Efficient partitioning of space for multiple UAS search in an unobstructed environment

Abstract

Abstract The advent of unmanned aerial systems (UAS) has created opportunities to replace expensive capital assets with these small, yet capable, platforms. Tasks that are identified as able to be performed by UAS may also benefit from the ability of a collection of UAS to operate in a cooperative and parallel manner. Parallelization means that several parts of the search area can be covered at the same time, reducing the overall task completion time. In this paper we investigate how to divide the search and rescue task in a way that it can be performed by a set of UAS. Our investigation covers the detection of multiple mobile objects by a collection of UAS. Three methods (two that are not informed by object location probabilities and one that is) for dividing the space are proposed, and their relative strengths and weaknesses investigated. A reduced aerial camera model facilitates the simulation of object detection. The topic is approached holistically to account for contingencies such as airspace deconfliction. Results are produced using simulation to verify the capability of the proposed method and to compare the various partitioning methods. Results from this simulation show that great gains in search efficiency can be made when the search space is partitioned using a method based on object location probability. For search areas of 3000 m

$$\times $$

×

3000 m detection probability gains of 2

$$\times $$

×

are achievable. For larger areas these gains can exceed 7

$$\times $$

×

.