CoGait WalkAssist: Design and Evaluation of a Motor-Free Lower-Limb Rehabilitation Robot for Natural Gait Training on Treadmills

Abstract

With the rising prevalence of lower-limb mobility impairments caused by stroke, neurological disorders, and accidents, there is an urgent need for innovative, cost-effective rehabilitation solutions. This paper presents the design, development, and experimental validation of CoGait, a novel gait rehabilitation robot that employs single-degree-of-freedom (DOF) multi-link mechanisms to enable natural walking patterns on commercial treadmills without independent actuators. CoGait transfers power from the treadmill to its six-link mechanisms, facilitating rehabilitation at low speeds (e.g., 0.4 m/s), which are critical for therapeutic efficacy. Key innovations include: reduced mechanical complexity (fewer linkages) compared to existing systems; ankle trajectory synthesis that accounts for both position and time, closely mimicking natural gait; self-synchronization with the treadmill, eliminating the need for external motor and control; enhanced safety and adjustability, with passive mechanisms for out-of-sagittal-plane motion and adaptability to patient anthropometrics. Functional experiments with healthy participants demonstrated that CoGait generates joint angle profiles closely resembling those of natural walking, achieved through low movement error metrics. The system recorded a root mean square error (RMSE) of 2.9° for the hip and 4.9° for the knee, alongside high correlation coefficients (CC) of 86.2% for the hip and 95.3% for the knee. These results indicate that CoGait is a promising, low-cost solution for lower-limb rehabilitation, particularly for elderly and mobility-impaired patients, and lays a strong foundation for future research in passive-assistive robotics.