| With the ongoing, rapid advances in sensor and communication technologies, it is expected that in the near future an enormous number of heterogeneous sensors, performing observations and detecting events in real-time, will be established. We envision a world-wide geographical information infrastructure for sensor networks. This infrastructure will allow: (1) sensors to publish their geographically referenced sensing resources; (2) users to find sensing resources in their areas of interest; and (3) users to access, understand and use these geographically-referenced sensing resources. We call this geographical information infrastructure for sensor networks the Spatial Sensor Web (SSW).; In order to build an SSW, there are two major challenges: (1) Interoperability, a challenge that stems from the heterogeneity of current sensor networks; and (2) Scalability, a challenge that stems from the enormous number of sensor networks and users participating in the SSW. Hence, understanding, designing and implementing such an interoperable and scalable SSW is the focus of this Ph.D. thesis.; As a proof of concept, part of this thesis focuses on developing an interoperable and scalable distributed GIService architecture, called GeoSWIFT, for the Sensor Web. GeoSWIFT is a software platform that, through open standards-based sensor-focused GIService interfaces, allows users/applications/sensors to, as a single unit, access and interchange Internet-scale heterogeneous sensing resources according to spatial queries.; Our GeoSWIFT design is interoperable in that it uses open standards-based Sensor Web GIService interfaces and information models. Through implementation and real-world deployment, we have proved that (1) GeoSWIFT can enable heterogeneous sensor networks, both in-situ and remote, to become interoperable sensing services, and (2) that new applications can be built by dynamically assembling interoperable sensing resources.; Our GeoSWIFT design is also scalable, as evidenced by the fact that we have used a fully decentralized Peer-to-Peer architecture to build GeoSWIFT. The architecture is scalable because: (1) it has no single point of failure; (2) its spatial query performs effectively even when the number of sensors increases; (3) system loads are distributed across the peer nodes; and (4) system control cannot be taken over by a single organization. |
