Charles Explorer logo
🇬🇧

Centralized Bayesian reliability modelling with sensor networks

Publication at Faculty of Mathematics and Physics |
2013

Abstract

The article concerns reliability estimation in modern dynamic systems. It introduces a novel approach, exploiting a network of several independent spatially distributed sensors, actively probing the monitored system.

A dedicated network element - the fusion centre - is then responsible for processing the information provided by sensors and evaluation of final reliability estimate. On the base of computational abilities of sensors, we propose two conceptually different reliability estimation scenarios: (1) the computationally cheaper dummy sensors scenario, in which the sensors send raw data to the fusion centre; and (2) the smart sensors scenario, when the data are processed locally by sensors, and the fusion centre subsequently merges their resulting information.

Bayesian paradigm was adopted for consistent information representation, its adaptive dynamic processing and fusion.