Robust 3-D AOA Localization Against Malicious Attacks in Non-Gaussian Noise
Qingli Yan, Hui‐Ming Wang, Yanping Chen, Cong Gao
Abstract
This paper addresses the robust angle-of-arrival (AOA) localization problem for three dimensional (3D) sensor networks in adversarial environment. The problem becomes more challenging in practical scenarios where attack information is unknown and background noise has heavy tailed property. With reasonable assumptions on the attack model and non-Gaussian noise, we propose a robust sparse regularization formulation by adopting the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">l</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> -norm as loss function to meet the need for heavy-tailed noise and nonconvex Cauchy function as penalty due to the sparsity of attack perturbations. To solve this nonconvex and nonsmoothed source localization problem, we develop a new smoothed Cauchy-regularized least absolution (SCLA) estimator based on the alternating direction of multipliers method (ADMM), and provide the sufficient convergence condition for SCLA. Unfortunately, the solution of SCLA is biased due to the correlation between noise vector and measurement matrix in the pseudolinear measurement equations. To handle this bias problem, an instrumental-variable based Cauchy-regularized least absolution (IVSCLA) estimator is finally proposed by introducing the instrumental-variable estimator into SCLA. Extensive simulation experiments demonstrate the performance of IVSCLA has advantages over state-of-the-art algorithms in robustness and effectiveness for different noise distribution.