An AI Planning-Based Approach to the Multi-Agent Plan Recognition Problem.

Multi-Agent Plan Recognition (MAPR) is the problem of inferring the goals and plans of multiple agents given a set of observations. While previous MAPR approaches have largely focused on recognizing team structures and behaviors, given perfect and complete observations, in this paper, we address potentially unreliable observations and temporal actions. We propose a multi-step compilation technique that enables the use of AI planning for the computation of the probability distributions of plans and goals, given observations. We present results of an experimental evaluation on a novel set of benchmarks, using several temporal and diverse planners.