1 Adaptive Power Control for 4G OFDMA systems on Frequency Selective Fading Channels Rainer Schoenen, Fei Qin, Department of Communication Networks (ComNets), Faculty 6, RWTH Aachen University {rs,fei}@comnets.rwth-aachen.de Abstract—Future cellular radio systems aim at maximizing the spectral efficiency. OFDMA radio resources are the scarce good with their dimensions bandwidth, time and space. Due to frequency selective fading the effective pathloss varies in all dimensions. Adaptive algorithms are available which allocate the best modulation and coding scheme depending on the expected SINR, as well as dynamic subchannel assignment which aims to choose the best subchannel for each user. This already boosts the performance of OFDMA system. However, these algorithms alone do not touch the transmitted power per subchannel. On the cell edge this is fine, but large areas are covered with a transmitted power exceeding the usually required SINR. In this paper we introduce a power control which saves power on the users within the cell. This leads to a reduced interference into neighbor cells, especially for future reuse one systems. Also some of the saved power can be used to boost transmissions at the cell edge. In this paper we introduce an adaptive power control concept and arrange it into a closed loop control system which contains blocks for all adaptive algorithms for modulation, power, subchannel usage and channel quality indication. Index Terms—Scheduling, Adaptive Power Control, CQI, FairSINR I. I NTRODUCTION R ADIO transmission power determines the coverage area of a radio cell. User terminals (UT) are assumed to be distributed evenly in a cell (constant user density), therefore UTs appear in all distance ranges from the base station (BS) from close to the BS to the cell border. This means a wide range of path loss values will appear. At a UT close to a BS (cell center users), there is typically plenty of received power P R (→ SINR) At the cell border, the SINR drops down to 0dB. Additionaly, fading is selective in frequency and time, so the received SINR also varies in these dimensions. Each subchannel (aggregation of OFDM subcarriers) is affected by this. For OFDMA, the decision on which subchannel to choose for which UT is taken by the Dynamic Subcarrier Assignment (DSA) task of a radio resource scheduler. It typically selects the best subchannel for each UT according to some metrics [1]. Resource scheduling like this requires channel state informa- tion (CSI). CSI is signaled as channel quality indication (CQI) from the UTs to the BS (or RN) [2], [3]. The variations of SINR within one subchannel are treated adaptively with Adaptive Modulation and Coding (AMC). It is usually performed by resource schedulers to chose a PhyMode (physical layer mode = Modulation and Coding Scheme) which optimally utilizes the available SINR [4] [5] Fig. 1. One Downlink TTI frame for LTE and approximate the Shannon bound (see Fig. 3). But there is also an inner region of the cell where the SINR is > 20dB and the highest PhyMode is always chosen. For these locations the transmit power is higher than necessary, which also means the interference into neighbor cells is much higher than required. Especially for future reuse one systems, the interference is an important task. Therefore, in addition to AMC and DSA [5], it is possible to regulate the output power on each transmitted subchannel selectively in frequency and time. This topic is rather new in the literature [6]. This Adaptive Power Control (APC) unit tries to compensate in the short-term for the fading notches and in the long term for the distance-caused path loss imbalance between UTs. In this paper it is assumed that the control is continuous and piecewise linear, with only an upper limit of the power per subchannel, but no (DAC) quantization and no lower limit. This paper discusses the control aspects of APC using a control theoretic view on the radio link. The APC strategy “FairSINR” is introduced which aims to provide each UT with the same SINR at the receiver. It is shown that this strategy and closed loop power control essentially mean the same. Section II defines the required aspects of Resource Schedul- ing, Section III focuses on the Adaptive Power Control aspect and Section IV integrates this as a building block into the control model. Section V presents simulation results of the APC performance.