Abstract
Most existing array processing algorithms are very sensitive to model uncertainties caused by the mutual coupling and sensor location error. To mitigate this problem, a novel method for direction-of-arrival (DOA) estimation and array calibration in the case of deterministic signals with unknown waveforms is presented in this paper. The analysis begins with a comprehensive perturbed array output model, and it is effective for various kinds of perturbations, such as mutual coupling and sensor location error. Based on this model, the Space Alternating Generalized Expectation-Maximization (SAGE) algorithm is applied to jointly estimate the DOA and array perturbation parameters, which simplifies the multidimensional search procedure required for finding maximum likelihood (ML) estimates. The proposed method inherits the characteristics of good convergence and high estimation precision of the SAGE algorithm. At the same time, it forms a unified framework for DOA and array perturbation parameters estimation in the presence of mutual coupling and sensor location error. The simulation results demonstrate the effectiveness of the algorithm.
Highlights
The problem of estimating direction-of-arrival (DOA) of multiple narrowband signals plays an important role in many fields, including radar, wireless communications, seismology, and sonar
It is well known that both EM algorithm and Space Alternating Generalized Expectation-Maximization (SAGE) algorithm have been applied to the question of DOA estimation without array imperfections
The proposed algorithm gives a unified framework for DOA estimation in the presence of mutual coupling and sensor location error
Summary
The problem of estimating direction-of-arrival (DOA) of multiple narrowband signals plays an important role in many fields, including radar, wireless communications, seismology, and sonar. Most of the existing methods only focus on certain type of array imperfection, such as mutual coupling [1,2,3,4], gain/phase uncertainty [5,6,7], and sensor location error [8, 9]. Through the augmentation scheme specified by the EM or SAGE algorithm, the complicated multidimensional search involved in maximizing likelihood functions can be simplified to one-dimensional search It was proved in [14] that SAGE converges faster than EM while retaining the advantages of numerical simplicity and stability due to its flexible augmentation scheme. Compared to the existing method, the proposed algorithm can achieve higher estimation precision; at the same time, it follows a unified framework to address the DOA estimation problem in the presence of array imperfections, with typical perturbations of mutual coupling and sensor location error taken into consideration.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
