Directional Measurements and Propagation Models at 28 GHz for Reliable Factory Coverage

Dmitry Chizhik, Jinfeng Du, Reinaldo A. Valenzuela, Dragan Samardzija, Stepan Kucera, Dmitry Kozlov, Rolf Fuchs, Juergen Otterbach, Johannes Koppenborg, Paolo Baracca, Mark Doll, Ignacio Rodriguez, Rodolfo Feick, Mauricio Rodriguez

Research output: Contribution to journalJournal articleResearchpeer-review

3 Citations (Scopus)
32 Downloads (Pure)

Abstract

Directional measurements of over 2600 links in four distinct factories at 28 GHz are used to formulate the path gain and azimuth gain models to allow reliable 90% coverage estimates. A simple theoretical model of path gain, dependent on ceiling and clutter heights, is found to represent path gain across the four factories with 4.4 dB root-mean-square error (RMSE), contrasted with 6.9 dB slope-intercept fit and 8.5-14.9 dB RMSE for 3GPP factory models. The model also did well against 3.5 GHz path loss data collected over 18 MHz bandwidth in one of the factories, with an RMSE of 3.3 dB. In nonline-of-sight (NLOS) conditions, scattering reduces available antenna azimuth gain from nominal value by up to 7.3 dB in 90% of links. Line-of-sight (LOS) blockage by a 1.7 m × 1 m obstacle in factory aisle leads to 7 dB signal reduction, attributed to availability of other paths. It is found that an access point (AP) using 25 dBm transmit power per polarization, with 23 dBi nominal gain and omnidirectional terminals, supporting 2× 2 MIMO in a 400 MHz bandwidth, can provide 130 Mb/s for 90% of factory locations within 50 m.

Original languageEnglish
JournalIEEE Transactions on Antennas and Propagation
Volume70
Issue number10
Pages (from-to)9596-9606
Number of pages11
ISSN0018-926X
DOIs
Publication statusPublished - 1 Oct 2022

Keywords

  • Factory
  • industrial Internet of Things (IIoT)
  • measurement
  • path loss
  • propagation

Fingerprint

Dive into the research topics of 'Directional Measurements and Propagation Models at 28 GHz for Reliable Factory Coverage'. Together they form a unique fingerprint.

Cite this