Ball and Plate System

A Ball & Plate System using Noise-immune PD Control

Introduction

This project's goal is to build a 2-DOF ball and plate system, which can balance a ping-pong ball on a flat plate.

To detect the position of the ball, We use a low-cost USB camera as the image sensor, and process the video in real time with OpenCV-Python on the PC.The coordinate is then transmitted to STM32 microcontroller where a noise-immune PD control loop is implemented, actuating 2 servos to control the pitch and roll of the plate to keep the ball in its centre.To reduce the oscillation caused by large noises, a noise-robust differentiator[1] is used to improved the PD control algorithm. Experiment results show that the system using this method has better robustness and rapidness.

Software Environment

• python 3.6 

• numpy 1.14.0

• opencv-python 3.4.0

Hardware

• a USB camera

• 2X analog servos

• STM32F103ZET6 microcontroller

Other Materials

• an acrylic plate

• several plastic sticks

• 3X universal joints

• 2X metal sticks with rod ends

• others

Reference

[1] http://www.holoborodko.com/pavel/numerical-methods/numerical-derivative/smooth-low-noise-differentiators