Nuisance Parameters and Location Accuracy in Log-Normal Fading Models

The ability to accurately locate devices in wireless communication networks is of growing interest. Many of the position algorithms deployed in such networks are completely, or partially, based on received signal strength (RSS) measurements. However, the role played by nuisance parameters in RSS-based positioning algorithms has been largely overlooked. In this work we focus on a key nuisance parameter of such algorithms, the propagation path loss exponent of the log-normal fading model. We show how this nuisance parameter can have a significant impact on the location accuracy of RSS-based algorithms which assume log-normal fading, and how geometrical distributions of the reference nodes can be found where the error in the location estimate becomes unbounded. The analysis and results reported here help identify algorithms and geometric configurations of the reference nodes where an a priori measurement campaign related to the propagation model would be important