A patient-specific muscle force estimation model for the potential use of human-inspired swing-assist rehabilitation robots

This study presents a patient-specific muscle force estimation model (PMFE) for the potential use with swing-assist rehabilitation robots. This model employs a static optimization algorithm based on a computer-based musculoskeletal model, which provides patient-specific anthropometric parameters and muscle moment arms. Joint moments and muscle forces estimated by the computed muscle control (CMC) method using kinematic data from six healthy adolescents at their comfortable speed were used to evaluate the PMFE. The simulation results show that muscle forces estimated from the PMFE match well with those from the CMC, and the PMFE has significant computational efficiency (the entire swing phase takes only 0.02 s). As a case study, a novel patient-specific biological command-based controller (PSBc), based on the PMFE, was developed to control a human-inspired swing-assist rehabilitation robot. The PSBc is evaluated by both simulation and robot experiment. The simulation results and the preliminary experiment results demonstrate that the PSBc could track both desired knee angles and desired forces. The results show the possibility of applying biomechanics methodology to the rehabilitation robot field.