A New Position Feedback Method for Manufacturing Equipment

Abstract

This paper presents an innovative real time 2-dimensional position feedback method, which processes visual input data from a target image on an actively-controlled planar pixel matrix. The objective is to demonstrate the ability to position an X-Y stage with high resolution, using direct position sensing of a dynamically controlled image. In order to achieve high spatial resolution using a pixel array as a target, an algorithm that processes both the geometric shape and the grayscale intensities of the image is implemented. The test platform consists of an X-Y stage carrying a Liquid Crystal Display (LCD) screen that is imaged by a stationary digital camera. The pixel intensities on the LCD screen are modified dynamically to provide 2-dimensional position command inputs that translate to the desired stage motion. The digital images acquired by the camera are used to provide position error feedback to the controller. Experimental results show that direct position sensing is possible to a certain degree of accuracy. However, in order to match today’s CNC machines’ accuracy levels further processing of the digital images is required. A noise reduction algorithm to filter the fluctuations of the measurements in the digital images is proposed as future work, as well as other considerations.