A VISION-GUIDED FUZZY-LOGIC CONTROL-SYSTEM FOR DYNAMIC PURSUIT OF A MOVING TARGET

We investigate the use of linguistic rules based on 'rules-of-thumb' e xperiences and engineering judgments to guide an industrial robot to f ollow a moving target using visual information. The problem has been f ormulated in the context of Prey Capture with the robot as a 'pursuer' and a moving object as a passive 'prey'. Such a formulation mimics th e function and capability of a natural being to pursue its prey. The f easibility of the fuzzy logic control strategy was verified experiment ally. The experiments, built upon the framework of a low-cost flexible integrated vision system developed at Georgia Institute of Technology , were performed on a vibratory feeder with robotic hand-eye coordinat ion. Experiments demonstrated that this approach does not require an a ccurate description of the dynamics of the pursuit process or that of the target's motion, nor does it require the goals and constraints of the system to be quantified as single numerical values. These attracti ve features of the control strategy mean it has significant potential in industrial applications where models of the controlled process are not available, but the operator's experience may be used as a guide to formulate the control rules. (C) 1998 Elsevier Science B.V.