors.Therefore,in this paper,a vehicle motion planning model is designed to achieve reasonable avoidance under the condition of emergency dilemma based on the public inclination of evasive steering(IES)and the considerations of moral,legal,economic,harm and other elements.The main contents in this research are as follows:(1)Constructing the vehicle motion planning model.Aiming at the ethical controversy of the avoidance of intelligent driving vehicles under the dilemma condition,the public inclination of evasive steering(IES)is modeled according to the investigation results of the public ethical tendency to avoid risk,then,the model is applied to the mapping of the vehicle motion planning engineering layer,in order to construct the vehicle motion planning model under the dilemma condition.During the modeling,the vehicle motion planning problem is abstracted as a constrained optimization problem,where the optimization goal is the weighted sum of IES,the tracking of reference trajectory and the change rate of the increment of control quantity,so that the intelligent driving vehicle can comply with the tendency of public ethical avoidance as much as possible while meeting the comfort requirements of passengers in the vehicle,and the constraint conditions include the considerations of vehicle dynamics model,tire adhesion,steering wheel angle to ensure the feasibility and stability of the planned trajectory.(2)Solving the optimization problem.After considering the high nonconvexity and nonlinearity of the constraint optimization problem,in this paper,the layered optimization method is used to solve the problem.Among them,the upper layer is the warm-start trajectory acquisition layer,where a sequence of candidate trajectories is obtained by means of curve interpolation and a suboptimal rough trajectory was selected according to the objective function that is similar to the optimization layer(lower layer)and the physical constraints such as vehicle velocity,acceleration,maximum radius of curvature,etc.The lower layer is numerical optimization layer,where the sequential quadratic programming(SQP)is used to start from the optimization results of the upper layer,and the optimization problem converges in a finite time.(3)Building the co-simulation system,and verifying the feasibility,the stability and ethic correctness of motion planning algorithm for trajectory planning.The motion planning algorithm and MPC trajectory tracking control algorithm are designed into a motion planning controller.In addition,with the help of vehicle dynamics simulation software named Carsim,the simulation scene and input and output variables of the software are defined.Then,the Carsim and the motion planning cotroller co-simulation platform was built by combining the above two modules,and the motion planning algorithm and trajectory tracking control analysis of the intelligent driving vehicle under the dilemma condition are verified.