Decentralized Coordination of Intelligent Robot Swarms

This paper presents a decentralized coordination algorithm for intelligent robot swarms that integrates behavior-based control, fuzzy logic for dynamic behavior weighting, and a virtual leader mechanism for scalable formation guidance. The algorithm enables each robot to independently combine three core behaviors—formation keeping, obstacle avoidance, and goal tracking—using weights derived from fuzzy inference based on real-time sensor data. A mathematical formulation of the behavior composition model and velocity control is provided. The system was validated in a Python-based simulation with 16 agents navigating around obstacles while preserving formation. Results demonstrate smooth, collision-free movement with less than 0.5 meters of average deviation from formation targets. Compared to existing methods, the proposed approach offers improved resilience, adaptability, and real-time performance without centralized control.