How to progress beliefs in continuous domains
Vaishak Belle, Hector J. Levesque · Edinburgh Research Explorer (University of Edinburgh) · 2014
When Lin and Reiter introduced the progression of basic action theories in thesituation calculus, they were essentially motivated by long-lived roboticagents functioning over thousands of actions. However, their account does notdeal with probabilistic uncertainty about the initial situation nor witheffector or sensor noise, as often needed in robotic applications. In thispaper, we obtain results on how to progress continuous degrees of beliefagainst continuous effector and sensor noise in a semantically correctfashion. Most significantly, and perhaps surprisingly, we identify conditionsunder which our account is not only as efficient as the filtering mechanismscommonly used in robotics, but considerably more general.