Resource allocation for dynamic TDD systems

Abstract

Mobile telecommunication system is in constant development, as the number of users and applications is steadily increasing. Fifth generation of mobile system (5G) is supposed to bring further improvements regarding data rates, reliability, energy efficiency and security. In this thesis we will focus on increasing data rate by applying time division duplex (TDD). In attempt to increase sum-rate, we propose a system where each cell is scheduled in downlink (DL) or uplink (UL) based on the traffic demand within the cell. The flexibility of such a system requires deploying TDD instead of nowadays commonly used frequency division duplex (FFD), because it allows dynamical adaptation to asymmetric traffic requirements. Despite advantages of dynamic TDD, it introduces higher inter-channel interference due to the fact that neighboring base stations are allowed to transmit in opposite directions. Due to this, we propose a method to avoid interference and maximize sum-rate. Two questions have to be answered, how to find scheduling and how to allocate transmit powers, which optimize a system in terms of maximal sum-rate. We will address these issues from the point of view where exists an central unit which oversees the whole system. Further, a decentralized method of optimization is proposed where scheduling and power allocation is done locally based just on limited communication with neighboring cells. Finally, we will do MATLAB simulation of all proposed algorithms and compare their results.