This Small Business Innovative Research (SBIR) Phase I project brings a unified, information-based design to wireless sensor network (WSN) infrastructure architecture. Because WSNs sacrifice reliability over the greater density of the sensors and lower network setup costs compared to traditional sensor networks, these networks are characterized by severe resource constraints and uncertainty regarding power, communication and computation. Such conditions can result in latency and lack of coordination between infrastructure message routing and other decisions that protect and preserve the network and the applications fusion and information seeking actions. The proposed unified architecture enhances a tight coupling between communication and application by influencing spanning tree algorithms to consider efficiency in inference as important as efficiency in communication, to limit information flows to those that enhance the fused information state through the feedback of the application level fusion and eliminate the redundant data at the sensor source, and to decompose, embed, and distribute the application among the processing nodes of the sensor network. The expected end result will enhance autonomous decision making with much reduced latency, making for more responsive, adaptable applications.
The broad impact of this project is to provide software for the backbone infrastructure in high value, high-risk applications of sensor networks. This architecture will enhance the development of sensor networks in remote, even hostile environments as well as impossible to wire and difficult sensing applications. The industrial range of wireless sensor networks is very broad, from military to medicine, and the uniqueness of applications unconstrained by wires is likely to grow as well. Additionally, the impact of this work will provide a better understanding of how the limits of communication resources lead to poorly coordinated and latent decisions and the knowledge of how to better organize communication and analysis within human organizations, especially in the homeland security and intelligence community.
