Location Based Beamforming in 5G Ultra-Dense Networks

Petteri Kela, Mário Costa, Jussi Turkka, Mike Koivisto, Janis Werner, Aki Hakkarainen, Riku Jäntti

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review


In this paper we consider transmit (Tx) and receive (Rx) beamforming schemes based on the location of the device. In particular, we propose a design methodology for the Tx/Rx
beamforming weight-vectors that is based on the departure and arrival angles of the line-of-sight (LoS) path between access-nodes (ANds) and user-nodes (UNds). A network-centric extended Kalman filter (EKF) is also proposed for estimating and tracking the directional parameters needed for designing the Tx and Rx beamforming weights. The proposed approach is particularly useful in 5G ultra-dense networks (UDNs) since the high-probability of LoS condition makes it possible to design geometric beams at both Tx and Rx in order to increase the signal-to-interferenceplus-noise ratio (SINR). Moreover, relying on the location of the UNd relative to the ANds makes it possible to replace fullband uplink (UL) reference signals, commonly employed for acquiring the channel-state-information-at-transmitter (CSIT) in time-division-duplex (TDD) systems, by narrowband UL pilots. Also, employing the EKF for tracking the double-directional parameters of the LoS-path allows one to reduce the rate at which UL reference signals are transmitted. Consequently, savings in terms of time-frequency resources are achieved compared to beamforming schemes based on full-band CSI. Extensive numerical results are included using a realistic ray-tracing based system-level simulator in ultra-dense 5G network context. Results show that position based beamforming schemes outperform those based on full-band CSI in terms of mean user-throughput even for highly mobile users.
Original languageEnglish
Title of host publication2016 IEEE 84th Vehicular Technology Conference, VTC Fall 2016 - Proceedings
Number of pages7
ISBN (Electronic)9781509017010
Publication statusPublished - 17 Mar 2017
MoE publication typeA4 Article in a conference publication
EventIEEE Vehicular Technology Conference - Montreal, Canada
Duration: 18 Sep 201621 Sep 2016
Conference number: 84

Publication series

NameIEEE Vehicular Technology Conference Proceedings
ISSN (Print)1550-2252


ConferenceIEEE Vehicular Technology Conference
Abbreviated titleVTC Fall
Internet address


  • 5G mobile communication systems
  • Channel state information
  • Communication channels (information theory)
  • Extended Kalman filters

Fingerprint Dive into the research topics of 'Location Based Beamforming in 5G Ultra-Dense Networks'. Together they form a unique fingerprint.

Cite this