A time domain approach to sleep modes in dense small cellular networks

Abstract

Data traffic demands in cellular networks continue to increase exponentially and operators are continuously upgrading their networks to meet the demand. The resulting capital and operational expenditures have limited revenues for the operating companies. Furthermore, this network denitrification associated energy costs. Network denitrification is thus not an appropriate approach to address energy efficiency (EE) objective despite it addressing the capacity objective. In this project, we use the system-level simulation to investigate different sleep mode mechanisms that can address both capacity and energy efficiency objectives in dense small cell networks. We derive a multi-user connectivity model, of uniformly distributed base stations (BSs) and users with Poisson Point Process (PPP) using MATLAB, which facilitates the study of the sleep mode mechanisms. Using traffic profiles of real-life cellular networks from AFRICELL Uganda, for the study case of Makerere and its neighbouring sites, we determined along-term traffic profile thatcanbeusedtoadaptenergyconsumptiontotheprevailingtrafficconditions. Following the traffic profiles, it was justified that a base station may have no traffic request for a certain period, which is an opportunity to save energy. On that basis, the selection criterion of candidates for sleep mode was a matter to consider. We therefore designed an adoptive approach that determines the required base station density in response to a variable long-term traffic profile by implementing conventional, random and constrained sleep mode mechanism Our analysis shows that significant energy savings are achievable when using the different sleep mode mechanisms but better still, constrained sleep mode mechanism out performs the random and conventional mechanisms in the simulated area.