摘要
针对煤矿井下狭窄空间(平均高度小于3 m,宽度小于5 m)内巷道清理机器人定位精度不足的问题,提出一种融合激光追踪测量与混合优化算法的关节误差补偿方法。通过改进型Denavit-Hartenberg(MDH)运动学模型,引入激光跟踪仪标定的关节间隙补偿项,构建含间隙正则项的混合粒子群-拟牛顿优化(HPSO-QN)逆解算法,实现全局搜索与局部优化的协同求解。仿真结果表明,补偿后机械臂工作空间拓展率分别为1.7%、0.8%、1.6%,蒙特卡洛模拟验证其安全作业边界精度提升显著。现场试验表明,所提方法使末端定位误差从74~126 mm降低至35~45 mm,误差补偿率50%~64.3%,重复定位精度标准差控制在±60 mm以内,满足井下防碰撞要求。研究成果提升了巷道清理机器人的自主化作业的定位精度。
To address the issue of insufficient positioning accuracy of roadway cleaning robot in confined spaces(average height less than 3 m and width less than 5 m)in underground coal mine,proposed a joint error compensation method that integrates laser tracking measurements with a hybrid optimization algorithm.Using the modified Denavit-Hartenberg(MDH)kinematic model,and introducing compensation term for joint gaps calibrated by the laser tracker,a hybrid particle swarm-analogue Newton optimization(HPSO-QN)inverse solution algorithm incorporating the gap compensation term was constructed to achieve synergistic approach for global search and local optimization.Simulation results show that the extension rates of the working space of the compensated robotic arm are 1.7%,0.8%and 1.6%,respectively.The Monte Carlo simulations confirm a significant enhancement in the accuracy of the robot′s safe working boundary.The field tests indicate that the proposed method reduces the end positioning error from 74-126 mm before compensation to 35-45 mm,achieving an error compensation rate of 50%-64.3%.The standard deviation of repeat positioning accuracy is maintained within±60 mm,meeting the collision avoidance requirements in underground.The research results improve the localization accuracy of the autonomous operation of the roadway cleaning robot.
作者
张幼振
路前海
李旭涛
李旺年
刘智
Zhang Youzhen;Lu Qianhai;Li Xutao;Li Wangnian;Liu Zhi(CCTEG Xi’an Research Institute(Group)Co.,Ltd.,Xi’an 710077,China)
出处
《煤矿机械》
2025年第9期43-48,共6页
Coal Mine Machinery
基金
国家重点研发计划项目(2022YFB4703600)。
关键词
巷道清理机器人
D-H参数法
定位精度
关节误差
混合逆解算法
roadway cleaning robot
D-H parametric method
localization accuracy
joint error
hybrid inverse solution algorithm
