神经网络滑模控制下肢外骨骼机器人的轨迹跟踪

Neural Network Sliding Mode Control for Trajectory Tracking of Lower Limb Exoskeleton Robot

针对下肢外骨骼机器人行走稳定性与步态轨迹跟踪控制问题,对下肢外骨骼机器人三连杆模型进行动力学建模与轨迹仿真。通过拉格朗日法建立下肢外骨骼机器人的动力学模型,设计了神经网络自适应滑模控制算法。引入神经网络,对下肢外骨骼机器人步态轨迹跟踪系统的不确定项进行逼近,在控制器中采用了改进的趋近律,使用李雅普诺夫稳定性理论进行了稳定性分析,并通过MATLAB对改进后的控制算法进行了仿真验证。仿真结果表明,采用该算法对具有关节摩擦和外界环境干扰的下肢外骨骼机器人进行轨迹跟踪时,具有较好的跟踪效果;通过改进的趋近律,能削弱系统的抖振。相比于基于计算力矩法的滑模控制,该控制算法有更好的跟踪效果,能应用到下肢外骨骼机器人行走的稳定性和步态轨迹跟踪控制中。

To solve the problems of walking stability and gait trajectory tracking control of the lower limb exoskeleton robot,dynamic modeling and trajectory simulation are carried out on the three-link model of the lower limb exoskeleton robot.On the basis of dynamic model of the lower limb exoskeleton robot established by Lagrange method,the neural network self-adaptive sliding mode control algorithm is designed.The neural network is introduced to approximate the uncertain terms of the gait trajectory tracking system of the lower limb exoskeleton robot.An improved reaching law is adopted in the controller,for which stability analysis is carried out using the Lyapunov stability theory,and for which simulation is verified by MATLAB using improved control algorithm.The simulation results show that it has a good tracking effect when the algorithm is used to track the trajectory of the lower limb exoskeleton robot with joint friction and external environment interference,and improved reaching law can weaken the chattering phenomenon of the system.Compared with the sliding mode control based on the calculated torque method,the control algorithm has better tracking effect and can be applied to the walking stability and gait trajectory tracking control of the lower limb exoskeleton robot.