GRAL: Localization of Floating Wireless Sensors in Pipe Networks

Abstract

Mobile wireless sensors are increasingly recognized as a valuable tool for monitoring critical infrastructures. An important use case is the discovery of leaks and inflows in pipe networks using a swarm of floating sensor nodes. While passively drifting along, the devices must track their individual positions so critical points can later be located. Since pipelines are often situated in inaccessible places, large portions of the network can be shielded from radio and satellite signals, rendering conventional positioning systems ineffective. In this paper, we propose a novel algorithm for assigning location estimates to recorded measurements once the sensor node leaves the inaccessible area and transmits them via a gateway. The solution is range-free and makes use of a priori information about the target pipeline network. We further describe two extended variants of our algorithm which use data of encounters with other sensor nodes to improve accuracy. Finally, we evaluate all variants with respect to various network topologies and different numbers of mobile nodes in a simulation. The results show that our algorithm localizes measurements with an average accuracy between 4.81% and 7.58%, depending on the variability of flow speed and the sparsity of reference points.