关键词: 无源定位, 时差定位, 修正泰勒级数算法, 修正牛顿算法, 正则化算法

Abstract:

In the Time Difference Of Arrival (TDOA) source localization model, based on the Taylor-series (TS) method and Newton (NT) method, this paper presents the Modified Taylor-series(MTS) method and the Modified Newton method(MNT), which solve the critical convergent problem caused by the bad initial value in the original algorithms. The proposed algorithms modify the ill-condition Hessian matrix caused by the bad initial value using the Tikhonov (TI) or the Diagonal Singular Value Decomposition technique (DSVD) in the Regularization theory. The regularization parameter which controls the properties of the regularized solution is determined by the L-curve method. Simulation results show that compared with the TS and NT methods, the proposed methods ensure that the solution of the iterative methods converges on the source location, improves the convergent probability and has a better capability to remove the local minima. The proposed methods also give superior performances of the location accuracy comparing with the closed-form algorithms in low SNR environment.

Key words: source localization, time-difference-of-arrival, modified Taylor-series method, modified Newton method, regularization method