In this project we will present an algorithm whose purpose is causing a load-bearing drone to plan a course from a pre-known starting point to a destination area, while avoiding flying over obstacles. The drone will identify the destination area and the obstacles by itself, and from the moment of takeoff at the starting point up until the drone’s landing the entire process happens automatically.
The drone’s path planning is done using the RRT* algorithm, which is a version of the RRT algorithm – Rapidly exploring Random Trees. The object identification is done using basic image processing techniques. The drone is self-controlling using a P controller.
The duration of the algorithm from takeoff to landing is around 1 minute for a square area with a side of around 2.5 meters.