Active target search from UAVs in urban environments

Abstract

In this paper we consider the problem of searching for a target from a camera-equipped unmanned aerial vehicle (UAV) flying in an urban area. Urban areas present challenges because buildings can hamper the ability to see regions on the ground. We describe an algorithm that constructs paths that take into account obstructions due to buildings or other large objects. The approach combines search trees and a particle filters to evaluate a large number of possible paths, while at the same time performing all the Bayes' filter innovations that would need to occur during the evaluation of each path.