Implementation of Fuzzy Control Algorithm in Two-Wheeled Differential Drive Platform
Chen Guoyi
NUS Control and Simulation Laboratory, Singapore 117583
Abstract
Designing and developing Artificial Intelligence controllers on separately dedicated chips have many advantages. This report reviews the development of a real-time fuzzy logic controller for optimizing locomotion control of a two-wheeled differential drive platform using an Arduino Uno board. Based on the Raspberry Pi board, fuzzy sets are used to optimize color recognition, enabling the color sensor to correctly recognize color at long distances, across a wide range of light intensity, and with high fault tolerance.
Keywords: fuzzy logic; fuzzy inference system; fuzzification; defuzzification; PID algorithm.
Designed and assembled by Team ArduiNUS
Due to the superiority of our ALEX in algorithm and hardware design,
this robot is preserved in the Digital Systems & Applications laboratory for future engineers to study.
Introduction
This First Year Project (FYP) Report encapsulates my comprehensive academic journey throughout my first year at the National University of Singapore, aiming to illustrate a careful integration of concepts from modules ME3243/EE3305 Robotics System Design, EE4305 Fuzzy/Neural Systems for Intelligent Robotics, and CG2111A Engineering Principles and Practice. Through the practical application of these modules on the Arduino Uno & Raspberry Pi (RPi) robotic platform (Referred to as Robot ALEX in module CG2111A), I have engineered a rescue robot governed by fuzzy control algorithms.
In Module CG2111A, the rescue site is conceptualized as a Maze, where both the rescue robot and human operator do not have prior knowledge of the Maze’s terrain. They can only identify the terrain through an onboard LiDAR system. The victim is represented by a green or red cylindrical object that can appear anywhere in the Maze. Since there are cylinders of other colors present in the Maze, the robot needs to determine the color of the cylinders to confirm whether they represent victims.
The fuzzy control algorithm optimizes the robot’s locomotion and ability to identify victims. Chapters 1 – 3 focus on locomotion, and Chapter 4 analyzes color recognition. Chapter 1 uses kinematic models and constraints to analyze the locomotion of the two-wheel differential drive platform, represented by the ALEX robot configuration. Chapter 2 employs the Proportional–Integral–Derivative (PID) control algorithm by using the equation of motion derived in Chapter 1 to optimize the ALEX robot’s movement. Chapter 3 implements fuzzy control algorithms to refine the control accuracy of the PID algorithm. When the robot approaches a victim, it will utilize an optical color sensor to assess the victim’s condition. Chapter 4 illustrates the optimization of the color-recognition process using a fuzzy algorithm. The acknowledgment, Guoyi’s First Adventure in Robotics, is the final part of this paper. It will narrate my learning journey in my Year 1 at the National University of Singapore.
