Fourth-Order Cumulants based Underdetermined 2-D DOA Estimation using Single AVS
Umesh Sharma, Monika Agrawal · Global Oceans 2020: Singapore – U.S. Gulf Coast · 2020
The prime motivation to use single Acoustic Vector Sensor for measuring acoustic intensity is lying in its compact size and the unambiguous two-dimensional direction of arrival estimation which is insensitive to the range, frequency and bandwidth of the impinging acoustic signal sources. While, the major limitation of its use for direction of arrival estimation stands with its limited degree of freedom leading to identification of the maximum two sources using conventional direction of arrival estimation methods. This paper proposes a compressive sensing based method utilising the fourth-order circular cumulants for two-dimensional direction of arrival estimation of multiple non-Gaussian signals. By using the proposed method, a single acoustic vector sensor be able to estimate two-dimensional direction of arrival of up to seven non-Gaussian statistical independent source signals corrupted by Gaussian noise but maximum four if all the sources lying in a single plane. Compared to existing fourth-order cumulants based methods, the proposed algorithm offers improved direction of arrival estimation performance such as higher spatial resolution, improved estimation accuracy and required lower signal to noise ratio and lesser number of snapshots to operate.