Analyzing localization errors in one-dimensional sensor networks

Abstract

One-dimensional sensor networks can be found in many fields and demand node location information for various applications. Developing localization algorithms in one-dimensional sensor networks is trivial, due to the fact that existing localization algorithms developed for two- and three-dimensional sensor networks are applicable; nevertheless, analyzing the corresponding localization errors is non-trivial at all, because it is helpful to improving localization accuracy and designing sensor network applications. This paper deals with localization errors in distance-based multi-hop localization procedures of one-dimensional sensor networks through the CramérRao lower bound (CRLB). We analyze the fundamental behaviors of localization errors and show that the localization error for a sensor is locally determined by network elements within a certain range of this sensor. Moreover, we break down the analysis of localization errors in a large-scale sensor network into the analysis in small-scale sensor networks, termed unit networks, in which tight upper and lower bounds on the CRLB can be established. Finally, we investigate two practical issues: the applicability of the analysis based on the CRLB and the optimal anchor placement.