Direct State-Space Models for Time-Varying Sensor Networks

T. J. MATARAZZO, S. N. PAKZAD

Abstract


This paper discusses data from sensor networks with time-variant configurations called dynamic sensor network (DSN) data, which have a higher capacity for storing spatial information than fixed sensor data. DSN datasets are applicable to highresolution mobile sensor networks and the processing of BIGDATA. The defining attribute of these data matrices is the presence of spatial discontinuities, which pose a modeling challenge. An indirect state-space model with user-selected states is presented to account for data matrices containing spatial discontinuities. With this approach, measurements from a large number of sensor nodes can be incorporated in a model with a relatively small size. General characteristics of DSN data are provided and a BIGDATA processing example is examined.

doi: 10.12783/SHM2015/379


Full Text:

 Subscribers Only