Fast radioactive hotspot localization using a UAV

We address the problem of hotspot seeking in an unknown radiation field using an Unmanned Aerial Vehicle (UAV) with limited resources. For on-the-spot investigation of accidental radiation releases, without a priori knowledge on the whereabouts of the source of radiation substances leakages, it is very difficult to navigate and return a UAV for fast hotspot localization. We propose a novel Hexagonal Tree (HexTree) based sampling algorithm to find such an optimal tour path based on the appropriate measurement locations. We make a realistic assumption on the environment, theoretically analyze the optimality of proposed algorithm, and numerically compare the performance with the existing method. The proposed algorithm gives faster convergence to the hotspot, an optimal exploration termination condition, and more informative locations while returning to the initial position than conventional random sampling based exploration and path smoothing algorithms.