Vision-Based Gesture-Driven Control Method for Rehabilitation Lower Extremity Exoskeleton Robots

Recent advances in Rehabilitation Lower Extremity Robotic Exoskeletons (RLEEX) have driven significant improvements in Human Robot Interaction (HRI) methods, including traditional physical interaction and bioelectrical signal approaches. However, most existing HRI methods rely on contact-based interaction, which intro duces challenges such as operational complexity, susceptibility to interference, risks associated with implantable devices, and comfort limitations. These issues hinder intuitive and natural interaction, potentially reducing patient motivation during rehabilitation. To address these limitations, this study proposes a novel non-contact gesture interaction control method for RLEEX, utilizing RGB monocular camera depth estimation. The method comprises three core steps: key points detection, gesture recognition, and distance assessment, integrating gesture information with augmented reality triggering technology to control RLEEX gait movements. Experimental results demonstrate that this approach effectively mitigates issues of poor comfort, low reliability, and high latency in contact-based HRI. Specifically, the non-contact method achieves a gesture-controlled exoskeleton motion accuracy of 94.11% and an average system response time of 0.615 seconds. This innovative non-contact HRI framework represents a pioneering advancement in RLEEX control interactions, offering a promising direction for future research and development in the field.