室内环境下基于边际约束的快速路径自主探索算法

Fast autonomous path exploration algorithm based on marginal constraint in indoor environment

为了提高移动机器人在室内未知环境的自主探索能力,实现移动机器人在探索目标点之间的安全、快速移动,提出一种基于边际约束的快速路径自主探索算法。首先,将机器人自主探索问题描述为部分可观测马尔可夫决策过程模型。之后,在传统的快速扩展随机树(RRT)算法基础上,将随机树的生长空间划分为边际四象限空间,结合启发式评估函数的评价。该算法加快了移动机器人在探索目标点之间的移动速度,同时减少了随机树的节点,降低了对内存空间的占用。通过Matlab仿真实验,在实验设定的仿真环境中,该算法比传统RRT算法在时间上缩短约了75%,节点数量减少了约80%,并在机器人操作系统的仿真实验中验证了算法的实用性。

In order to improve the autonomous exploration ability of mobile robot in unknown indoor environment and realize the safe and fast movement of mobile robot among the exploration target points,a fast autonomous path exploration algorithm based on the marginal constraint is proposed.First,the robot’s autonomous exploration problem is described as partially observable Markov decision process model.Then,based on traditional rapidly-exploring random trees(RRT)algorithm,the growth space of the random tree is divided into marginal four-quadrant space combined with the evaluation of heuristic evaluation function.This algorithm accelerates the moving speed of mobile robot among the exploring target points,reduces the nodes of random tree,and occupies less memory space.Matlab simulation experiments are carried out in the environment set up by the experiment,which show that the algorithm shortens the time by about 75%and reduces the number of nodes by about 80%,compared with traditional RRT algorithm.Moreover,the practicability of the algorithm is verified in the simulation experiment of robot operating system.