Mechanism design and motion control of a parallel ankle joint for rehabilitation robotic exoskeleton

Comparing with hip and knee, the design of exoskeleton ankle is much more difficult due to the strict requirements of smaller space, better rigidity and heavier load. A novel ankle exoskeleton with 3-RPS (Revolute- Prismatic-Spherical) parallel mechanism, which can fully sustain the heavy load of human body with good dynamic and kinematic performances, has been conducted to assist rehabilitation of the physically weak persons. The 3-RPS parallel mechanism of ankle joint is optimized in detail. The skin surface electromyographic (sEMG) signals of muscles are applied as main input signals. By preprocessing the sEMG signals, a new neuro-fuzzy controller is developed to predict the user´s motion and control the robotic exoskeleton in real time. The experimental results prove that EMG-based neuro-fuzzy controller is effective, and the parallel ankle with higher stiffness and lighter weight meets the kinematical and dynamical requirement for rehabilitation.