Learning Post-Stroke Gait Training Strategies by Modeling Patient-Therapist Interaction

Lower-extremity hemiparesis, characterized by weakness or partial paralysis on one side of the body, significantly impacts a vast majority of stroke survivors. As a result, physical therapy intervention is crucial for the successful recovery and rehabilitation of a patient's gait cycle. During these sessions, physical therapists implement corrective forces to the patient’s limb, aiding in gait retraining. This hands-on therapy serves as the primary treatment option for stroke survivors, due to the unique gait requirements for each patient. Previous attempts of creating a lower-limb exoskeleton have been attempted but often face challenges in their ability to tailor to each patient's specific needs.

 

Researchers at Arizona State University in collaboration with researchers at Barrow Neurological Institute have developed a virtual impedance model for predicting therapist faciliatory forces in gait rehabilitation. Working directly with physical therapists, the team utilized a wearable sensing system which includes a custom-made force sensing array to accurately measure the assistive forces applied to the patient’s leg. As a result, a virtual model was created to accurately capture high-level therapist behaviors over the course of a full training session.  

 

This model provides a new approach to encode the decision-making process of physical therapists into a human-robot interaction framework for gait rehabilitation.

 

Potential Applications

Benefits and Advantages