Skip to Main Content
As the ubiquitous society matures, mobile phones are equipped with agents, which support users in selecting the appropriate course of actions, for example, by giving guidance on where to change trains. However, since human society becomes more and more complex and sophisticated, agents capable of understanding the situation of each user (action plan, preferences, location, etc.) and providing support accordingly are still at the research stage. This paper defines the normal action model from the user's action history, proposes the configuration of an agent that can propose an alternative action and lead the user to take that action when the user encounters a situation not conceived in the normal action plan. In addition, the TupleSpace is applied to inter-agent communication. It has been confirmed using an evaluation system that tuple-based communication can be used to communicate information about an unexpected event.