DEVELOPEMENT OF GAIT GENERATION SYSTEM FOR A LOWER LIMB PROSTHESIS USING MEASURED GROUND REACTION FORCES

Lower limb dynamic models are useful to investigate the biomechanics of the knee and ankle joints, but many systems have several limitations which includes simplified forces, non physiologic kinematics and the complicated interactions between the foot and the ground. Many approaches in control of prosthetic devices use prediction algorithms to estimate ground reaction forces (GRF), which can degrade the performance and the efficiency of the devices due to calculation errors. In this study, the variation of the GRF during different gait cycles was investigated in the design of an adaptive fuzzy controller for a dynamic model of an active ankle-knee prosthesis, the efficiency of the controller was tested for walking gait, stair ascent and descent. Real experimental kinematics of the lower limb and GRF measured by forces platforms were selected as the controller inputs and fuzzy reasoning was used to determine the adequate torques to actuate the prosthetic device model. The capacity of the active prosthesis and the designed controller to provide walking, stair ascent and descent cycles was tested by comparing the gait kinematics to those provided by a healthy subject