Luftschar

UAV Cooperative Swarm Formation

This project proposes a novel approach to resilient navigation for a drone swarm operating in GNSS (Global Navigation Satellite System)-denied environments. The core idea is to leverage Swarm Intelligence where individual drones cooperatively enhance the localization accuracy of the entire group, using IMU sensors / Ultra-wideband radui / mono-camera.

The solution focuses on inter-drone data exchange to establish a shared understanding of relative positions, minimizing reliance on external localization signals.

• Individual Localization: Each drone uses its onboard sensors (e.g., a monocular camera) to independently estimate its position (Visual Odometry/SLAM).

• Cooperative Enhancement: Drones continuously exchange small data packets containing the product of onboard Edge AI processing. This processed data includes information about key visual features or relative pose estimates of neighboring drones.

• Relative Position Awareness: By fusing its own sensor data with the relative information received from neighbors, each drone significantly improves its absolute and relative positional awareness (e.g., cooperative SLAM or filtering, GPS).

1. Low-Bandwidth Communication: The system is designed for low-overhead communication, sharing only feature descriptors or processed data, making it highly scalable and robust against jammed communications.

2. Hardware Focus: While a hardware demonstration is feasible within the hackathon timeframe, the focus will be on delivering detailed system architecture, and robust proof-of-concept simulation/description of the algorithm and data flow. The key deliverable is the information (processed feature data) that UAV-neighbors in swarm would share.