Open AccessDesign of an Intelligent Controller for Above Knee Prostheses based on an Adaptive Neuro-Fuzzy Inference System
Author(s) -
Dhirgaam A. Kadhim,
Mithaq Nama Raheema,
Jabbar S. Hussein
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/671/1/012066
Subject(s) - adaptive neuro fuzzy inference system , knee joint , sagittal plane , controller (irrigation) , gait , computer science , gait analysis , ground reaction force , exoskeleton , work (physics) , prosthesis , simulation , kinematics , fuzzy logic , physical medicine and rehabilitation , artificial intelligence , engineering , fuzzy control system , medicine , mechanical engineering , surgery , anatomy , physics , classical mechanics , biology , agronomy
The number of Above Knee (AK) amputees has increased in recent years and this has led to a need for urgent work on the design of proper lower limb prostheses. Lower limb prosthetics can be divided into active and passive devices. However, passive prosthetics cannot fully provide the natural motion of a healthy leg, and the technologies used in active prosthetics with knee joints are often far too expensive for amputees in developing countries such as Iraq. In this paper, an active lower limb prosthesis with an efficient knee joint is thus designed. Two strategies were used to collect data for gait cycle analysis of the leg in the sagittal plane: the first was based on the use of a force platform device to obtain the foot ground force according to the foot position (x, y), while the second utilised a video-camera based system to examine knee joint angles. The obtained data were all sent to an intelligent controller that uses an Adaptive Neuro-based Fuzzy Inference System (ANFIS). The ANFIS controller determines the ground force, mimicking the moment of the active knee with a DC motor and flexion-extension angle values. The experimental data for the motion of the knee joint were collected in the Gait Laboratory, then transformed to joint angles using the ANFIS controller. The results show excellent response in the proposed ANFIS controllers in terms of determining angle and moment values of the knee joint with a very low RMS error of 0.006.