Quantizer Design for RSSI-based Target Localization in Sensor Networks
Ahmad A. Ababneh, Fadi R. Zghoul and Lutfa Akter
In RSSI-based target localization, estimation error depends on several factors (e.g., location of sensors, target energy). Furthermore, when quantization is involved, localization accuracy largely depends on the number of allocated bits and quantization thresholds used by each sensor. In this paper, we consider the problem of meeting estimation error requirements at different points of interest within the deployment area as best as possible under a bit rate budget constraint. Due to the complexity of the problem, we first address the bit allocation problem by proposing three different sequential allocation algorithms. Simulation results indicate that the error performance of proposed algorithms is within 2% − 20% of that of the optimal solution. Secondly, we propose the ESAQ adaptive scheme that incorporates target parameters and error profile of each sensor in quantization threshold selection. Simulation results show that the ESAQ outperforms both the EDQ and EBHQ thresholding methods found in literature.
Keywords: Sensor networks; localization; sensor placement; sensor relocation.