一种移动机器人自主避障与导航方法  

A Obstacle Avoidance and Navigation Method for the Mobile Robot

作　　者：储星[1] 文桂林[1] 卢远志[1] 

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室特种装备先进设计技术与仿真教育部重点实验室,长沙410082

出　　处：《机械科学与技术》2016年第6期939-945,共7页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金重点项目(91120306/F03);高等学校博士学科点专项科研基金课题项目(20120161130001);国家科技支撑计划项目(2012BAH09B02)资助

摘　　要：提出了一种移动机器人的避障与导航方法。该方法充分考虑了人类感知、推理和决策的不确定性以及人类避障与导航的特点,基于云模型构建了一种同时具备随机性和模糊性的高鲁棒性的拟人决策算法,主要包括障碍物避让规则库、目标吸引规则库、方位偏差和距离前件正态云模型、转角后件正态云模型以及通行函数的设计,提高了移动机器人的人工智能水平,从而降低了系统对传感器和执行器性能的苛刻要求。通过理论分析计算和建模仿真试验,验证了所提出新方法的有效性。A novel method of obstacle avoidance and navigation is presented for mobile robots. Based on cloud models,the uncertainty of human being perception,reasoning and decision as well as the characteristics of human being collision avoidance and navigation are taken into account adequately. An anthropopathic decision making algorithm with the randomness meanwhile fuzziness is constructed based on cloud models,mainly includes the designs of rules for obstacles avoidance and object attracting,Gaussian cloud model of azimuth deviation,distance offset and steering angle,pass function. The method is possible to meet with the requirements of low computational cost and low precision for sensors as well as improvement of the artificial intelligence（ AI）. Theoretical analysis and simulation experiment show that the proposed method is effective.

关 键 词：移动机器人 避障 导航 云模型 决策 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]