Complex Domain Super MDS: Computationally Efficient Localization via Ranging and Angle Information

We address the localization problem from ranging and angle information with the aim of deriving an algorithm that achieves high accuracies at lower computational demands than the state of the art. We focus in particular on the multidimensional scaling (MDS) approach, offer modified versions of the super MDS (SMDS) scheme, which has been shown to be able to robustly exploit both ranging and angle information for superior localization accuracies. The new algorithm is obtained by first recasting the SMDS formulation as a linear system in the complex domain, allowing for multiple maximum ratio combiner (MRC) variations of the SMDS to be designed, which in turn enables the adaptation to different conditions in terms of availability of information. Amongst various possible implementations of the new approach, we describe two in particular which are directly compared to the original SMDS: a singleton, and a cooperative variation. The singleton alternative is found to outperform the original in terms of accuracy at single-target scenarios, while the cooperative version is shown to exhibit significantly lower complexity in larger networks.