Model-free based neural network control with time-delay estimation for lower extremity exoskeleton

Abstract

A model-free based neural network control with time-delay estimation (TDE-MFNNC) for lower extremity exoskeleton is presented in this paper. The lower limb exoskeleton which has 5 DOFs for each leg is established in Solidworks as a virtual prototype which is used as a platform to build the control system in SimMechanics. In an attempt to get the effective tracking trajectory, a neural network and time-delay estimation technique is added to model-free based iPD controller. The proposed controller is simulated and tested on virtual prototype to comparing with PD controller, neural network and model-free based iPD controller. This comparative study validates TDE-MFNNC as more stable and effective than the traditional controllers. Further, the kinematic model of lower limb exoskeleton added to exoskeleton also proves the stability and effectiveness of the proposed controller.