Use of Machine Vision for Feedback Computation in Motion Control Systems

Abstract

Abstract Machine vision systems use cameras to capture images and process image signals to interpret and analyze the received signals by computers or other signal processing devices. An important application of machine vision is to control the position of moving objects or the moving part of an object (e.g., a motor) using cameras (by image processing). Image processing is aimed at determining positions and transmitting relevant data via a processor to a controller (i.e., a digital control system) implemented on a microcontroller to control an actuator’s position. This study utilized the linear Lorentz actuator, as it has a simple structure, small size, and no clearance or output force fluctuation. The operating frequency of a system is also a key parameter because the execution time of a control process is directly associated with the system performance. The operating frequency is the frame rate, which is 60 Hz (60 fps). This study tracked the motion of a linear electromagnetic actuator, determined position and displacement in each frame, and calibrated the real position as the position feedback to implement a position controller for the actuator to eliminate the error signal (i.e., the difference between the real position and reference actuator position). Furthermore, a digital control system designed with two zeros and two poles along with an integrator was employed to eliminate the steady-state error.