Multi-objective optimization of Force Transmission Quality and Joint Misalignment of a 5-Bar Knee Exoskeleton

Abstract

The ability to follow a normal anatomical motion is one of the fundamental design requirements of wearable exoskeleton. Human knee motion is a combination of sliding and rolling actions, which can not be replicated by a simple kinematic pair with a fixed rotation axis. In this paper, a polycentric robotic knee exoskeleton is designed based on the 5-bar mechanism (5- BM) to reproduce knee motion. The Genetic Algorithm finds the optimum parameters of the 5-BM by minimizing a weighted cost function which consists of the average Joint Force Index (JFI) and the misalignment between the centre of rotation of the exoskeleton and the user’s knee. An average and maximum ICR error of 0. 16mm and 0. 43mm is obtained by the optimized 5-BM. It yielded an average actuation toque of 7. 33N.m and JFI of 4.88 compared to 217. 73N.m and 6.04 obtained by the widely used 4-BM.