A bio-inspired swarm UAV framework integrating thermal sensing and optimization-based coordination for efficient...

Area of Science:

• Robotics and Autonomous Systems
• Artificial Intelligence
• Search and Rescue Technology

Background:

• Search and rescue (SAR) operations face challenges in disaster zones, including time constraints, difficult terrain, and responder safety.
• Unmanned Aerial Vehicles (UAVs) with thermal imaging offer potential for aerial SAR, but current methods struggle with efficient area coverage and redundancy.
• Coordinated UAV swarms are needed to overcome limitations of single-drone approaches in complex SAR environments.

Purpose of the Study:

• To propose and evaluate a modular, bio-inspired swarm UAV framework for real-time thermal-based SAR.
• To enable autonomous, cooperative exploration of disaster areas using a shared thermal confidence map.
• To introduce and validate a novel Exploration Score metric for assessing swarm search efficiency.

Main Methods:

• Development of a bio-inspired swarm UAV framework with intelligent agents using optimization algorithms.
• Implementation in a high-fidelity PX4 + Gazebo simulation for real-time thermal detection and multi-drone coordination.
• Evaluation of ten bio-inspired algorithms, including Particle Swarm Optimization (PSO), Grey Wolf Optimizer, and Ant Colony Optimization, using the Exploration Score metric.

Main Results:

• Particle Swarm Optimization (PSO) achieved the highest Exploration Score (0.67), outperforming other tested algorithms.
• The PSO-enabled swarm covered 80% of the search area within 60% of the mission time with low redundancy (0.25).
• The swarm maintained balanced inter-drone separation (15-20m), demonstrating adaptive and energy-efficient search capabilities.

Conclusions:

• The proposed bio-inspired swarm UAV framework effectively enhances the speed and efficiency of thermal-based SAR missions.
• The Exploration Score metric provides a unified measure for comparing UAV swarm coordination strategies in SAR.
• This research offers a valuable benchmarking platform for future advancements in autonomous SAR systems.