Distributed Localization Protocol for Routing in a Noisy Wireless Sensor Network

To be efficient, routing protocols in wireless sensor networks suppose to take into account local interactions, noise and collisions. In this work, we propose a routing protocol suitable for noisy environments. Using the qualitative localization algorithm (QLoP) in order to select sensors with a better signal-to-noise ratio, we build a logical topology on which we apply a routing protocol. We show that received signal strength indicator (RSSI) is not a good metric to evaluate proximity. In order to provide a better proximity metric, we introduce the qualitative localization protocol: QLoP. QloP does not use any anchor or dedicated hardware like GPS. Each node builds a qualitative distance according to the 2-hop neighborhood information. Thus, this algorithm allows to determine coarsely the proximity of the neighbors which are classified as very close, close or far. We highlight the routing performances in a noisy environment and compare it to a classical routing protocol on a flat topology. Delivery rate, throughput and average distance reachable in those extremes conditions are appreciably enhanced.