Bisector angle estimation in a nonsymmetric multipath radar scenario

M. D. Zoltowski*, Ta-Sung Lee

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

In the classical problem of low-angle radar tracking, echoes return to the array via a specular path as well as by a direct path, with the angular separation between the two ray paths a fraction of a beamwidth. The performance of any bearing estimation scheme in this scenario is dependent on the phase difference between the direct and specular path signals at the centre of the array. The beamspace domain maximum likelihood (BDML) bearing estimator is a recently developed three-beam extension of the sum and difference beam technique employed in conventional radar. Nonsymmetric BDML breaks down when the phase difference is either 0° or 180°. In contrast, symmetric BDML, in which the point angle of the centre beam is the bisector angle between the two ray paths, can theoretically handle any phase difference, with the 0° case giving rise to the best performance. A simple, closed-form bisector angle estimator is developed based on characteristic features of the 3 × 3 forward-backward averaged beamspace correlation matrix when the centre pointing angle is the true bisector angle. In this way, a 2-D parameter estimation problem is decomposed into two successive 1-D parameter estimation problems: estimation of the bisector angle, followed by estimation of the target bearing. Simulations are presented assessing the performance of the new bisector angle estimator and comparing the performance of symmetric BDML employing the new estimator with other ML based bearing estimation schemes in a simulated low-angle radar tracking environment.

Original languageEnglish
Pages (from-to)559-570
Number of pages12
JournalIEE Proceedings, Part F: Radar and Signal Processing
Volume138
Issue number6
DOIs
StatePublished - 1 Jan 1991

Fingerprint Dive into the research topics of 'Bisector angle estimation in a nonsymmetric multipath radar scenario'. Together they form a unique fingerprint.

Cite this