摘要
全自主机器人的导航控制未对目标对象定位,无法获取有效目标路径,导致控制方法精度低且效率差,提出全自主机器人视觉信息采集与导航控制方法。该方法采集机器人信息数据,根据相关数据以及粗定位和精定位的方法精确目标物的位置,在Q算法的帮助下得到最优控制策略,结合基于Boltzmann探索方案生成机器人的下一步运动角度,根据奖惩值避开障碍物,通过不断训练机器人不断得到最优导航动作,实现全自主机器人的导航控制。实验结果表明,所提方法控制的轮转角和航向角均与理想角一致,横向位移偏差达到了0m,跟踪目标在Y轴最高偏差值为0.5m,因此,该方法导航控制效果好和目标对象跟踪精度高。
The navigation control of the fully autonomous robot fails to locate the target object and cannot obtain the effective target path,resulting in low precision and poor eficiency of the control method,a visual information acquisition and navigation control method for fully autonomous robots is proposed.This method collects the information data of the robot.The method is accurate to the position of the target,and with the help of the Q algorithm,the optimal control strategy is obtained.The next movement angle of the robot is generated based on the Boltzmann exploration scheme,and obstacles are avoided according to the reward and punishment value.By constantly training the robot to get the best navigation action.Optimal navigation action to realize the navigation control of fully autonomous robots.The experimental results show that the wheel rotation and heading Angle controlled by the proposed method are consistent with the ideal Angle,the lateral displacement deviation reaches Om,and the maximum deviation of the tracking target in the Y-axis is O.5m.Therefore,the proposed method has good navigation control effect and high trackingaccuracy.
作者
祝毅鸣
刘鑫
ZHU Yi-ming;LIU Xin(Sias University,He'nan Xinzheng 451150,China;School of Business,Central South University of China,Hu'nan Changsha 410083,China)
出处
《机械设计与制造》
北大核心
2023年第8期214-218,共5页
Machinery Design & Manufacture
基金
河南省2020年民办普通高等学校学科专业建设资助项目(教办政法[2020]162号,计算机科学与技术专业)。
关键词
全自主机器人
视觉技术
信息采集
导航控制
定位
Fully Autonomous Robot
Vision Technology
Information Collection
Navigation Control
Positioning
