CPG-Based Industrial Manipulators on AVR System for PTP Applications

Abstract

This paper intends to develop a cost-effective model of industrial manipulators on AVR system, which incorporates high-precision and real-time trajectory transition to perform repetitive Point-to-Point (PTP) tasks. We opt for Central Pattern Generator (CPG), a bio-inspired network that generates rhythmic trajectories without the aid of any sensory feedback. To implement this controller in AVR system, we have proposed an AVR compatible model of CPG; a custom cosine function is designed for numerical efficiency, and several techniques are outlined to ease implementation. A simulation is carried out on three degrees of freedom (DOF) manipulator as an exponent of effectiveness of the proposed methods. From simulation result, it appears that the robot generates smooth and precise trajectories in Cartesian space and has no problem dealing with either random perturbations or abrupt changes in control parameters.