Design an intelligent hybrid position/force control for above knee prosthesis based on adaptive neuro-fuzzy inference system

Abstract

This paper reviews the position/force control approach for governs an efficient knee joint in an active lower limb prosthesis, and the interfacing current control algorithm with human gate parameter is inserted. Two techniques are used to collect gait cycle data of leg: first, the foot ground force is obtained by the force platform device based on its position (x, y), then data of knee joint angles is recorded by using a video-camera device. The collected information is sent and used in the proposed intelligent controller. This intelligent control system used an adaptive neuro-fuzzy inference system (ANFIS) circuit in addition to the proportional integral derivative (PID) controller. This hybrid ANFIS-PID control system simulates and provides the ground force values. The experimental results show an excellent response and lower root mean square error (RMSE) compared with each of PID and ANFIS controller that implemented for a similar purpose. In summary, the results showed acceptably stable performance of the proposed position/force controller based on hybrid ANFIS-PID system. It can be concluded that the finest performance of the controlled force, as quantified by the RMSE criteria, is perceived by the proposed hybrid scheme depending on the controller intelligent decision circuit.