Project SKY-NET is an advanced Multi-Agent System designed for Urban Air Mobility (UAM) traffic management. As cities transition toward autonomous air taxis, the challenge shifts from flying a single drone to managing thousands of drones in a dense 3D airspace without a centralized "Air Traffic Control."
Develop a decentralized collision avoidance system that allows individual UAVs to detect, negotiate, and resolve path conflicts in real-time using Velocity Obstacle (VO) geometry and predictive pathing.
- Multi-Agent Deconfliction: Swarm-based logic where drones "talk" to neighbors to resolve head-on and crossing conflicts.
- Dynamic Geofencing: Real-time creation of "No-Fly Zones" around emergencies or tall buildings.
- Path Priority Engine: Giving priority to Emergency Medical Service (EMS) drones over commercial deliveries.
- 3D Air-Corridor Simulation: Modeling urban "Sky-Lanes" above high-density city grids.
- Languages: Python (Coordination Logic), C++ (Fast Vector Math)
- Algorithms: A* Pathfinding, RVO2 (Reciprocal Velocity Obstacles)
- Visualization: Matplotlib / PyGame for 2D/3D Traffic Viz
- Communication: MQTT / DDS (Data Distribution Service) simulation
src/: Deconfliction algorithms and Multi-Agent coordination code.maps/: GeoJSON/JSON city grid layouts and landing pad (Vertiport) data.sim/: Traffic flow simulators and collision-rate analysis tools.docs/: UTM (Unmanned Traffic Management) standard papers and architecture.
- Zero-Collision Rate in simulations with 50+ agents.
- Path Efficiency: Minimize detour time to < 15% of the direct route.
- Latency: Resolve a 3-way conflict in under 50ms.
Developed by Yogesh E S
Aerospace Portfolio - Project #5