Navigation of a Fuzzy-Controlled Wheeled Robot Through the Combination of Expert Knowledge and Data-Driven...

Area of Science:

• Robotics
• Artificial Intelligence
• Control Systems

Background:

• Autonomous navigation in unknown environments presents significant challenges for wheeled robots.
• Existing methods often struggle with dynamic obstacle avoidance and efficient path planning.

Purpose of the Study:

• To propose a robust and efficient navigation scheme for wheeled robots operating in unknown environments.
• To enhance robot autonomy through integrated behaviors and advanced control strategies.

Main Methods:

• Developed a navigation scheme combining obstacle boundary following (OBF), target seeking (TS), and vertex point seeking (VPS) behaviors.
• Designed a fuzzy controller (FC) for OBF using constrained multiobjective optimization and ant colony optimization.
• Implemented fuzzy proportional-integral-derivative (PID) and proportional-derivative (PD) controllers for TS behavior.
• Introduced a novel behavior supervisor for seamless switching between navigation behaviors.
• Proposed a robot localization method fusing encoder and infrared sensor data with a particle filter.

Main Results:

• The proposed navigation scheme demonstrated feasibility and advantages through simulations and experiments.
• The integrated behaviors and supervisor effectively managed robot navigation in unknown terrains.
• The particle filter-based localization method provided accurate robot positioning.

Conclusions:

• The developed navigation scheme offers a reliable solution for autonomous wheeled robot navigation in complex, unknown environments.
• The combination of advanced control techniques and sensor fusion enhances navigation performance and robustness.