基于模型预测控制的协作焊接双机械臂轨迹跟踪算法
|
摘要
焊接双机械臂协作焊接时,存在从动机械臂对主机械臂末端位置的轨迹跟踪的精确性问题,课题组提出了基于模型预测控制算法的主/从位置协调控制方法。根据所建立的六自由度双机械臂的运动学模型,建立末端位置与关节角度的变换关系以及机械臂的运动预测模型;通过上述模型,对主机械臂采用位置控制方法,并通过从动机械臂末端的三维激光扫描仪测距传感器所获取的主机械臂位置及方向,对从动机械臂采用基于模型预测控制算法的位置控制;根据机械臂关节角度变换旋转运算序列及末端位置的预测模型,通过动态矩阵控制算法,由当前时刻的位置状态及下一时刻位置输入状态对未来某时域内的位置输出状态进行预测,从而实现期望的位置跟踪。最后,采用仿真实验测试来验证该算法的实用有效性,结果表明:从动机械臂末端在有较小超调的情况下快速达到稳态,实现对期望轨迹的跟踪;与传统PID控制算法相比,模型预测控制算法能够更快速、更稳定地达到对期望轨迹的跟踪效果;该算法对期望轨迹的跟踪误差在有小幅波动的情况下可保持在±0. 05 mm之内,在无波动的情况下可迅速趋近于零。基于该算法,从动机械臂可根据主机械臂末端动态位置信息更加有效地实现对期望轨迹的实时跟踪。
Research group presents a master-slave position coordinated control method based on model predictive control algorithm( MPC),to improve the trajectory tracking accuracy of slave arm according to the master arm end position when two manipulators are coordinating motion to execute welding tasks. According to the kinematics model of 6-DOF dual manipulators,transformation relationship between end position and joint angle and motion predictive model of manipulator were built; Through models mentioned above,position control was adopted for master arm and position control based on MPC algorithm was adopted for the slave arm where a distance measuring sensor with 3D laser scanner was installed in the end position to obtain position and direction information of master arm; On the basis of rotation operation sequence of joint angle transformation and predictive model of end position for manipulators,the current position state and the position input state of next step were used for predicting position output state during some future time domain with dynamic matrix control algorithm,to realize the desired trajectory tracking. Finally,simulation testing for effectiveness of this algorithm was implemented,and then result shows that slave arm end with smaller overshoot can reach steady state rapidly to realize desired trajectory tracking; compared with traditional PID control algorithm,MPC algorithm can achieve desired trajectory tracking more rapidly and more stable; tracking error of desired trajectory based on this algorithm can be kept within ± 0. 05 mm with slight fluctuation or approach to zero rapidly without fluctuation.Based on this algorithm,slave arm can more effectively realize real-time tracking desired trajectory according to dynamic information from the end position of master arm.
Research group presents a master-slave position coordinated control method based on model predictive control algorithm( MPC),to improve the trajectory tracking accuracy of slave arm according to the master arm end position when two manipulators are coordinating motion to execute welding tasks. According to the kinematics model of 6-DOF dual manipulators,transformation relationship between end position and joint angle and motion predictive model of manipulator were built; Through models mentioned above,position control was adopted for master arm and position control based on MPC algorithm was adopted for the slave arm where a distance measuring sensor with 3D laser scanner was installed in the end position to obtain position and direction information of master arm; On the basis of rotation operation sequence of joint angle transformation and predictive model of end position for manipulators,the current position state and the position input state of next step were used for predicting position output state during some future time domain with dynamic matrix control algorithm,to realize the desired trajectory tracking. Finally,simulation testing for effectiveness of this algorithm was implemented,and then result shows that slave arm end with smaller overshoot can reach steady state rapidly to realize desired trajectory tracking; compared with traditional PID control algorithm,MPC algorithm can achieve desired trajectory tracking more rapidly and more stable; tracking error of desired trajectory based on this algorithm can be kept within ± 0. 05 mm with slight fluctuation or approach to zero rapidly without fluctuation.Based on this algorithm,slave arm can more effectively realize real-time tracking desired trajectory according to dynamic information from the end position of master arm.
引文
[1]李开霞,丁玲,赵江海,等.双臂机器人末端一体化关节控制与驱动单元设计[J].机电工程,2017,34(9):1019-1023.
[2]HOGAN N.Impedance control:an approach to manipulation[J].Asme Transactions Journal of Dynamic Systems Measurement and Control,1985,107(1):304-313.
