Direction-of-Arrival Estimation for Nested Acoustic Vector-Sensor Arrays Using Quaternions

Yi Lou, Xinghao Qu, Dawei Wang, Julian Cheng

Research output: Contribution to journalArticlepeer-review

15 Scopus citations

Abstract

There is an increasing interest in direction-of-arrival (DOA) estimation using nested arrays composed of vector sensors. Considering acoustic vector sensors (AVSs) commonly used in underwater applications, this article proposes a novel algorithm, called augmented nested quaternion-multiple signal classification (ANQ-MUSIC), to perform DOA estimation. By judiciously arranging the multicomponent outputs of AVSs, we model the received signals from the entire nested AVS array (NAA) as a quaternion observation vector in a compact way to reduce computational complexity. Next, we formulate a quaternion-based difference co-array (QDCA) model via vectorizing the quaternion covariance matrix (QCM). Based on the obtained insights from the QDCA model, we derive a suitable QCM, which is constructed by applying the spatial smoothing (SS) technique. Finally, classical quaternion-MUSIC (Q-MUSIC) is logically introduced to estimate the DOA parameters. In simulations, we take into account nonuniform received noise and intercomponent correlated (ICC) noise, which may occur in practical underwater environments. The results demonstrate that the proposed method shows superiority in angular resolution and achieves a desirable tradeoff between estimate accuracy and computational burden, besides showing robust performance in the above test scenarios.

Original languageEnglish
Article number4204714
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume61
DOIs
StatePublished - 2023

Keywords

  • Acoustic vector sensors (AVSs)
  • direction-of-arrival (DOA) estimation
  • nested arrays
  • quaternion theory
  • subspace-based methods

Fingerprint

Dive into the research topics of 'Direction-of-Arrival Estimation for Nested Acoustic Vector-Sensor Arrays Using Quaternions'. Together they form a unique fingerprint.

Cite this