The detection and localization of an event through a sensor network is a topic that has attracted considerable attention recently because of many potential applications. Typically, these decisions are taken by conveying the sensor measurements to a sink node that processes the data and provides an estimate. However, the presence of a sink node creates a bottleneck that is the cause of potential congestions and it poses problems of scalability. In this work, we propose a decentralized decision scheme that is capable to achieve optimal decisions without requiring a fusion center. The network is composed of a set of mutually coupled oscillators, where each node is coupled only to the nearest nodes. We show how to achieve optimal detection for both deterministic and random signals by properly selecting the parameters of the coupling mechanism. Furthermore, if the nodes know their own positions and the network is connected, we show how to make each node able to perform a totally distributed energy-based source localization. © 2006 IEEE.
Decentralized detection and localization through sensor networks designed as a population of self-synchronizing oscillators / PESCOSOLIDO, Loreto; BARBAROSSA, Sergio; SCUTARI, GESUALDO. - 4:(2006). (Intervento presentato al convegno 2006 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2006 tenutosi a Toulouse; France nel May 14 – 19, 2006) [10.1109/ICASSP.2006.1661135].
Decentralized detection and localization through sensor networks designed as a population of self-synchronizing oscillators
PESCOSOLIDO, Loreto;BARBAROSSA, Sergio;SCUTARI, GESUALDO
2006
Abstract
The detection and localization of an event through a sensor network is a topic that has attracted considerable attention recently because of many potential applications. Typically, these decisions are taken by conveying the sensor measurements to a sink node that processes the data and provides an estimate. However, the presence of a sink node creates a bottleneck that is the cause of potential congestions and it poses problems of scalability. In this work, we propose a decentralized decision scheme that is capable to achieve optimal decisions without requiring a fusion center. The network is composed of a set of mutually coupled oscillators, where each node is coupled only to the nearest nodes. We show how to achieve optimal detection for both deterministic and random signals by properly selecting the parameters of the coupling mechanism. Furthermore, if the nodes know their own positions and the network is connected, we show how to make each node able to perform a totally distributed energy-based source localization. © 2006 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
