Parallel tracking and mapping with multiple cameras on an unmanned aerial vehicle

This paper presents the development and testing of a new multi-camera system designed for tracking and mapping aboard a small unmanned aerial vehicle. In order to be resistant to failure from loss of tracking, the large field of view cameras are positioned in such a way as to cover a large portion of the vehicle's surroundings. The proposed algorithm is tested indoors, and it is found that the system achieves good performance. More importantly, it is shown that the system is resistant to failure by comparing it to a system with fewer cameras that loses tracking during certain portions of the test.