Work

Fuzzy Logic-Based Navigation for Mobile Robots

Fuzzy Logic
Mobile Robotics
MATLAB
Simulink
Navigation

Designed and simulated a fuzzy logic-based controller for the navigation of a differential drive mobile robot. Implemented the model in MATLAB/Simulink to optimize robot movement and decision-making.

Fuzzy Logic Robot Navigation

GitHub Repository

Overview

This project explores the implementation of a fuzzy logic-based navigation system for a mobile robot using a differential drive mechanism. The objective was to design a controller capable of guiding the robot from a starting position to a target location while avoiding obstacles and optimizing movement efficiency.

Technologies Used
  • Programming & Simulation: MATLAB, Simulink
  • Control Technique: Fuzzy Logic Controller
  • Robot Model: Differential Drive Robot
  • Input Variables:
    • Error in X position
    • Error in Y position
    • Error in orientation (Theta)
  • Output Variables:
    • Speed of the left wheel
    • Speed of the right wheel
Implementation Details
  1. Development of Fuzzy Logic Controller
    • Designed a fuzzy inference system using the Mamdani model.
    • Defined fuzzy sets for input and output variables.
    • Established membership functions for distance errors and angular errors.
    • Created rule-based decision-making for movement control.

  1. Kinematic Modeling of the Robot

    • Implemented the kinematic model for a differential drive robot.
    • Simulated real-time movement of the robot based on velocity inputs.
    • Integrated odometry-based localization.

  2. Simulation in MATLAB/Simulink

    • Created a simulation environment to test different navigation scenarios.
    • Visualized robot trajectories, angular adjustments, and velocity changes.
    • Conducted multiple test cases to fine-tune fuzzy rules.
Results & Findings
  • Successfully achieved smooth and adaptive navigation using fuzzy logic.
  • The robot effectively corrected its trajectory in response to positional errors.
  • The fuzzy control system showed improved stability in comparison to conventional PID controllers.
  • Achieved realistic trajectory tracking under simulated conditions.
Future Improvements
  • Integration with Machine Learning: Implementing neuro-fuzzy learning to enhance adaptation.
  • Hardware Implementation: Deploying the controller on a real mobile robot platform.
  • Obstacle Avoidance: Enhancing the system with ultrasonic or LiDAR sensors.
  • Cloud-Based Monitoring: Logging navigation data for remote analysis.

Contributors

  • Imad-Eddine NACIRI
  • Achraf Berriane
  • Errouji Oussama