With the development of miniaturized sensor and communication technologies, there is an ever-growing demand for algorithms to derive high-level spatiotemporal events from large amounts of sensed data. Our previous work has already defined a decentralized, in-network approach to identifying topological relation changes between continuously evolving regions monitored by a geosensor network. However, our previous work relies on a number of strong continuity assumptions, concerning the temporal granularity of sensor observations and the type of region deformations. This paper presents an improved algorithm which demonstrates how these key simplifying assumptions can be relaxed. Empirical testing of the algorithm demonstrates how this algorithm can operate at higher levels of scalability than both the previous algorithm, and smart centralized alternatives.
↧