A Localization Method Using Frequency Difference of Arrival with Multiple Moving Observers

To address the localization challenges posed by the highly nonlinear nature of frequency difference of arrival (FDOA) equations in passive narrowband signal localization, this paper proposes a multiple-hypothesis Gauss-Newton (MNGN) iterative localization method. Initially, the method calculates multiple potential initial source positions using a single FDOA measurement. Subsequently, the Gauss-Newton (GN) algorithm is employed to iteratively refine these initial estimates while evaluating the cost function at each iteration. Finally, the solution with the lowest cost function value is selected as the final localization result. The proposed approach demonstrates superior localization performance compared to particle swarm optimization (PSO) and simulated annealing (SA) methods. It achieves Cramér-Rao lower bound (CRLB) accuracy under moderate Gaussian noise conditions while significantly reducing computational complexity.