Adaptive Iterative Learning Control design for RUPERT IV

Abstract

An adaptive robot control strategy combining PID-based feedback and Iterative Learning Controller (ILC) is proposed for performing passive reaching tasks for RUPERT™ (Robotic Upper Extremity Repetitive Therapy), a pneumatic muscle driven rehabilitation robotic device. Additionally, a fuzzy rule-base is used to estimate the ILC learning rate to achieve an optimized learning. The preliminary test results from two able-bodied subjects demonstrate that the proposed controller can provide consistent performance for different subjects performing different reaching tasks.