基于采样隐马尔可夫模型的移动机器人路径规划  

Path Planning of Mobile Robot Based on Sampled Hidden Markov Model

作　　者：白晓兰[1] 张振朋 周文全 袁铮 任鹏博 张炳贵 BAI Xiaolan;ZHANG Zhenpeng;ZHOU Wenquan;YUAN Zheng;REN Pengbo;ZHANG Binggui(School of Mechanical and Power Engineering,Shenyang University of Chemical Technology,Shenyang 110142,China)

机构地区：[1]沈阳化工大学机械与动力工程学院,沈阳110142

出　　处：《组合机床与自动化加工技术》2023年第7期50-56,共7页Modular Machine Tool & Automatic Manufacturing Technique

摘　　要：为提高移动机器人路径规划的学习和预测能力,使得规划的路径长度和转弯较小,能够动态躲避行人,提高搜索效率,提出一种基于采样隐式马尔可夫模型(sampled hidden markov model,SHMM)的移动机器人路径规划方法。首先,将2D网格环境下的轨迹划分为位置(x,y)、航向(θ)、线速度(v l)四个维度,建立机器人轨迹预测模型;其次,利用KL散度(kullback-leibler divergence)将建立的HMM运动模型进行降采样;最后,用K-means聚类因子、SHMM因子和转角因子改进蚁群算法进行路径规划。仿真验证了基于SHHM的移动机器人规划的路径长度短、转角少,并具有学习和预测能力,在路径规划地同时能很好地躲避行人。In order to improve the learning and prediction ability of mobile robot path planning,make the planned path length and turn smaller,dynamically avoid pedestrians,and improve the search efficiency,a mobile robot path planning method based on the sampled hidden markov model(SHMM)is proposed.Firstly,the trajectory in 2D grid environment is divided into four dimensions:position(x,y),heading(θ)and linear velocity(v l),and the trajectory prediction model of the robot is established;Secondly,the HMM motion model is de sampled using KL divergence;Finally,K-means clustering factor,SHMM factor and corner factor are used to improve ant colony algorithm for path planning.The simulation results show that the SHHM based mobile robot planning has short path length,few corners,learning and prediction capabilities,and can avoid pedestrians while planning the path.

关 键 词：学习预测 路径规划 马尔可夫模型 蚁群算法 

分 类 号：TH165[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]