1: \begin{abstract}

2:   Real-world network applications must cope with failing nodes,

3:   malicious attacks, or, somehow, nodes facing corrupted data ---

4:   classified as outliers. One enabling application is the geographic

5:   localization of the network nodes. However, despite excellent work

6:   on the network localization problem, prior research seldom

7:   considered outlier data --- even now, when already deployed networks

8:   cry out for robust procedures. We propose robust, fast, and

9:   distributed network localization algorithms, resilient to high-power

10:   noise, but also precise under regular Gaussian noise. We use the

11:   Huber M-estimator as a difference measure between the distance of

12:   estimated nodes and noisy range measurements, thus obtaining a

13:   robust (but nonconvex) optimization problem. We then devise a convex

14:   underestimator solvable in polynomial time, and tight in the

15:   inter-node terms. We also provide an optimality bound for the convex

16:   underestimator. We put forward a new representation of the Huber

17:   function composed with a norm, enabling distributed robust

18:   localization algorithms to minimize the proposed underestimator. The

19:   synchronous distributed method has optimal convergence rate and the

20:   asynchronous one converges in finite time, for a given

21:   precision. The main highlight of our contribution lies on the fact

22:   that we pay no price for distributed computation nor in accuracy,

23:   nor in communication cost or convergence speed. Simulations show the

24:   advantage of using our proposed algorithms, both in the presence of

25:   outliers and under regular Gaussian noise: our method exceeds the

26:   accuracy of an alternative robust approach based on L1 norms by at

27:   least 100m in an area of 1Km sides.

28: \end{abstract}

29: