动力学模型不确定的上肢外骨骼自适应滑模控制  

Adaptive Sliding Mode Control for the Upper-limb Exoskeleton with Uncertain Dynamic Model

作　　者：刘健 弓正菁 张燕 卢宁 LIU Jian;GONG Zheng-jing;ZHANG Yan;LU Ning(School of Artificial Intelligence and Data Science,Hebei University of Technology,Tianjin 300130,China)

机构地区：[1]河北工业大学人工智能与数据科学学院,天津300130

出　　处：《控制工程》2023年第1期54-61,共8页Control Engineering of China

基　　金：国家自然科学基金资助项目(61773151,61703135);河北省自然科学基金资助项目(F2018202279)。

摘　　要：为了解决上肢外骨骼系统中动力学模型不确定情况下的轨迹跟踪问题,提出了一种自适应滑模控制策略。该控制策略采用径向基函数(radial basis function,RBF)神经网络对动力学模型进行估计,并根据估计值设计自适应滑模控制律。利用李雅普诺夫稳定性判据证明控制系统的稳定性。为验证控制策略的有效性,使用Kinect传感器获取人体上肢动作关节角度变化曲线,并使用SolidWorks建立人体和上肢外骨骼模型,同时将其导入Adams中,进行Adams和MATLAB联合仿真。最后,仿真结果表明,实际关节角度与期望关节角度的误差在收敛后小于0.01,即使在动态模型不确定的情况下,所提控制策略也能获得优异的控制效果。An adaptive sliding mode control strategy was proposed to solve the trajectory tracking problem in the upper limb exoskeleton system with uncertain dynamics model.The control strategy estimated the dynamic model by using a radial basis function(RBF)neural network and designs an adaptive sliding mode control law based on the estimated value.The Lyapunov theory was employed to prove the stability of control system.In order to verify the effectiveness of the control method,Kinect sensors were used to obtain the joint angle curve of the human upper limb movement.SolidWorks was utilized to build the human body and the upper limb exoskeleton model,which were imported into Adams.Adams and MATLAB were adopted for joint simulation.Finally,the simulation results show that the error between the actual joint angle and the expected joint angle is less than 0.01 after convergence,and the proposed control strategy can obtain excellent control effect even when the dynamic model is uncertain.

关 键 词：动力学仿真 滑模控制 上肢外骨骼 RBF神经网络 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]