Indirect Radio Interferometric Localization via Pairwise Distances

The Radio Interferometric Positioning System (RIPS), introduced by Maroti et. al. [1], provides a means for very accurate sensor localization with very minimal device hardware requirements. To avoid stopping in a significant number of locally optimal location solutions, RIPS employs a genetic optimization localization algorithm. This paper proposes an indirect localization algorithm which first estimates pairwise distances, and then uses them to estimate coordinates via distributed weighted multidimensional scaling (dwMDS). The pairwise distances are iteratively improved and coordinates re-estimated. While suboptimal, simulations show the proposed method can achieve 50 cm RMS location errors. The algorithm provides a tradeoff between computational complexity and accuracy, and may enable distributed RIM-based localization.