Decision-Theoretic Planning for Autonomous Robotic Surveillance

In this paper, we introduce a decision-theoretic strategy for surveillance as a first step towards automating the planning of the movement of an autonomous surveillance robot. In our opinion, this particular application is interesting in its own right, but it also provides a test-case for formalisms aimed at dealing both with (low-level) sensor, localisation, and navigation uncertainty and with uncertainty at a more abstract planning level. After a brief discussion of our view on surveillance, we describe a very simple formal model of an environment in which the surveillance task has to be performed. We use this model to illustrate our decision-theoretic strategy and to compare this strategy with other proposed strategies. We treat several simple examples and obtain some general results.