High-Level Parametric Gait Modeling for Assistive Robotics

Through a myriad of physiological ailments, many people find themselves with gait pathologies which affect their daily lives. Using gait assistance robotics, those affected may regain their balance, strength, and ultimately, their confidence in their own volition. To assist, exoskeleton devices have been in development for some time now. However, an adaptive control algorithm is required to allow for dynamic usage. Direct motor control using bio-signals has been implemented in research and clinical studies with success, however, in bio-signals such as electromyography (EMG), current technologies dissuade implementation of these in direct motor control, primarily due to the high level of variance, nonstationary nature of the signals, and the sensitivity of electrodes to external disturbances such as sweat and friction. For this reason, this study looks to bypass the motor control, and perform high-level gait parameter estimation to allow for an adaptive gait assistance device which would be adaptive to the user and at the same time, robust enough to allow for real life applications, and end-user products.