Gait-Generation Strategy for Lower Limb Exoskeleton Based on Central Pattern Generator
Wen Duan,Weihai Chen,Jianhua Wang,Jingmeng Liu,Zhongcai Pei,Jianer Chen
DOI: https://doi.org/10.1109/tmech.2024.3367348
2024-01-01
IEEE/ASME Transactions on Mechatronics
Abstract:To improve human–machine interactions, many studies have replicated the human motor nervous system to control lower limb exoskeletons. However, this approach is hindered by its intricacy and the disparities between human and machine capabilities, leading to suboptimal adaptability and constrained practicality. This article presents a gait-generation strategy for lower limb exoskeletons utilizing central pattern generators (CPGs). This method offers a simplified model compared with the traditional CPG methods, significantly easing the parameter optimization process. The strategy is bifurcated into two segments. The initial segment encapsulates the human motor nervous system, constructs a musculoskeletal structure, and generates a basic bionic gait. The subsequent segment employs adaptive oscillators to assimilate and refine this basic gait through weak coupling theory. It further broadens the application range by integrating variational controllers and state observers. We implemented this strategy in a bench-type lower limb exoskeleton, allowing adjustments in walking speed and integrating a preference control algorithm for personalized exoskeleton management. Experimental evaluations confirmed the efficacy of this control system. The results indicated that our system can overcome the limitations of fixed training modes in bench-type lower limb exoskeletons and can adapt to the unique gait characteristics of various users, thereby facilitating gait customization.
automation & control systems,engineering, electrical & electronic, manufacturing, mechanical