Experimental Evaluation of Data-driven Signal Level Estimation in Cellular Networks

Melisa Maria Lopez Lechuga, Troels Bundgaard Sørensen, Istvan Kovács, Jeroen Wigard, Preben E. Mogensen

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

4 Citationer (Scopus)

Abstract

Estimating accurately the signal levels that a user equipment experiences along a movement route is a key step in the process of providing and guaranteeing the required service quality. Obtaining accurate location-specific estimations of the signal level is challenging due to its random variations. In this paper we investigate the use of aggregated measurements from multiple User Equipments (UE) to estimate the serving Reference Signal Received Power (RSRP) that the user will experience along a route. We use LTE measurements obtained in rural and urban areas from drive tests and analyze the dependence of data variability. Results show that the accuracy of data-driven estimation is impacted significantly by the variability in the underlying data due to UE orientation, UE characteristics and their immediate environment. With compensation for a subset of these effects the standard deviation of the estimation error can be lowered from an overall approximately 8 dB down to 4dB.
OriginalsprogEngelsk
Titel2021 IEEE 94th Vehicular Technology Conference (VTC2021-Fall)
Antal sider6
ForlagIEEE
Publikationsdato2021
Artikelnummer9625559
ISBN (Trykt)978-1-6654-1369-5
ISBN (Elektronisk)978-1-6654-1368-8
DOI
StatusUdgivet - 2021
Begivenhed2021 IEEE 94th Vehicular Technology Conference (VTC2021-Fall) - Norman, USA
Varighed: 27 sep. 202130 sep. 2021

Konference

Konference2021 IEEE 94th Vehicular Technology Conference (VTC2021-Fall)
Land/OmrådeUSA
ByNorman
Periode27/09/202130/09/2021
NavnIEEE Vehicular Technology Conference. Proceedings
ISSN1090-3038

Fingeraftryk

Dyk ned i forskningsemnerne om 'Experimental Evaluation of Data-driven Signal Level Estimation in Cellular Networks'. Sammen danner de et unikt fingeraftryk.

Citationsformater