项目名称: 混合速度障碍法及其在移动机器人动态运动规划中的应用研究
项目编号: No.61305117
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 仲训昱
作者单位: 厦门大学
项目金额: 25万元
中文摘要: 基于速度障碍原理的碰撞分析方法,是研究动态复杂环境下移动机器人运动规划的一个新方向。本项目通过对速度障碍法的深入分析,发现其运动"摆动 "问题根源所在,提出用于多机器人协调的互动策略;模仿人的视距调整行为,综合计算碰撞时间和碰撞距离建立对障碍物大小的自适应调整方法;在速度变化空间中进行信息映射,将运动规划转化为一种多目标优化问题,在多步可达速度窗口(可以确保安全并克服规划的"短视"性)中,采用快速搜索获得使目标函数值最优且满足所有约束条件的速度指令。这些解决运动"摆动"、避碰"保守"、规划"短视"、速度"优化选择"问题的一系列新思路和方法,将形成一种完善的混合速度障碍法理论;并针对实际应用,建立机器人运动约束关系的分析方法与信息不确定性的处理方法。这不仅对速度障碍法的发展具有重要理论意义,同时对促进各领域移动机器人的应用开发,特别是对无人车、无人艇和服务机器人的自主导航具有切实的指导作用。
中文关键词: 移动机器人;运动规划;速度障碍;自主导航;避障
英文摘要: Collision analysis method based on the principles of velocity obstacle, is a new direction of the study of mobile robots motion planning in the dynamic complex environments.This project 1) finds the problem of motion "oscillation" and proposes an interactive strategy for multi-robot coordination through the deep analysis of the velocity obstacle method; 2) will design adaptive modulation method to the size of the obstacle by imitatiton the behavior of line-of-sight adjustment of human with the computation of collision time and collision distance; 3) mapping the information in velocity variation space, transforms the motion planning problem into a multi-objective optimization problem, making planning in the multi-step reachable velocity window, gets the desired velocity command using a quick search method and with optimal value of the objective function satisfied all constraints. A series of the new ideas and new methods of these solutions for the problems of motion "oscillation", "conservative" collision avoidance, "short-sighted" planning and the "optimal selection" of the desired velocity, will form a complete hybrid velocity obstacle theory. And establish the relationship between motion constraints of the wheeled differential-drive mobile robot, and the analysis methods of information uncertainty for practica
英文关键词: mobile robot;motion planning;velocity obstacle;autonomous navigation;obstacle avoidance