Distributed Algorithm for Hybrid TOA/DOA-based Source Localization

We consider source localization in a wireless sensor network in the absence of a fusion center, where each sensor makes hybrid time-of-arrival (TOA) and direction-of-arrival (DOA) measurements. Specially, in this paper, we focus on the bounded noise case. Distributed location estimation is posed as a consensus problem. The average consensus method in the literature can be used for a linear fusion estimator which can be written as the ratio of two averages. The disadvantage of this average consensus method is that it can only used for linear model and needs to wait for the final convergence to achieve an acceptable accuracy. However, generally its convergence rate is low. Alternatively, we present a convex projection consensus method where we show that the resulting estimator at each sensor converges to a common point lying in the intersection of convex regions that are determined based on local sensor measurements and also the minimum number of sensors required to meet a certain estimation error tolerance is derived. Numerical results are presented showing the convergence properties and root mean squared error under both the average consensus and the convex projection methods.