Automated Indoor Drone Flight with Collision Prevention

This paper details a system that provides positioning, obstacle detection and avoidance, pathfinding, and energy monitoring suitable for indoor operation designed for the Crazyflie 2.0 drone. The positioning subsystem was achieved by a hybrid of Received Signal Strength Indicators and Dead Reckoning. Obstacle detection and avoidance used an IR sensor to detect objects in the drone’s path, allowing the drone to look for alternate paths. Pathfinding used a modified Node Array A* algorithm implemented in a 3D model of the testing area within the Unity engine. Finally, energy monitoring used the drone’s built-in Python library to log voltage values and were sent to the Unity system, which reroutes the drone upon detecting a low battery voltage. The system was able to provide a basic autonomous navigation system that prioritizes safety of the drone and its flying environment.