Modélisation énergétique de réseaux à antennes distribuées pour les communications sans-fil indoor. (Energy Characterisation of Distributed Antenna Systems for Indoor Wireless Communications)

The increasing number of mobile terminals in home and in-building networks for indoor environments including offices, train stations and airports entails demanding bit-rate requirements with unprecedented increase in energy consumption. Considering the energy required per user terminal to access the Internet, more than half of the total energy used from the core network to the end user terminal is consumed in the home / in-building network. In the purpose to limit the energy consumed within an in-building infrastructure, whilst enabling higher bit-rates, a promising approach is to provide the wireless coverage by sharing the transmitted power among several distributed antennas. Therefore, a distributed antenna system (DAS) using radio-over-fiber (RoF) links is considered in the present study. One objective is to model and compute the system energy efficiency, as defined by the consumed power to the effective bit-rate ratio, and to compare it to that of a conventional single antenna system. The simulation results based on both power consumption measurements in Wi-Fi networks and power consumption modeling of radio-over-fiber links, show that a RoF-DAS can be up to three times more energy efficient than a conventional system for the same total radiated power. For the deployment of distributed antenna systems in indoor areas, a method is proposed to optimize their energy efficiency. Knowing the probability of presence of the terminals within a given area, the proposed method can determine the number of distributed antennas and the total radiated power. Finally, the present thesis studies algorithms for setting home equipment in standby mode and switching off distributed antennas when no connection is active. These algorithms add on to the aforementioned modeling and enable further reduction of energy consumption in home networks by more than 45% for an eight-hour day.