Vision Based Cooperative Distributed Collision Avoidance Estimation in a Dynamic Environment for a Heterogeneous UAV Formation

In this paper we present design process of a video based onboard system that needs to detect and avoid static obstacles by the members of a heterogeneous UAV formation. After motivating the importance and relevance of the research topic, we introduce the UAV platforms which will be used with the developed system. A saliency algorithm is used to address the problem object collision avoidance. Furthermore, we expose the demands needed for image stabilization to obtain a working algorithm. This relates directly with the UAV dynamics and with the performances of the gimbal that holds the video sensor. Finally, numerical simulations are performed to evaluate the suitability of the designed system for practical applications.