Multi-Agent Oriented Tactical Plan Recognition Method with Uncertainty

Abstract

Plan recognition can be considered as a problem of dynamic classification of multiple hypotheses. It is studied in Artificial Intelligence in a variety of research domains. In this paper, the plan recognition theory is first simply introduced. A plan recognition model based on multi-agent systems is set up and then a plan recognition algorithm based on template matching is proposed. Finally, an instance of tactical plan recognition is analyzed and some predicates and operators for expressing the multi-agent tasks are introduced, which shows the feasibility of using the model to solve the problem of multi-agent plan recognition.