Research on Plan Recognition Based on Misleading Action Processing

Abstract

As the application field of plan recognition expands continuously, the traditional plan recognition methods have hindered the application of plan recognition in hostile environment, which ignored the defects of misleading action. In this paper, a plan recognition algorithm based on the process of misleading action is proposed under the framework of plan knowledge graph. The method makes it possible to infer the action that is not observed by introducing the concept of the original reliability into the plan knowledge graph, and makes it possible to process the misleading action and wrong operation of the hostile agent by introducing the recognition process of the misleading action, which the method based on plan knowledge graph cannot achieved. The experimental results showed the algorithm in this paper greatly improved the accuracy of plan recognition by the process of the misleading action and it was with good performance. The algorithm has great significance for the application of plan recognition in the adversarial environment such as the intrusion detection, tactical planning, intelligence games etc.