| With the continuous development of autonomous driving technology,the study of complex traffic scenarios is a research difficulty.Compared with unsignalized intersections,vehicles in U-turn scenarios have to deal with more complex traffic scenarios,such as the gap selection of conflicting vehicles,traffic sign recognition,signal light stage judgment,etc.Autonomous vehicles need to quickly and accurately identify and respond to these factors in order to safely cross the U-turn intersections.In order to ensure the safety and efficiency,it is necessary to conduct research on U-turn scenarios of autonomous driving.Trajectory planning and motion control of vehicles are the key technologies to realize autonomous driving.Therefore,the combination of urban U-turn scenarios and autonomous driving planning and control research is particularly important.Thus,this paper takes the vehicles in U-turn scenarios as the research object and optimizes the methods of trajectory planning and control.The specific work is as follows:(1)Establish a decision-making model for U-turn scenes.First,the trajectory characteristics and conflict characteristics of U-turn vehicles were analyzed through actual investigation data.At the same time,to ensure the safety of autonomous vehicles,the problem of how U-turn vehicles merge into oncoming traffic lanes is considered.The vehicle decision-making model is constructed for single or multiple U-turn vehicles in specific scenarios and the driving status is switched according to different driving conditions using a finite state machine.(2)Optimize the trajectory planning model for U-turn vehicles.To solve trajectory planning problem,a Frenet frame is first established,and the reference line is smoothed.In path planning,the dynamic programming algorithm is used to find solution and then the solution is optimized by quadratic programming to achieve the planning goal of not colliding with any obstacle and satisfying constraints.In speed planning,the planned path is assigned with a speed,and corresponding constraint conditions are added based on the actual road conditions to ensure the safety,comfort,and feasibility.By separating the path and speed,the trajectory planning module converts the complex 3D problem into a simpler 2D problem,achieving trajectory planning for U-turn vehicles.Then,the driving scenario designer module in Matlab is used to build the planning scenario,and the optimized trajectory planning method is useful.(3)Implement trajectory tracking for U-turn vehicles.To achieve accurate tracking of the planned path,the model predictive control method is optimized in Uturn scenarios.Considering the radius constraint of U-turn vehicles and the influence of kinematic models,the deviation state equation is established based on a two-degreeof-freedom vehicle model,and the trajectory is solved based on the rolling optimization method.(4)Load a joint simulation model.To verify the performance of the proposed planning and control algorithms for U-turn scenarios,Car Sim was used to build the vehicle dynamic model,with Simulink serving as the simulation signal interface.The results show that the trajectory planning algorithm can generate collision-free trajectories and the motion control module has good control performance,and the method proposed in this article is superior to the control effect of the linear quadratic regulator method in terms of time consumption. |
