DEVELOPMENT OF A THREE DIMENSIONAL VISION ALGORITHM FOR DIRECT TRANSFORMATION FROM IMAGE SPACE TO ROBOT JOINT SPACE

A DIRECT TRANSFORMATION FROM IMAGE COORDINATES TO ROBOT JOINT COORDINATES WILL BE DEVELOPED TO SERVE AS THE BASIS FOR A BINARY VISION SYSTEM FOR SIX DEGREE OF FREEDOM ROBOT CONTROL. A SYSTEM IDENTIFICATION ALGORITHM WILL ENABLE THE SYSTEM TO BE USED AS A TURN-KEY SYSTEM WITH ANY ROBOT WHICH ACCEPTS JOINT COORDINATE INPUT. THE DIRECT TRANSFORMATION FROM IMAGE SPACE TO ROBOT JOINT SPACE WILL BE ACCOMPLISHED BY DEVELOPING A NONLINEAR MODEL OF THE ROBOT JOINT COORDINATES AS FUNCTIONS OF THE CORRESPONDING IMAGE DATA FROM TWO IMAGE SENSORS. ALGORITHMS CURRENTLY USED IN RESEARCH AND COMMERCIAL APPLICATIONS TRANSFORM IMAGE DATA TO OBJECT SPACE. THE ROBOTS KINEMATIC EQUATIONS MUST THEN BE SOLVED TO GET ROBOT JOINT COORDINATES. THE DIRECT TRANSFORMATION FROM IMAGE SPACE TO ROBOT JOINT SPACE WILL ELIMINATE THIS STEP, RESULTING IN REDUCED COMPUTATION TIME. THE SYSTEM IDENTIFICATION ALGORITHM ALLOWS TRUE TURN-KEY OPERATION WHICH ELIMINATES EXPENSIVE CUSTOM ENGINEERING AND PROGRAMMING FOR AN INSTALLATION.