REDUCTION OF JERK THROUGH OPTIMIZATION OF A KNEE ASSISTIVE DEVICE DESIGNED USING FOUR-BAR CONTROLLED COMPLIANCE ACTUATOR

摘要

Two major responsibilities are taken by a knee joint during level ground walking of a human being. At stance phase, knee joint is locked in order to provide stability against ground reaction force, and in swing phase, knee flexes for ground clearance of the foot. In this study, a four-bar controlled compliance actuator (FCCA) has been configured for assisting the knee joint during walking. The main focus of FCCA design is the reduction of required mechanical power through a compliance actuation strategy and amplification of motor power. The proposed design consists of two linear spring-damper systems, out of which, one controls the locking of knee during stance phase and another takes care of knee flexion at swing phase. Without a proper selection of stiffness and damping coefficient for both the springs, the knee may be subjected to jerk during flexion, which may give rise to discomfort to the user and consequently, he/she may fall during walking. This study aims to ensure smooth knee operation in both stance phase as well as swing phase by assigning the proper spring stiffness and damping coefficient for both the springs. The responsibility is given to a non-traditional optimization tool, namely particle swarm optimization (PSO). The optimization is carried out using a co-simulation approach between Matlab and ADAMS. The aim of PSO (run in Matlab) is to minimize both the frequency as well as amplitude of the jerk by finding a suitable set of design variables, that is, spring stiffness and damping coefficient of two spring-damper systems.