A study on the implementation of a ball and plate system using lw-rcp and machine vision based on odroid

초록

In this paper, we propose the architecture of a ball-and-plate system as an educational platform that can help students comprehensively study control engineering, robotics, machine vision, and embedded systems. In addition, we describe how to implement the sub-systems and present an example of an education plan using the proposed system. In the proposed system, we use an LW-RCP (Light Weight - Rapid Control Prototyping) system, which is lab-built, to help students implement the control algorithm. An RCP system makes it easy for students who are not familiar with u-Controller or the C language to implement controllers. We also use Odroid, a low-cost SBC (single-board computer), and a single camera to calculate the position of the ball using a machine vision algorithm. In this structure, the measurement subsystem is more durable compared with contact-based sensing systems and students can test their own machine vision algorithms for better measurements. We adopt the Stewart platform, a type of parallel robot, to generate roll-and-pitch actuation for the ball-and-plate system. By adopting the Stewart platform, we can handle more complicated control problems for movement that requires more than two degrees of freedom. Additionally, we implement an LQ controller, which is a typical model-based controller, and apply it to the ball-and-plate system. Finally, through experiments, we illustrate the usefulness of the proposed system. © ICROS 2020.

키워드