The paper first analyzes different multi-cell modelling approaches. Compared to a realistic London environment, hexagonal cell modeling’s received SINR is 5-7dBs more optimistic. However, the shape of SINR profiles for a variety of modelling approaches are similar and a back-off factor can be used to yield 80% correlation between the performance profiles. The second part of the paper characterizes the tradeoff between energy saving and coverage probability using up-to-date traffic data. The results show that for a certain minimum coverage threshold, a sleep mode threshold that yields the lowest energy consumption can be found. The energy savings achieved are 13-42%, and the key result is that increasing the number of cells in sleep mode also reduces the interference, and improves the coverage probability across the area. The third part of the paper analyzed the energy saving potential of low-power-nodes (LPNs) deployed in a heterogeneous configuration and integrated this technique with sleep mode. It was found that random deployment can degrade the network-wide capacity, but articulated deployment can significantly improve the capacity. The combined solutions yielded an energy saving of 15-46% in a realistic London scenario.
Acknowledgement
The work reported in this paper has formed part of the EPSRC Knowledge Transfer Account and Green Radio Core 5 Research Programmer of the Virtual Centre of Excellence in Mobile and Personal Communications, Mobile VCE. The authors would like to acknowledge the invaluable support given by: John Turk, David Lister, Eric Murray of Vodafone; Sunil Vadgama and Mythri Hunukumbure of Fujitsu. Fully detailed technical reports on this research are available to Industrial Members of the Mobile VCE. www.mobilevce.com