Sparse particle filtering for modeling space-time dynamics in stochastic sensor network

Wireless sensor network has emerged as a key technology for monitoring space-time dynamics of complex systems, e.g., wearable electrocardiogram (ECG) sensor network. However, traditional ECG sensor networks demand reliable sensor readings through sensor-skin contacts, which greatly deteriorate the system comfortability and wearability. In order to realize a highly wearable sensing system, we propose a new strategy of stochastic sensor networks that relax the hardware constraints and allow a random subset of sensors to transmit information intermittently at dynamically varying locations within the network of sensors. Nonetheless, stochastic sensor network gives rise to spatially-temporally big data. Space-time interactions bring substantial complexity in the scope of the modeling. Hence, stochastic sensor network calls upon the development of novel analytical methods and tools to support the design and harness the uncertain information for decision making. This paper presents a new approach of sparse particle filtering to model spatiotemporal dynamics of big data in the distributed and stochastic sensor network. First, we utilized the parametric kernel-weighted regression model to represent spatial patterns. Further, the parameters of spatial model are transformed into a reduced-dimension space, and then sequentially updated with the recursive Bayesian estimation when new sensor observations are available over time. As such, spatial and temporal processes closely interact with each other. Experimental results on real-world data and different scenarios of stochastic sensor networks demonstrated the effectiveness of sparse particle filtering to support the stochastic design and harness the uncertain information for modeling space-time dynamics of complex systems.