Field-based robust admittance control of robots for isokinetic exercise

In recent years, isokinetic exercise has garnered significant attention for its application in muscle injury rehabilitation and muscle strength enhancement. However, existing research primarily focuses on passive isokinetic motion control for single-joint training, which limits its application scope, functionality, and, importantly, its safety guarantees. This paper proposes a novel velocity-field-based robust isokinetic admittance control scheme for multi-joint training. In the proposed control strategy, an isokinetic-attraction field based admittance control framework is developed in the outer loop to generate the desired velocity. Concurrently, a novel power-optimization non-singular integral terminal sliding mode controller (PONITSMC) is introduced in the inner loop to enable the robot to accurately track the desired admittance velocity. The PONITSMC incorporates a higher-order power proportional term and a lower-order integral term, which are designed to enhance both convergence speed and steady-state accuracy. Furthermore, a saturation function based adaptive variable-gains super-twisting algorithm (SF-AVGSTA) is developed for the PONITSMC to effectively mitigate chattering and address unmodeled dynamics and disturbances. The effectiveness of the proposed algorithm is validated through both simulations and physical experiments on a 2-DOF manipulator.