摘要
为了提高串联机器人的末端绝对精度,本文首先采用轴线测量法识别机器人D-H参数模型,继而将D-H参数转化为最小完整连续运动学(CMMK)模型参数,并进行非线性优化,以解决D-H参数模型奇异和冗余带来的非线性优化不易收敛问题,最后,将优化后的CMMK模型参数转化为工业标准的D-H模型参数,再经过补偿后将其作为设计模型以获得更高精度的定位.通过在MOTOMAN-MH80机器人上进行试验,该方法能确实有效地识别机器人的杆件参数,未标定前机器人的位置误差只能达到2 mm左右,而标定后降至0.7 mm左右,精度提高了将近70%.本文方法通过轴线测量获取机器人模型参数初值,避免了对机器人进行理论建模的过程,与运动学回路法相比,具有较高的通用性;采用最小完整连续运动学模型进行标定,能有效解决D-H模型奇异性、非连续、不易收敛到正确值的问题.
This paper intends to improve the absolute accuracy of the end-effector of serial robot. Firstly, the axis measurement method is used to identify the D-H（Denavit-Hartenberg） model parameters, and the D-H model parameters are transformed into the complete, minimal and model-continuous kinematic（CMMK） model parameters. Then, the parameters are optimized by the nonlinear method to solve the non-convergence problem of nonlinear optimization, which is caused by the singularity and redundancy of the D-H parameter model. Finally, the optimized CMMK model parameters are transformed into the industrial D-H model parameters, according to which the theoretical D-H parameter model is modified to accomplish high-accuracy calibration and position control. In experiments on MOTOMAN-MH80 robot, the robotic link parameters are effectively identified with the proposed method, and the positioning error decreases from 2 mm to 0.7 mm after calibration,ithe positioning accuracy is improved by about 70%. The proposed method can be used in more cases comparing with the kinematic loop method, because it obtains initial values of robot model parameters by axis measurement method and avoids the theoretical modeling process. Plus, problems of singularity, discontinuity and non-convergence of D-H model are solved effectively through the CMMK model based calibration.
出处
《机器人》
EI
CSCD
北大核心
2016年第3期360-370,共11页
Robot
基金
国家自然科学基金(51575332)
机械系统与振动国家重点实验室课题(MSV2015010)
上海市教育委员会科研创新项目(2014Z10280034)
关键词
串联机器人
绝对精度
轴线测量法
最小完整连续运动学模型
serial robot
absolute accuracy
axis measurement method
CMMK（complete
minimal and model-continuous kinematic） model
