Predicting the Effects of Walker Height and Weight Support on Assisted Gait Using Physics-Based Predictive Simulations

Walker-assisted gait is widely used in clinical rehabilitation for individuals with muscle weakness and balance impairments. This study presents a first step toward developing a predictive simulation framework that integrates a 3D full-body musculoskeletal model driven by muscle torque generators within an optimal control problem. We calibrated the muscle torque generators model using experimental isometric and isokinetic data from a healthy participant, obtained from a biodex dynamometer equipment. To assess the predictive capability of the framework, we evaluated the effects of walker height and percentage of body weight support on walker-assisted gait patterns by running nine predictive simulations across varying walker configurations. Results showed that effects of walker height were well predicted (e.g., elbow flexion increased with walker height from a mean value of 81.59 deg to 93.86 deg), while effects of weight support were only partially predicted (e.g., upper body joints did not show a clear trend with changes in weight support). Results suggest that developing a detailed hand-walker interaction model would significantly improve the realism of the simulations. This study provides an important step toward optimizing walker-assisted gait through simulation-based design and personalization.