Direct position determination of multiple sources using a moving virtual interpolation array

Abstract

Direct position determination (DPD) refers to determining the target position directly without estimating intermediate positioning parameters. Compared to the traditional two-step methods, it avoids parameter correlations and significantly enhances the algorithm's adaptability to low Signal-to-Noise Ratio (SNR) conditions. This paper uses coprime arrays to investigate direct positioning in a motion single-station passive localization system. Addressing issues where current algorithms fail to fully utilize array aperture and perform poorly in low snapshot scenarios, this paper proposes a motion single-station DPD algorithm based on virtual interpolated arrays. The proposed algorithm first uses the $l_0$ atomic norm to estimate the covariance matrix after filling gaps in the difference co-array. Then, the MVDR (Minimum Variance Distortionless Response) method is applied to fuse covariance estimates for localization. Additionally, we derive the Cramér-Rao lower bound. Numerical simulations validate the algorithm's performance, demonstrating its ability to maximize the degrees of freedom provided by coprime arrays and achieve superior performance in scenarios with short snapshots.