Detection and Tracking of Phenomena Clouds using Service Oriented Sensor Networks

In this paper, we propose distributed algorithms for the detection and tracking of phenomena clouds such as gas leaks and oil spills, using service oriented sensor networks (SOSNs). A SOSN represents each of its sensors as a service in a service-oriented architecture (SOA), that allows their dynamic discovery and composition into applications. SOSNs emerged due to a need for sensor networks which can be easily programmed and customized and are primarily deployed in smart environments such as homes, warehouses and automobiles. As opposed to the fully distributed nature of wireless sensor networks (WSNs), SOSNs consist of both centralized and distributed components. Currently, the two main approaches to phenomenon detection and tracking are based on either centralized processing of sensor streams or distributed in-network processing using WSNs with ad-hoc networking. Neither of these approaches works well in the case of SOSNs due to high performance requirements and scalability issues of centralized algorithms, and the fact that SOSNs typically do not use ad-hoc networking due to a lack of widely adopted standard protocols. The algorithms described in this paper utilize both the centralized and distributed aspects of SOSNs to perform real-time detection and tracking of phenomenon clouds while minimizing the cost and occurrence of networking and processing bottlenecks. As validation of our approach, we show through simulation experiments, that our proposed algorithms result in significantly lower resource usage and average power consumption per node as compared to other contemporary systems.