Towards agile multilayer cellular networks

Location : IMT Atlantique, Rennes


Supervisors : X. Lagrange, L. Nuaymi and G. Texier (IMT Atlantique)


The spectrum used for mobile services is increasingly extended and this trend will be confirmed in the coming years: in addition to the traditional band between 700 and 2600 MHz, higher frequencies are being deployed (3.4 GHz) or planned (26 GHz) to eventually reach THz. This evolution does not only concern radio transmissions but has an influence on the network. The very low coverage resulting from very high frequencies leads to multi-layer networks  (also called hierarchical cell structure, HCS)  that mix small and large cells. This is a paradigm shift that has been identified as early as 2013 [And13], but has not yet been translated into reality because the layers interact weakly. The close integration of radio links with different characteristics is already possible with traditional technologies [Lag14] but more can be done. The virtualization of radio functions makes it possible to have agile networks whose radio resources are adapted to demand and where the traditional notion of a cell disappears, as a terminal is connected to several access points that vary over time.

Subject of the thesis

The objective of the thesis is to define different architectures in the access network and strategies of localization of the treatments in order to jointly optimize the performances (throughput, latency, …) and the energy consumption [LCX20]. The analysis will be based on models, but may use learning techniques if necessary.

The approach used is to take advantage of the possibilities offered by  Cloud Radio Access Network technology (C-RAN). The processing functions at the different network layers can be located in different places [Su19]. Processing closer to the user generally reduces latency but the counterpart is a lower mutualization and thus an increase of the cost. In addition, the load and the types of services used vary over time. Optimization must therefore take into account the state of demand. Finally, the separation of functions into several execution sites can lead to an increase in signaling. Global optimization must therefore be carried out. A promising approach has been proposed in [GCL18].

Required skills for the candidate

  • excellent background in networks, especially low layers (physical layer, MAC layer, cellular networks)
  •  background in mathematics (probability, optimisation)
  • scientific writing
  • and the traditional skills for a thesis (imagination, precision, tenacity, listening ability, team spirit)


[And13]       Andrews, J. G. (2013). Seven ways that HetNets are a cellular paradigm shift. IEEE communications magazine, 51(3), 136-144.

[Lag14]        Lagrange, X. (2014, April). Very tight coupling between LTE and Wi-Fi for advanced offloading procedures. In 2014 IEEE Wireless Communications and Networking Conference Workshops (WCNCW) (pp. 82-86). IEEE.

[GCL18]      A. Garcia-Saavedra, X. Costa-Perez, D. J. Leith and G. Iosifidis, “FluidRAN: Optimized vRAN/MEC Orchestration,” IEEE INFOCOM 2018 – IEEE Conference on Computer Communications, Honolulu, HI, 2018, pp. 2366-2374, doi: 10.1109/INFOCOM.2018.8486243.

[Su19]          R. Su et al., “Resource Allocation for Network Slicing in 5G Telecommunication Networks: A Survey of Principles and Models,” in IEEE Network, vol. 33, no. 6, pp. 172-179, Nov.-Dec. 2019, doi: 10.1109/MNET.2019.1900024

[LCX20]       Li, Y. N. R., Chen, M., Xu, J., Tian, L., & Huang, K. (2020). Power Saving Techniques for 5G and Beyond. IEEE Access, 8, 108675-108690.


Applicants must submit a CV, a cover letter and the official academic records for their bachelor and masters education. It is a requirement to hold a masters or an equivalent degree for being considered for this position. At least two references (name, position, e-mail, and telephone number) should be included in the application. Send the files to cand-doct-b5g at