IEICE Information and Communication Technology Forum
Localizing Sensors from their Responses to Targets
PDF download (1.1MB)
The spatial relations between sensors placed for target detection can be inferred from the responses of individual sensors to the target object even if the location of object is unknown. In this talk, I show an algorithm for estimating locations of sensors based only on their responses to target objects. In the algorithm, two sensors are assumed to be connected when they simultaneously detect an object. Thus, new pieces of proximity (connectivity) information are obtained whenever an object passes over the area in which the sensors are deployed. This proximity information can be aggregated and transformed into a two dimensional map of sensors by exploiting an algorithm of the connectivity-based sensor localization. This algorithm works even if the locations of the sensors are unknown at first. A map of the locations of the sensors is gradually built up as a number of objects pass over the area in which the sensors are deployed. In the talk, I show that this simple algorithm can yield surprisingly accurate estimates of the locations of sensors.