MAV-swarms: Unmanned aerial vehicles stabilized along a given path using onboard relative localization

A novel approach for stabilization and navigation of swarms of Micro Aerial Vehicles (MAVs) along a predefined path through a complex environment with obstacles is introduced in this paper. The method enables to control large MAV swarms (in literature also called UAV swarms) based only on onboard sensors and without any inter-vehicle communication. The proposed method relies on visual localization modules carried by all MAVs, which provide estimation of the relative positions of neighbours in the swarm. Guess on the positions of the neighbouring MAVs and information on the relative positions of obstacles are integrated into swarm stabilization via Reynolds' Boids model. The performance of the complex system is shown in various numerical simulations and in experiments with a fleet of MAVs in the paper. Presented experimental results with the multi-MAV swarm were conducted in indoor and outdoor environment, and without using any external global localization system such as Vicon motion capture system or GPS localization.