Estimating density of wireless networks in practice

Density of a wireless network drastically impacts its performance. Adapting the networking protocols at run-time to density changes, which may not be predictable in advance, may improve the network performance. Estimating the density of a wireless network is the challenge we address in this paper. A wireless node may locally estimate the network density by measuring the received signal strength (RSS) of packets sent by its neighbours. However, RSS is prone to large- and small-scale fading, and this phenomenon negatively affects the accuracy of density estimators. In this study, we validate the existing RSS-based density estimators by controlled laboratory experiments conducted in the FIT IoT-LAB test-bed located in Lille, France. Further, we propose a new density estimator that is a fusion of existing estimators. Controlled laboratory experiments showed that the average absolute percentage deviation of the new density estimator is around 1 to 10 percent and the fusion approach overcomes the deficiencies of the existing RSS-based estimators.