An Improved Variable Step Size LMSTDE Algorithm Based on Fourth-order Cumulant

Abstract

Two-stage positioning method is widely used in multi-star positioning for satellite interference sources. The estimating accuracy of time difference of arrival (TDOA) has an important impact on the positioning results. Under the influence of correlated Gaussian noise and low signal-to-noise ratio (SNR), the estimating accuracy of TDOA is seriously reduced. Therefore, this paper proposes an improved least mean square time difference estimation (LMSTDE) algorithm based on fourth-order cumulant. Firstly, the signal is preprocessed by using the self fourth-order cumulant and cross fourth-order cumulant to filter out the correlated Gaussian noise. Then, a new variable step size model based on arctangent function is proposed to improve the estimating accuracy under low SNR. Simulation results show that, compared with other methods, the proposed algorithm can improve the convergence speed and the estimating accuracy of TDOA under the correlated Gaussian noise and low SNR.