Observability and independence in distributed sensing and its application

Abstract

This paper addresses the observability and independence of a system monitored using networked distributed sensors under the assumption that some sensing results are indiscriminable. We theoretically analyze the independence of sensing templates, that is, whether sensing systems are not redundant, and the observability of such systems, that is, whether we can determine the state of a system from the sensing results. Application examples to which the results given in this paper can be applied are: (1) end-to-end network quality monitored by end users in which the route actually used is not known and (2) object detection/non-detection by binary sensors randomly deployed without positioning mechanisms.