[3]MEER D W,HSIA T C.Experiment in cooperative manipulation of flexible objects[D].Stanford,CA:Stanford University,1995:25.
[4]JUNG S.Neural network impedance force control of robot manipulator[J].IEEE Transactions on Industrial Electronics,1988,45(3):451-461.
[5]AGRAVANTE D J,CLAUDIO G,SPINDLER F,et al.Visual servoing in an optimization framework for the whole-body control of humanoid robots[J].IEEE Robotics and Automation Letters,2017,2(2):608-615.
[6]HUANG Panfeng,YUAN Jianping,LIANG Bin.Adaptive slidingmode control of space robot during manipulating unknown objects[C]//IEEE International Conference on Control&Automation.Guangzhou,China:IEEE,2007:2907.
[7]AGHILI F.Adaptive control of manipulators forming closed kinematic chain with inacurate kinematic model[J].IEEE/ASME Transactions on Mechatronics,2013,18(5):1544.
[8]CHENG Wei,WANG Xuelin,MA Haiyan.Master-slave force control based on grey GM(1,1)model of robot gripper[C]//IEEEInternational Conference on Advanced Mechatronic Systems.Beijing,China:IEEE,2015:22-24.
[9]邹涛,丁宝苍,张端.模型预测控制工程应用导论[M].北京:化学工业出版社,2010:120-135.
[10]刘永信,王玲琳,韩晓爽,等.双机械臂协调控制综述[J].内蒙古大学学报(自然科学版),2017,48(4):471-480.
[11]朱齐丹.六自由度机械臂逆运动学算法[J].机器人技术与应用,2014(2):12-18.
[12]JIANG Ming,FAN Mingqu,LI Aimin,et al.Coordination control of dual-arm robot based on modeled predictive control[C]//IEEEInternational Conference on Real-time Computing and Robotics.Angkor Wat,Cambodia:IEEE,2016:495-499.
[13]李国勇.智能预测控制及其MATLAB实现[M].北京:电子工业出版社,2010:255-298.
[14]CORKE P.Robotics,vision and control:fundamental algorithms in MATLAB[M].Berlin:Springer Publishing Company,2013:189-280.
[2]HOGAN N.Impedance control:an approach to manipulation[J].Asme Transactions Journal of Dynamic Systems Measurement and Control,1985,107(1):304-313.
[3]MEER D W,HSIA T C.Experiment in cooperative manipulation of flexible objects[D].Stanford,CA:Stanford University,1995:25.
[4]JUNG S.Neural network impedance force control of robot manipulator[J].IEEE Transactions on Industrial Electronics,1988,45(3):451-461.
[5]AGRAVANTE D J,CLAUDIO G,SPINDLER F,et al.Visual servoing in an optimization framework for the whole-body control of humanoid robots[J].IEEE Robotics and Automation Letters,2017,2(2):608-615.
[6]HUANG Panfeng,YUAN Jianping,LIANG Bin.Adaptive slidingmode control of space robot during manipulating unknown objects[C]//IEEE International Conference on Control&Automation.Guangzhou,China:IEEE,2007:2907.
[7]AGHILI F.Adaptive control of manipulators forming closed kinematic chain with inacurate kinematic model[J].IEEE/ASME Transactions on Mechatronics,2013,18(5):1544.
[8]CHENG Wei,WANG Xuelin,MA Haiyan.Master-slave force control based on grey GM(1,1)model of robot gripper[C]//IEEEInternational Conference on Advanced Mechatronic Systems.Beijing,China:IEEE,2015:22-24.
[9]邹涛,丁宝苍,张端.模型预测控制工程应用导论[M].北京:化学工业出版社,2010:120-135.
[10]刘永信,王玲琳,韩晓爽,等.双机械臂协调控制综述[J].内蒙古大学学报(自然科学版),2017,48(4):471-480.
[11]朱齐丹.六自由度机械臂逆运动学算法[J].机器人技术与应用,2014(2):12-18.
[12]JIANG Ming,FAN Mingqu,LI Aimin,et al.Coordination control of dual-arm robot based on modeled predictive control[C]//IEEEInternational Conference on Real-time Computing and Robotics.Angkor Wat,Cambodia:IEEE,2016:495-499.
[13]李国勇.智能预测控制及其MATLAB实现[M].北京:电子工业出版社,2010:255-298.
[14]CORKE P.Robotics,vision and control:fundamental algorithms in MATLAB[M].Berlin:Springer Publishing Company,2013:189-280.