An experimental testbed for intelligent control

Abstract

In this paper we discuss the development of an experimental testbed for research in intelligent robotic systems. Our system, an air hockey playing robot, consists of a direct drive robot arm, DSP controller, DataCube vision system, CCD camera and air table, is an excellent vehicle to facilitate our research in the integration of machine learning with nonlinear and adaptive control theory and real-time visual servoing. As a first application we present a new result on adaptive visual servoing with online calibration for this system.