Optimized confidence weights for localization algorithms with scarce information

A method to derive weights to be used in distance-based multi-dimensional scaling (MDS) source localization algorithms under scarce information is discussed. In particular, a family of weighing function is derived with basis on small-scale statistics and the parameter that drives the choice of a particular weighing function out of such a family is optimized with basis on an information-theoretical criterion. It is found that, under the condition of scarce information, the proposed weighing strategy outperforms the alternative of utilizing an approximation of the maximum-likelihood (ML) weighing strategy.