A microcontroller based motion control system for robotic eyes

Abstract

Robot with the vision capability is an interesting research topic which has many practical applications. The goal of this project is to build a robotic eyes system based on the motion control system. The robotic eyes mimic the human eyes to follow the moving object. It could be applied into the surveillance systems or humanoid robots. It also might be implemented into other system which could help human to do some difficult tasks such as exploring the ocean or space, capturing the battle. This system consists of two webcams attached to four DC motors and a microcontroller system which is connected to a computer. The images are captured by the webcams and processed by the image-processing program in the computer. This program detects the object and calculates how large the angles which the motors need to rotate. Then the data is sent to the microcontroller. The microcontroller will receive the data and to run the motors to the desired positions so that the webcams can keep track of the movements of the object like our eyes. This project focuses on the motion control part to rotate the robotic eyes precisely. One driven and decoder circuit was built successfully to control four motors at the same time. Proportional-Integral (PI) controller and Model Predictive Controller (MPC) were implemented to achieve a closed loop control. The performance and comparison between two controller were described in this thesis. In the result, the robotic eyes system created is able to detect the object and trace it in two dimenstions with considerable accuracy. This thesis covers the background, hardware and software implementation, as well as testing results and discussions of the project. Finally, some future practical implementations of the system are also discussed.