Study of Inverted Pendulum with Pneumatic Servo Control by Image Track Method

• Main Authors: Che-Ming, Kuo, 郭哲銘
• Other Authors: Ching-Kuo, Lin
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/37823881687987863729

碩士 === 華梵大學 === 機電工程研究所 === 93 === The aim of this paper is to develop an image system as the feed back of inverted pendulum with pneumatic control. At present, the technique of image is widely applied to different purposes. Image is often utilized as a feed back basic in a great number of non-contact detects. This is because image facilities are cheaper, highly replaceable as well as expansible, and the accuracy of detect is adjustable freely. The image taken with a CCD (Charge Coupled Device) video camera is 2-D. There exist inherent drawbacks in dynamic image system while it is used in detection, such as the number of image is limited to 30 per second, image retention emerges while a moving object is taken, and image distortion caused by lens. In this study, a CCD video camera with a manual shutter was used to reduce image retention of moving objects. LSE (Least Squares Estimation) was used to correct distortion, and it was combined with bilinear interpolation for distortion restoration. This paper proposed a pattern recognition filter to quickly identify the figure and central coordinate position of objects, as well as a module classification to classify objects quickly, which, in combination with threshold, could separate objects from noise quickly. The concept of dynamics engineering mechanics was used to design an image estimator and a fuzzy controller to control pneumatic platform, and carry out an experiment on the swing upward and balance of inverted pendulum.