A human with muscle weakness can wear a rehabilitation device, e.g. exoskeleton, to receive support during walking. However, how can this support be adjusted in real-time to the intention of the user? The given support should be intuitive, user-specific, compliant and robust to sensor noise. We suggest the use of an adaptive oscillator which can detect the frequency and phase of the user’s gait. The designed oscillator-based controller is capable to adjust the support in real time, based on reliable joint angle measurements. The adaptive oscillator determines the position and velocity reference trajectories, based on the measured joint angles. The controller uses these reference trajectories to attract the current angle respectively velocity. This is part II of a two-part paper. In part I of this paper the controller is optimized using simulations. In part II the performance of the optimized oscillator-based controller is determined for the ankle joint using specially designed pneumatic ankle foot orthoses (PAFO). The PAFO actuate the plantar exion movement with pneumatic muscles. The performance of the controller is analyzed measuring the activity of the Tibialis Anterior and Gastrocnemius Medialis with electromyography (EMG). The oscillator-based controller found the same reference joint angles and velocities in real time compared to the optimization. The desired exoskeleton torques are comparable to the simulations, but the amplitude is lower. The support decreased the EMG activity of the Gastrocnemius Medialis. However, the lack of transparency of the PAFO increased the EMG of the Tibialis Anterior. Overall the oscillator-based controller can give a user-specic support by detecting the gait frequency and learning the gait trajectory. The specially designed PAFO can measure the performance of controllers on humans objectively. The EMG of the Gastrocnemius Medialis can decrease more, when the transparency of the PAFO is increased.

Категория: Научные статьи | Нет комментариев »

Комментарии