A fast algorithm for wireless sensor networks localization based on Metric least squares scaling

Accurate and low-cost wireless sensor networks localization is now a very popular research topic. Multidimensional Scaling (MDS) localization techniques have been successfully applied to the problem of node locating. In this paper, a relaxation algorithm which is based on Metric least squares scaling (LSS), a MDS technique, is devised for sensor location estimation. We investigate several variants of Metric LSS and compare their rate of convergence with relaxation algorithm. Simulation results show that our method achieves convergence faster than others at the same expense of computational cost. At last, the simulation results of localization accuracy comparison are given.