Localization Using Modified Stochastic Proximity Embedding Under Correlated Shadowing

Localization is the process of finding the location coordinates of a node. Distances from nodes with known coordinates is required for this computation. In most of the literature, the errors in these distance measurements are assumed to be independent. However, in the real world this does not hold true. There is a need to design algorithms for the case when the errors are correlated. In this work, the Stochastic Proximity Embedding algorithm is modified to provide improved performance under such correlated errors. Simulation studies are conducted to evaluate the performance of this algorithm. Semi Definite Programming approach and the original Stochastic Proximity Embedding Algorithm are used for comparison.