Development of a wake sensing capability within the onboard AI of an Unmanned Aerial Vehicle towards formation flying

It is well known that many species of large birds, especially migrating ones, fly together in a symmetric V-shaped formation. It is most likely to save energy that they adopt such behavior. Through a biomimetic approach, humans applied this observation to military aircrafts and Unmanned Aerial Vehicles (UAV's) to benefit the significant energy savings from formation flights. It is important to mention that for safety reasons, this principle is not applied to civil aviation yet. The most challenging problem in formation flight is to know the follower's position in the leader's wake, this is called "wake sensing". The goal of this master thesis is to develop a wake sensing capability for UAV's using artificial intelligence. The first step will be to modify the source code of JSBSim, an open-source flight simulator, in order to incorporate the physics of the wake generated by the leader drone. Then, multiple wake sensing algorithms will be tested using the NVIDIA Jetson TX2, a powerful microcomputer, optimal for artificial intelligence deployment.