A Fast-Performing Error Simulation of Wideband Radiation Patterns for Large Planar Phased Arrays With Overlapped Subarray Architecture

We present an efficient solution for the fast computation of wideband radiation patterns of large planar arrays with overlapped subarray architecture, subject to beamforming errors. We obtain very good results fully hosting the simulation on a desktop platform, and even better results implementing a...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on antennas and propagation 2014-04, Vol.62 (4), p.1779-1788
Hauptverfasser: Ricciardi, Gerald F., Connelly, James R., Krichene, Hedi A., Ho, Minhtri T.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We present an efficient solution for the fast computation of wideband radiation patterns of large planar arrays with overlapped subarray architecture, subject to beamforming errors. We obtain very good results fully hosting the simulation on a desktop platform, and even better results implementing a small portion of code on commercially-available graphical processing units (GPUs). Errors can be introduced throughout the beamforming chain and include random phase/magnitude errors, element failures, and surface deformation errors. At the core of the computational architecture are interchangeable primitives that quickly compute the entire far-field subarray pattern, subject to errors and user-defined tapers. The primitives are routines that address the general case of arrays with non-uniformly spaced elements (to model surface distortion) and the special case of arrays with uniformly-spaced elements. Both grid-type primitives are CPU hosted, while only the general non-uniform grid primitive type is GPU hosted due to its excellent run-time performance. The simulation quickly generates wideband patterns for the entire forward-looking hemisphere with sufficient resolution to accurately evaluate directivity and sidelobe metrics. Using only the non-uniform grid type hosted on the GPU, we show significant run-time improvement over both CPU-hosted implementations: 18.0 × over the general non-uniform grid and 5.7 × over the uniform grid.
ISSN:0018-926X
1558-2221
DOI:10.1109/TAP.2013.2255576