Simulating BDI-Based Wireless Sensor Networks

``Autonomic systems´´ merge advancements in the field of multi-agent software design, dynamic analysis, and decentralized control in order to assist designers in constructing complex distributed systems. Wireless Sensor Networks (WSN) represent such systems, and may benefit from autonomic system designs that target distributed nodes in diverse and changing environments that interact over a wireless communication channel for decentralized problem solving. Multi-agent system techniques have been recently applied to WSN´s; however, due to hardware limitations nodes (agents) are not fully deliberative (or strong) reasoning systems. Since hardware increases rapidly it is expected that such systems may eventually be viable. In this paper we provide a generic, extensible, and deliberative simulator for testing interactions in autonomous WSN´s. The belief, desire, intention (BDI) agent model of Rao is used, as well as the Agentspeak language, and the Jason framework. Results from two simple WSN test scenarios show how (simulated) BDI agents might perform basic WSN functions.