Architecting multi-agent systems with distributed sensor networks

In this paper, we propose a cognitive architecture for an intelligent agent and define an agent model that can be used to conceptualize the design of multi-agent systems. Sensor networks can be considered as dynamic and adaptive systems that can support multi-agent applications. We present a hybrid, multi-layered, agent-oriented architecture for distributed sensor networks. We suggest an agent model that integrates the software agents of the application with the hardware agents of the physical environment. We describe a target tracking experiment as a multi-agent sensor network application, to demonstrate the efficacy of approach. We conclude the paper with an insight to future challenges.