Sources localization through matrix completion via Nystrom completion

Abstract

In this paper, the completion of missing measurements in a squared distances matrix through Nystrom completion algorithm have been investigated. this missing occurred due to limitation of power when the sensors are deployed in a large area. The Nystrom algorithm has overcome the classical multidimensional scaling in a low and moderate signal to noise ratio, in addition it performed well as the number of missing entries increased. The plotted figures show admissible consequences for the proposed algorithm.