A Body–Machine Interface for Assistive Robot Control in Spinal Cord Injury: System Description and Preliminary Tests

Abstract

Motor impairments, particularly spinal cord injuries, impact thousands of people each year, resulting in severe sensory and motor disabilities. Assistive technologies play a crucial role in supporting these individuals with activities of daily living. Among such technologies, body–machine interfaces (BoMIs) are particularly important, as they convert residual body movements into control signals for external robotic devices. The main challenge lies in developing versatile control interfaces that can adapt to the unique needs of individual users. This study aims to adapt for people with spinal cord injury a novel control framework designed to translate residual user movements into commands for the humanoid robot Alter-Ego. After testing and refining the control algorithm, we developed an experimental protocol to train users to control the robot in a simulated environment. A total of 12 unimpaired participants and two individuals affected by spinal cord injury participated in this study, which was designed to assess the system’s applicability and gather end-user feedback on its performance in assisting with daily tasks. Key metrics such as the system’s usability, accuracy, performance, and improvement metrics in navigation and reaching tasks were assessed. The results suggest that assistive robots can be effectively controlled using minimal residual movements. Furthermore, structured training sessions significantly enhance overall performance and improve the accuracy of the control algorithm across the selected tasks.