【作者】 乔应； 赵立军； 刘清河； 聂文鹏； 潘丞； 郑龙泽；

【Author】 Qiao Ying;Zhao Lijun;Liu Qinghe;Nie Wenpeng;Pan Cheng;Zheng Longze;Harbin Institute of Technology(Weihai);

【机构】 哈尔滨工业大学(威海)；

【摘要】 针对结构化场景下的传统解耦规划方法无法充分考虑动态障碍物信息,导致最终轨迹陷入次优解的问题,本文基于可充分考虑静态、动态障碍物信息的T-S-L三维空间,提出一种可直接规划轨迹的方法。首先采用三维动态规划算法,设计包含障碍物、平滑性、偏移全局路径以及参考速度的代价函数,并搜索决策轨迹;然后利用三次样条函数对决策轨迹进行插值和平滑处理;最后考虑平滑性、道路边界等约束,采用二次规划使目标函数达到极值,对初始轨迹进行平滑处理。在仿真实验部分,设计了两种工况对本文所提出的算法进行可行性验证,并在Ubuntu/CARLA中,结合跟踪控制层分别在静态、动态障碍物场景中对算法进行闭环仿真验证。通过上述仿真实验验证,本文提出的算法能使规划轨迹有效地处理静态障碍物和动态障碍物,并保证驾驶的舒适性。更多还原

【Abstract】 In order to solve the problem that the traditional decoupling planning method in structured scenarios cannot fully consider the dynamic obstacle information,resulting in the final trajectory falling into a suboptimal solution,this paper proposes a method that can directly plan the trajectory based on the T-S-L three-dimensional space that can fully consider the static and dynamic obstacle information.Firstly,the three-dimensional dynamic programming algorithm is used to design the cost function including obstacles,smoothness,offset global path and reference velocity,and the decision trajectory is searched,then the cubic spline function is used to interpolate and smooth the decision trajectory,and finally,considering the constraints of smoothness and road boundary,the quadratic programming is used to make the objective function reach the extreme value,and the initial trajectory is smoothed.In the simulation experiment,two working conditions are designed to verify the feasibility of the proposed algorithm,and the algorithm is simulated and verified in Ubuntu/CARLA combined with the tracking control layer in static and dynamic obstacle scenarios.Through the above simulation experiments,the proposed algorithm can effectively deal with static and dynamic obstacles and ensure the driving comfort.更多还原