Sparse particle filtering for modeling space-time dynamics in distributed sensor networks

Wireless sensor network has emerged as a key technology for monitoring space-time dynamics of complex systems, e.g., environmental sensor network, battlefield surveillance network, and body area sensor network. As a result, distributed sensing gives rise to spatially-temporally big data. Realizing the full potentials of distributed sensing calls upon the development of space-time modeling of measured signals in the dynamically-evolving physical environment. However, space-time interactions bring substantial complexity in the scope of the modeling, due to the need to investigate spatial correlation, temporal correlation, as well as how space and time interact. Most of previous research considers either spatially-varying time series model or temporally-varying spatial model. This paper presents a new approach of sparse particle filtering to model spatiotemporal dynamics of big data in distributed sensor network. Notably, we developed a compact kernel-weighted regression model of spatial patterns. Further, the parameters of spatial model are transformed into a reduced-dimension space, and thereby sequentially updated with the recursive Bayesian estimation when new sensor observations are available. As such, spatial and temporal processes closely interact with each other. Experimental results on real-world data from wearable ECG sensor network showed that the proposed methodology outperforms traditional methods and effectively models space-time dynamics in distributed sensor networks.