分级随机采样弱随机RRT算法及在移动机器人运动规划中的应用

针对移动机器人运动规划过程中,采用快速扩展随机树(RRT)算法采样效率低,寻找临近节点计算量大,及非线性反馈控制器不受系统模型动态约束的问题, 提出一种新的基于分级随机采样与扩展的弱随机RRT算法,同时设计快速限幅非线性反馈控制器,保证运动规划过程中机器人始终满足系统模型动态约束。首先,在迭代伊始结合节点评价策略建立节点的选取集合;其次,按照规定顺序选取扩展节点并随机选择扩展方向,将计算得到的新子节点连接到随机树完成扩展;然后,对初始路径进行规划,采用快速限幅非线性反馈控制器计算机器人在路径点上的控制序列和位姿,实现移动机器人的运动规划;最后,通过仿真验证了该算法的有效性。结果表明:提出的分级随机采样弱随机RRT算法不依赖最近节点的选取,相比RRT算法缩短了求解时间,提高了迭代速度。

Hierarchical Random Sampling Weak Random RRT Algorithm and Application for Motion Planning of Mobile Robot

Aiming at the problems of rapid expansion random tree (RRT) has low sampling efficiency and large amount of node searching for nearest node, and nonlinear feedback controller is not subject to the dynamic constraints of the system model in mobile robots motion planning. A new weak RRT based on hierarchical random sampling and expansion is proposed, and a fast limiting amplitude nonlinear feedback controller is designed to ensure the robot can satisfy the dynamic constraints of the system model during the motion planning. Firstly, nodes to be expanded set are established at the beginning of the iteration in conjunction with the node evaluation strategy. Secondly, the nodes to be expanded are selected according to the prescribed order and the expansion direction is randomly selected, then the calculated new child node is connected to the random tree to complete the expansion. Thirdly, the initial path is planned by calculating the control sequence and posture of the path point of the mobile robot via the fast limiting amplitude nonlinear feedback controller to realize the motion planning of the mobile robot. Finally, the effectiveness of the proposed algorithm is verified through the simulations. The RRT based on hierarchical random sampling and expansion doesn’t depend on the selections of the nearest node, which reduces the solving solutions time of the RRT and increases the iteration speed.