Human-in-the-Loop Control of a Hip Assistive Exoskeleton Based on Cross-Limb Virtual Force Transfer

Abstract

In industries where various unavoidable heavy lifting tasks are prevalent, workers are at high risk of developing musculoskeletal disorders. It is of great research value to relieve the physical stress of workers at work, reduce the risk of musculoskeletal diseases, and improve the work efficiency of workers. This article introduces a hip joint assistive exoskeleton robot that can provide a maximum assist torque of up to 80 Nm. We propose a human-in-loop control scheme for cross-limb virtual force transfer, with the human as the control leader. The exoskeleton is controlled based on surface electromyography (EMG) signals. The signal is filtered, normalized, and fed into a neural network model to estimate human joint torque. The estimated virtual force is used as the inner loop force control trajectory to assist the user. The assist torque can be adaptively adjusted according to the size of the grabbed heavy object. This solution collects the signals from the upper limbs, controls the exoskeleton to assist the lower limbs, and adjusts the assistance level in real time according to the preferences of user. The experimental results demonstrate that when using exoskeleton assistance, the muscle activity of the thighs of the subjects can be reduced by up to 43.82%.