Multi-Robot Swarm for Search and Rescue in Disaster Zones
Designed and simulated a decentralized multi-robot swarm using MATLAB to autonomously navigate, search, and relay data in dynamic disaster zones. The project combines Voronoi-based partitioning, repulsive field navigation, and energy-aware planning to create a scalable, robust system suitable for search and rescue missions in uncertain and hazardous environments.
1. Distributed Coordination and Area Coverage
- Simulated a swarm of 20 robots in a 200×200 m environment with 5 dynamically expanding hazard zones.
- Implemented Voronoi partitioning for efficient area allocation and collision-free navigation.
- Enabled autonomous local decision-making using distributed control with no centralized controller.
- Achieved over 95% total area coverage within 250 steps under tight communication and sensing constraints.
2. Hazard Avoidance and Adaptive Behavior
- Modeled disaster zones as growing repulsive fields, with each robot adjusting paths in real time to maintain safety.
- Used local sensors and gradient-based repulsion logic to detect and steer clear of expanding hazards.
- Validated performance through disaster heatmaps and real-time trajectory visualizations.
3. Energy-Aware Motion Planning
- Developed a control algorithm prioritizing coverage and efficiency, with energy depletion modeled per time step.
- Robots retained over 80% of energy by the end of the mission through optimized path planning and task allocation.
- Ensured longer mission lifetimes without recharging, ideal for real-world deployments.
4. Results and Applications
- Demonstrated strong coverage, safe hazard avoidance, and energy conservation in a simulated post-earthquake urban setting.
- Scalable to larger swarms or heterogeneous systems (e.g., drones, ground robots).
- Applications include search and rescue, environmental monitoring, battlefield reconnaissance, and planetary exploration.
Simulation Visualizations
Swarm behavior navigating dynamic hazard zones.