功能性电刺激下的关节自适应运动控制研究

Adaptive Control of Joint Movement Induced by Electrical Stimulation

针对功能性电刺激(Functional electrical stimulation,FES)下外部干扰和肌肉疲劳对关节运动的影响,提出了一种神经网络自适应滑模控制方法以获得更加精确的关节运动.本文建立了电刺激下的关节运动模型,在此模型的基础上设计了滑模控制律,利用径向基神经网络在线逼近系统不确定特性,并通过Lyapunov方法设计了径向基神经网络的自适应律,以电刺激所产生的膝关节运动控制为例,通过仿真和实验研究验证了该神经网络滑模控制方法相对于传统的滑模控制来说,不仅可以准确地控制电刺激而获得期望的关节运动,而且当关节运动受到外部干扰和肌肉疲劳的影响时,还可自适应地对此进行补偿,有效地调节电刺激强度以获得准确的关节运动.

This paper presents a neuro sliding mode control method of electrical stimulation for accurate electrically- induced joint movement by compensating the effects of external disturbances and muscle fatigue during stimulation. The sliding mode control law is rested on an electrically-induced musculoskeletal model. The adaptive control law of the radial basis function network which is used to approximate system modeling uncertainties is derived through the Lyapunov function. This proposed method is evaluated by adaptive control of electrical stimulation to achieve expected knee movements, especially in the presence of external disturbances and muscle fatigue. Both simulation and experimental studies indicate that the proposed adaptive control method is effective and feasible to compensate deviations of joint movement resulting from external disturbances and muscle fatigue.