On Learning and Exploiting Time Domain Traffic Patterns in Cellular Radio Access Networks

  • Jordi Pérez-RomeroEmail author
  • Juan Sánchez-González
  • Oriol Sallent
  • Ramon Agustí
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9729)


This paper presents a vision of how the different management procedures of future Fifth Generation (5G) wireless networks can be built upon the pillar of artificial intelligence concepts. After a general description of a cellular network and its management functionalities, highlighting the trends towards automatization, the paper focuses on the particular case of extracting knowledge about the time domain traffic pattern of the cells deployed by an operator. A general methodology for supervised classification of this traffic pattern is presented and it is particularized in two applicability use cases. The first use case addresses the reduction of energy consumption in the cellular network by automatically identifying cells that are candidates to be switched-off when they serve low traffic. The second use case focuses on the spectrum planning and identifies the cells whose capacity can be boosted through additional unlicensed spectrum. In both cases the outcomes of different classification tools are assessed. This capability to automatically classify cells according to some expert guidance is fundamental in future networks, where an operator deploys tenths of thousands of cells, so manual intervention of the expert is unfeasible.


Classification Cellular networks 5G Radio access network management 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    World Economic Forum. Enabling Transformation: Information and Communications Technologies and the Networked Society (2009)Google Scholar
  2. 2.
    Ericsson White Paper. More than 50 billion connected devices, February 2011.
  3. 3.
    Fallgren, M., Timus, B. (eds.): Scenarios, requirements and KPIs for 5G mobile and wireless system, Deliverable D1.1. of the METIS project, May 2013Google Scholar
  4. 4.
    El Hattachi, R., Erfanian, J. (eds.): NGMN 5G White Paper, NGMN Alliance, February 2015Google Scholar
  5. 5.
    Ericsson, Big Data Analytics. White paper, August 2013Google Scholar
  6. 6.
    Thomas, R.W., DaSilva, L.A., MacKenzie, A.B.: Cognitive networks. In: 1st IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks, (DySPAN), pp. 352–360 (2005)Google Scholar
  7. 7.
    Ramiro, J., Hamied, K.: Self-Organizing Networks. Self-planning, self-optimization and self-healing for GSM, UMTS and LTE. John Wiley & Sons (2012)Google Scholar
  8. 8.
    Banerjee, A.: Advanced Predictive Network Analytics: Optimize your Network Investments and Transform Customer Experience. White Paper, Heavy Reading (2014)Google Scholar
  9. 9.
    Han, J., Kamber, M.: Data Mining Concepts and Techniques, 2nd edn. Elsevier (2006)Google Scholar
  10. 10.
    Wilson, R.A., Keil, F.C.: The MIT Encyclopedia of the Cognitive Sciences. MIT Press (1999)Google Scholar
  11. 11.
    GSMA, Mobile’s Green Manifesto, 2nd edn., June 2012Google Scholar
  12. 12.
    RapidMiner Studio.
  13. 13.
    GPP workshop on LTE in unlicensed spectrum, Sophia Antipolis, France, June 13, 2014.
  14. 14.
    Al-Dulaimi, A., Al-Rubaye, S., Quiang, N., Sousa, E.: 5G Communications Race: Pursuit of More Capacity Triggers LTE in Unlincensed Band. IEEE Vehicular Technology Magazine 10(1), 43–51 (2015). doi: 10.1109/MVT.2014.2380631 CrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2016

Authors and Affiliations

  • Jordi Pérez-Romero
    • 1
    Email author
  • Juan Sánchez-González
    • 1
  • Oriol Sallent
    • 1
  • Ramon Agustí
    • 1
  1. 1.Universitat Politècnica de Catalunya (UPC)BarcelonaSpain

Personalised recommendations