Analysis of path loss exponent error in ranging and localization of wireless sensor network

Abstract

This paper presents the results of study and analysis on the effects and influences caused by inaccurately measured environmental variables for ranging and localization purposes. These variables include path loss exponent, lognormal noise, and received signal strength measured at reference distance. Three cases of variable settings are included in the analysis for comparison: the actual, underestimated, and over-estimated values. The analysis involves two radio wave propagation models: the basic log-distance path loss model, and its extension log-normal path loss model. Log-distance path loss model is used in simulation, and log-normal path loss model is used in real-case experiment. Excluding the influence of log-normal shadowing and various fading effects, simulation reveals the enlargement and decrement of estimated length / space scale. Real-case experiment reveals the strengthening and weakening effects of log-normal variation due to under-estimated and overestimated path loss exponent. Concluding simulation and real-case experiment, it shows that over-estimated path loss exponent reduces the effects of log-normal noise but it also reduces the maximum reachable range in estimation. If choices are allowed between over-estimated and underestimated path loss exponent, the first choice is preferred.