Robustness Analysis of Source Localization Using Gaussianity Measure
Kun Yan, Hsiao‐Chun Wu, S. Sitharama Iyengar · 2008
Nowadays, the source localization has been widely applied for wireless sensor networks. The Gaussian mixture model has been adopted for maximum-likelihood (ML) source localization schemes. However, this model does not match the statistics of the real data in practice. In this paper, we study the probability density function of the sensor signals and demonstrate that the distribution is not Gaussian. We propose to employ the Gaussianity test based on the bootstrap algorithm to quantify the departure of Gaussianity for the received signals added with different kinds of noise. Our proposed Gaussianity test can be used as the robustness figure for evaluating the prevalent ML source localization schemes.