Auto Swarm

The boys

Inspiration Traditional drone delivery systems are limited by central control, GPS dependency, and poor adaptability in unpredictable environments. I wanted to design a resilient, scalable solution that mimics how nature solves complexity — with autonomous swarms.

What it does Auto Swarm is a decentralized drone delivery system where each drone communicates with others via a mesh network. The swarm self-organizes, adapts to obstacles, and can deliver payloads without GPS, enabling operations in GPS-denied zones like forests, warzones, or disaster areas.

How I built it Used ESP32-based boards with mesh communication to create a lightweight mesh network. Each drone is equipped with a Pixhawk flight controller, VL53L0X distance sensor, an MPU6050 for orientation, and custom algorithms for swarm behavior. I incrementally tested flight logic, inter-drone communication, and real-time coordination.

Challenges I ran into

Achieving reliable mesh communication in dynamic environments
Stable Connection to telemetry module
Fine-tuning motor control and PID stability without external positioning
Power management and ensuring consistent sensor feedback mid-flight
Debugging swarm logic in real-time was tricky without a centralized system

Accomplishments that I proud of _ Built two functional units

Built a fully functional swarm communication protocol
Achieved a stable test flight on a single drone

What I learned

Swarm intelligence requires simplicity in individual agents but robust inter-agent logic
Real-time sensor fusion is critical to maintaining flight stability

What's next for Auto Swarm