Practical active force control with iterative learning scheme applied to a pneumatic artificial muscle actuated robotic arm

Abstract

The precise motion control of a pneumatic artificial muscle (PAM) actuated system poses a great challenge to researchers due to the inherent nonlinearities, time-varying parameters, and high sensitivity to payload of the PAM mechanism. This paper highlights the effective practical implementation of an active force control (AFC) technique incorporating an iterative learning (IL) algorithm known as AFCAIL applied to a two-link planar robotic arm actuated by a pair of PAMs. The iterative learning is primarily used as a technique to compute the best value of the estimated inertia matrix of the robot arm required for the AFC loop that is complemented with a conventional proportional-integral-derivative (PID) control. An experimental rig utilizing a hardware-in-the-loop simulation (HILS) configuration was designed and developed based on suitable hardware and software installation. A number of experiments were carried out to validate the theoretical counterpart considering the independent joint control and coordinated motion control of the system for a given operating and loading conditions. The results of the experimental works verify the effectiveness and robustness of the proposed PAM actuated AFCAIL scheme in executing a number of trajectory tracking tasks.