面向动态不确定场景的自动驾驶车辆时空耦合分层轨迹规划研究

Study on Spatio-temporal Coupled Hierarchical Trajectory Planning of Autonomous Vehicles for Dynamic Uncertain Scenarios

针对智能车辆在结构化道路上,处理动态复杂场景能力较弱以及实时性差的问题,基于可达集的方法设计一种分层式时空耦合的轨迹规划方法,完成车辆在动态不确定场景下的轨迹规划。首先,根据自车与周围的障碍物状态信息,预测障碍物在未来一段时间内的位置分布概率,基于可达集的方法计算各时刻的可达区域,得到最优行驶走廊及一条初始轨迹。其次,利用二次规划的方法在最优行驶走廊内根据初始轨迹进行优化,求解出一条平滑轨迹,并对该轨迹进行跟踪行驶。最后,利用PreScan、CarSim和Matlab软件搭建仿真平台,在动态复杂交通场景下进行仿真分析,并在实车平台上进行避障测试。结果表明,所设计的规划方法能够有效处理动态不确定场景,在保证安全性的前提下能够规划出高效的通行轨迹,同时也能兼顾预测准确度与实时性。

In order to solve the problem of weak ability of intelligent vehicles to handle dynamic complex scenes and poor real-time performance on structured roads,a hierarchical spatio-temporal coupled trajectory planning method is designed based on the reachable set method to complete the vehicle’s dynamic uncertainty scenarios trajectory planning.Firstly,based on the status information of the vehicle and surrounding obstacles,the location distribution probability of obstacles in the future is predicted,and the reachable area at each time is calculated based on the reachable set method,and the optimal driving corridor and an initial trajectory are obtained.Secondly,the quadratic programming method is used to optimize the initial trajectory within the optimal driving corridor,and a smooth trajectory is obtained,and the trajectory is tracked.Finally,a simulation platform is built using PreScan,CarSim and Matlab software,simulation analysis is conducted under dynamic and complex traffic scenarios,and obstacle avoidance testing is conducted on the real vehicle platform.The results show that the designed planning method can effectively handle dynamic uncertain scenarios,plan efficient traffic trajectories while ensuring safety,and can also take into account prediction accuracy and real-time performance.