Deterministic versus Probabilistic Methods for Searching for an Evasive Target

Authors

  • Sara Bernardini Royal Holloway University of London
  • Maria Fox King's College London
  • Derek Long King's College London
  • Chiara Piacentini University of Toronto

DOI:

https://doi.org/10.1609/aaai.v31i1.11054

Keywords:

UAVs, planning, CP, probabilistic reasoning, POMDPs

Abstract

Several advanced applications of autonomous aerial vehicles in civilian and military contexts involve a searching agent with imperfect sensors that seeks to locate a mobile target in a given region. Effectively managing uncertainty is key to solving the related search problem, which is why all methods devised so far hinge on a probabilistic formulation of the problem and solve it through branch-and-bound algorithms, Bayesian filtering or POMDP solvers. In this paper, we consider a class of hard search tasks involving a target that exhibits an intentional evasive behaviour and moves over a large geographical area, i.e., a target that is particularly difficult to track down and uncertain to locate. We show that, even for such a complex problem, it is advantageous to compile its probabilistic structure into a deterministic model and use standard deterministic solvers to find solutions. In particular, we formulate the search problem for our uncooperative target both as a deterministic automated planning task and as a constraint programming task and show that in both cases our solution outperforms POMDPs methods.

Downloads

Published

2017-02-12

How to Cite

Bernardini, S., Fox, M., Long, D., & Piacentini, C. (2017). Deterministic versus Probabilistic Methods for Searching for an Evasive Target. Proceedings of the AAAI Conference on Artificial Intelligence, 31(1). https://doi.org/10.1609/aaai.v31i1.11054

Issue

Section

AAAI Technical Track: Reasoning under Uncertainty