基于动态运动基元的移动机器人路径规划

针对二维环境下移动机器人路径规划问题,提出一种基于动态运动基元(DMPs)的路径规划方法。该方法首先用手柄来控制机器人,记录机器人的运动轨迹,然后将记录的轨迹作为DMPs算法的学习样本,利用轨迹样本进行训练获得DMPs算法的模型参!,从而实现机器人的路径规划。当运动环境中有障碍物时,在已有学习基础上通过设计耦合因子规划出避障路径,在此基础上,改变机器人运动的目标位置,可以完成对新目标的泛化推广。仿真和实验结果表明,DMPs算法在移动机器人路径规划上具有可行性。

Path planning for mobile robots based on dynamic movement primitives

A novel path planning algorithm based on dynamic movement primitives (DMPs) for a mobile robot working in 2D environment is proposed.The algorithm can be described below:Firstly,the robot's trajectory is recorded while the robot is controlled by a handle.Then the recording trajectory is used as a learning sample of DMPs,and the autonomous path planning of the robot is realized,which utilizes the model parameters obtained by training the sample trajectory.In the environment with obstacles,the avoidance path can be produced through designing a coupling factor on the basis of the existing study.Based on the autonomous path planning,the generalization comes true to new goals while the original goal is changed.The simulation and experimental results show that the DMPs algorithm is feasible on a mobile robot path planning.