Bias Compensation Method for 3-D AOA-TMA with Uncertainty in Sensor Positions

Feifei Pang, Kutluyil Doǧançay, Haiyan Wang, Xiaohong Shen

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

This article focuses on the problem of target motion analysis (TMA) in 3-D space using angle of arrival (AOA) measurements taken by a moving sensor whose positions are inaccurate. The AOA measurement consists of a pair of azimuth and elevation angles in 3-D where azimuth and elevation angles are used jointly to estimate target motion parameters. To start with, the maximum likelihood estimator (MLE), the pseudolinear estimator (PLE), the weighted instrumental variable estimator (WIVE), and the selective-angle-measurement (SAM)-WIVE with uncertainty in sensor positions are derived. Under the effect of sensor position errors, the WIVE would be no longer asymptotically unbiased due to the correlation between the instrumental variable (IV) matrix and the pseudolinear noise vector. As a result, the bias of the WIVE is theoretically analyzed by modeling the sensor position error as the additional azimuth and elevation noises. Then, a new bias-compensated SAM-WIVE (BCSAM-WIVE) is derived based on the bias analysis, which can mitigate the bias of the PLE and IV-based methods caused by both AOA measurements and sensor position errors. Moreover, the Cramer-Rao lower bound (CRLB) is derived when sensor positions are not precisely known. Simulation results validate the superior performance of the proposed BCSAM-WIVE over other existing methods.

Original languageEnglish
Pages (from-to)14482-14492
Number of pages11
JournalIEEE Sensors Journal
Volume24
Issue number9
DOIs
StatePublished - 1 May 2024

Keywords

  • 3-D angle of arrival (AOA) target motion analysis (TMA)
  • bias compensation
  • pseudolinear (PL) estimation
  • sensor position uncertainty

Fingerprint

Dive into the research topics of 'Bias Compensation Method for 3-D AOA-TMA with Uncertainty in Sensor Positions'. Together they form a unique fingerprint.

Cite this