Continuous Control Method for Lower Limb Rehabilitation Robot Based on Self-learning Strategy

Traditional lower limb rehabilitation robots employ discrete control methods, which are unable to decode human intention in real-time. To enhance human-machine interaction capabilities for patients with lower limb motor dysfunction, this study focuses on continuous control decoding of rehabilitation robots based on a self-learning strategy. An online classification model was developed, employing a multimodal brain-computer interface paradigm. The LGBM-LMS classification model was constructed to classify continuous control commands for the robot, converting them into control instructions for continuous control of the lower limb rehabilitation robot via electroencephalogram (EEG) signals. Experimental results indicate that this method reduces computational complexity and model redundancy, with an average recognition rate improvement of 17.82% per group.