Event-triggered adaptive control for upper-limb robot-assisted passive rehabilitation exercises with input delay

Abstract

In this article, an adaptive tracking control strategy is designed for uncertain electrically driven end-effector type upper-limb rehabilitation robots subject to an input delay and a limited bandwidth channel. This control scheme is implemented to perform upper-limb passive rehabilitation training for different subjects. Primarily, dynamic analysis of the rehabilitation robot is carried out using the Euler–Lagrange principle, which incorporates motor dynamics to allow the voltage-based control commands as desirable in practical implementations. Thereafter, an adaptive backstepping control law with input delay compensation is designed to estimate the unknown dynamical parameters of the rehabilitation robot during the training sessions. Furthermore, a Lyapunov-based triggering mechanism is developed to deal with the limited bandwidth challenge and reduce the transmissions over the network. The experimental validation is conducted for different scenarios, and a comparison study is carried out with two time-triggered control schemes to investigate the potential of the proposed approach. From the experimental runs and the comparative analysis, the proposed control scheme is found to achieve a promising tracking performance with input delay compensation. Moreover, a significant saving in the network resources is attained during the passive rehabilitation training of the subjects.