On Controlling a Pediatric Lower-Limb Exoskeleton Using Finite-State Machine and Electroencephalography

Abstract

There are approximately 330,000 American children aged 5 to 17 who suffer from ambulatory difficulties that adversely impact their quality of life. Robotic-Assisted Gait Training (RAGT) has been shown to be an effective intervention for motor-related conditions in children with mobility disabilities. Compared to traditional therapy, RAGT enables longer duration, higher intensity training sessions, more precise movement patterns, and decreased physical demands on therapists. Furthermore, the advent of portable, powered exoskeletons, while still relatively rare in pediatric applications, has significantly increased accessibility to treatment, allowing for more frequent therapy sessions. Integrating that with a brain-computer interface (BCI) could further optimize the learning process and accelerate motor function recovery. To address this need, the Laboratory for Non-Invasive Brain-Machine Interface Systems at the University of Houston in collaboration with Center for Wearable Exoskeletons at TIRR Memorial Hermann has developed a Pediatric Lower-Extremity Gait System (P-LEGS). This dissertation aims at building a control system for P-LEGS that can utilize explicit manual input or electroencephalography (EEG) signals to command the device to generate movement trajectories. The process of developing, evaluating, and prototyping the various levels of control for the system, as well as the proposed movement intent decoding pediatric BCI, is delineated herein. This dissertation contributes to the pediatric rehabilitation community by introducing a novel exoskeleton system that can provide customizable Robotic-Assisted Gait Training and function as a mobility assistive device. Furthermore, it advances the fields of neural engineering and neuroscience through the development of a pediatric brain-computer interface capable of decoding movement intent.