基于混合摩擦的柔性关节机器人跟踪控制研究
|
摘要
针对柔性与摩擦的存在导致机器人关节跟踪控制效果不理想的问题,考虑柔性与摩擦建立完善的动力学模型,并设计神经网络反演复合控制器。动力学模型中以弹性扭簧模型描述关节柔性,以包含静摩擦、库伦摩擦、黏性摩擦和Stribeck摩擦的混合摩擦模型描述关节摩擦。针对柔性关节的特点设计反演控制器,并在此控制器的基础上以RBF神经网络算法实现摩擦等非线性项补偿。最后通过MATLAB仿真验证所设计控制器的有效性,并分析不同摩擦模型对跟踪精度的影响。实验结果表明:与PD控制器相比,采用所构建的复合控制器缩短了响应时间且减小了跟踪误差。该控制器可以实现柔性关节机器人轨迹的良好跟踪,提高了跟踪控制效果。
For the problems that the joint tracking control effect of the robot is not ideal due to the existence of flexibility and friction.In this paper,establishing a complete dynamic model including flexibility and friction,and designing a neural network backstepping compounded controller.In the dynamic model,the flexibility of the joint is described by the elastic torsion spring model,and the joint friction is described by a mixed friction model including static friction,Coulomb friction,viscous friction,and Stribeck friction.The backstepping controller is designed for the characteristics of flexible joints,and based on this controller,the RBF neural network algorithm is used to realize friction and other nonlinear compensation.Finally,the effectiveness of the designed controller is verified by MATLAB simulation,and the influence of different friction models on the tracking accuracy is analyzed.The results show that compared with the PD controller,the constructed compounded controller shortens the response time and reduces the tracking error.The controller can achieve a good tracking of the trajectory of the flexible joint robot and improve the tracking control effect.
For the problems that the joint tracking control effect of the robot is not ideal due to the existence of flexibility and friction.In this paper,establishing a complete dynamic model including flexibility and friction,and designing a neural network backstepping compounded controller.In the dynamic model,the flexibility of the joint is described by the elastic torsion spring model,and the joint friction is described by a mixed friction model including static friction,Coulomb friction,viscous friction,and Stribeck friction.The backstepping controller is designed for the characteristics of flexible joints,and based on this controller,the RBF neural network algorithm is used to realize friction and other nonlinear compensation.Finally,the effectiveness of the designed controller is verified by MATLAB simulation,and the influence of different friction models on the tracking accuracy is analyzed.The results show that compared with the PD controller,the constructed compounded controller shortens the response time and reduces the tracking error.The controller can achieve a good tracking of the trajectory of the flexible joint robot and improve the tracking control effect.
引文
[1]谷勇霞,张玉玲,赵杰亮,等.柔性机械臂动力学建模理论与实验研究进展[J].中国机械工程,2016,27(12):1694-1703.
[2]刘业超,金明河,刘宏.柔性关节机器人基于柔性补偿的奇异摄动控制[J].机器人,2008,30(5):460-466.
[3]张晓霞.基于Backstepping的几类机器人控制研究[D].秦皇岛:燕山大学,2012.
[4]Albu-Schaffer A,Gerd Hirzinger.A globally stable state feedback controller for flexible joint robots[J].Advanced Robotics,2001,15(8):799-814
[5]潘冬.空间柔性机械臂动力学建模分析及在轨抓捕控制[D].哈尔滨:哈尔滨工业大学,2014.
[6]王雪竹,李洪谊,王越超,等.柔性关节机器人高精度自适应反步法控制[J].信息与控制,2016,45(1):1-7.
[7]Chu M,Jia Q,Sun H.Backstepping Control for Flexible Joint with Friction Using Wavelet Neural Networks and L2‐Gain Approach[J].Asian Journal of Control,2017,10.1002/asjc.1602.
[8]Spong M W.Reviewarticle:Modeling and control of elastic joint robots[J].Mathematical&Computer Modelling,1989,12(7):912-912.
[9]丁千,翟红梅.机械系统摩擦动力学研究进展[J].力学进展,2013,43(1):112-131.
[10]柴杰,江青茵,曹志凯.RBF神经网络的函数逼近能力及其算法[J].模式识别与人工智能,2002,15(3):310-316.
[2]刘业超,金明河,刘宏.柔性关节机器人基于柔性补偿的奇异摄动控制[J].机器人,2008,30(5):460-466.
[3]张晓霞.基于Backstepping的几类机器人控制研究[D].秦皇岛:燕山大学,2012.
[4]Albu-Schaffer A,Gerd Hirzinger.A globally stable state feedback controller for flexible joint robots[J].Advanced Robotics,2001,15(8):799-814
[5]潘冬.空间柔性机械臂动力学建模分析及在轨抓捕控制[D].哈尔滨:哈尔滨工业大学,2014.
[6]王雪竹,李洪谊,王越超,等.柔性关节机器人高精度自适应反步法控制[J].信息与控制,2016,45(1):1-7.
[7]Chu M,Jia Q,Sun H.Backstepping Control for Flexible Joint with Friction Using Wavelet Neural Networks and L2‐Gain Approach[J].Asian Journal of Control,2017,10.1002/asjc.1602.
[8]Spong M W.Reviewarticle:Modeling and control of elastic joint robots[J].Mathematical&Computer Modelling,1989,12(7):912-912.
[9]丁千,翟红梅.机械系统摩擦动力学研究进展[J].力学进展,2013,43(1):112-131.
[10]柴杰,江青茵,曹志凯.RBF神经网络的函数逼近能力及其算法[J].模式识别与人工智能,2002,15(3):310-316.