Optimizing Power and Resource Management for Multiuser MIMO/OFDM Systems Ying Jun Zhang and Khaled Ben Letaief § Center for Wireless Information Technology Electrical and Electronic Engineering Department The Hong Kong University of Science and Technology Clear Water Bay, Kowloon, Hong Kong Email: yjzhang@ust.hk , eekhaled@ee.ust.hk Abstract- Fast adaptive transmission has been recently recognized as an essential method to exploit the system diversity and enhance power/spectral efficiency in time-varying channels. In this paper, we propose adaptive transmission algorithms for multiuser MIMO/OFDM systems. The algorithms adaptively control multiple-access, modulation, and transmit power to exploit the channel variation in space, frequency, and user domains. The ultimate objective is to optimize the overall power efficiency while ensuring the fulfillment of every user’s QoS requirements. A major challenge in the problem lies in the optimization complexity due to the presence of non-linear, non-convex constraints. The proposed algorithms successfully reduce the optimization complexity and solve the noise enhancement problem that exists in multiuser systems with non-orthogonal signatures. Numerical results show that significant improvement in power/spectral efficiency is achievable with reasonable computational complexity when the above schemes are applied. I. INTRODUCTION With the explosive growth of wireless communications, the development of fast adaptation algorithms in both the physical and MAC layer is essential to provide high-speed transmission over scarce frequency spectrum. Orthogonal Frequency Division Multiplexing (OFDM) is a very promising technique to provide a high-performance physical layer thanks to its ability to combat ISI (intersymbol interference) and flexibility for adaptive frequency channel control [1, 2]. Recently, there has been increasing interest in multiuser resource allocation for OFDM systems. In [3-5], multiuser resource allocation schemes with QoS guarantee have been proposed in the context of OFDM. Spectral efficiency is significantly improved by the algorithms therein due to the exploitation of multiuser diversity. MIMO (Multiple Input Multiple Output) antenna systems have been recently acknowledged as one of the most promising techniques to achieve dramatic improvement in system performance and capacity [6, 7]. In [8, 9], adaptive multiuser transmission for MIMO systems has been investigated to maximize the system capacity through power control. Nevertheless, none of these works take into consideration every single user’s QoS (Quality of Service). In other words, although the data rate-sum is maximized, some user’s data rate might be very low for certain time intervals due to poor channel conditions. In this paper, we propose adaptive multiuser resource allocation algorithms for MIMO/OFDM systems that involve adaptive multiple-access control, adaptive modulation, and power control. The objective is to optimize the power efficiency while ensuring the fulfillment of each user’s QoS requirements including the bit rate and BER (bit error rate). In [3-5], different users are separated by transmitting on different subcarriers. In this paper, multiple users are allowed to transmit on the same subcarriers because they can be separated in the space domain by multiple antennas. This, however, introduces CCI (co- channel interference). The existence of CCI makes our problem much more complicated than that in [3-5]. In contrast to [8, 9], this paper proposes algorithms that can specify the allocation of subcarriers, power and modulation modes for every user. Moreover, the instantaneous QoS provisioning makes the optimization problem in this paper much more challenging. The complexity of the power-efficiency optimization problem is extremely high because of the presence of the nonlinear, non- convex constraints resulting from CCI and QoS guarantee. To deal with the prohibitively high complexity, low complexity algorithms are proposed. The rest of this paper is organized as follows. In Section II, the system model is described. An SVD-based (singular value decomposition) SDMA scheme is introduced in Section III. The joint optimization problem is then formulated in Section IV. In this case, zero forcing (ZF) is applied to decouple the joint optimization problem into simpler single user optimization problems. The problem of power-enhancement that exists in the ZF-based algorithm is solved in Section V by proposing another algorithm that introduces orthogonality in the frequency domain to those users who lack separability in space. The performance of both algorithms is demonstrated in Section VI. Finally, the paper is concluded in Section VII. II. SYSTEM MODEL In this paper, we investigate a multiple access system based on the use of a MIMO/OFDM structure. Assume that there are K users in the system, with k J denoting the number of transmit antennas for user k. Let m k x denote the data symbol transmitted by user k on subcarrier m and m k p , denotes the relevant transmit power. m k x is then multiplied by a transmit weight vector m k T to form the data symbols to be transmitted by the k J antennas on subcarrier m. Assume that there are L receive antennas at the BS and that l m r , is the data received by antenna l on subcarrier m. ] , , , [ . 2 , 1 , L m m m r r r L is then weighed and combined by a receive § This work is supported in part by the Hong Kong Research Grant Council under Grant #HKUST6164/02E GLOBECOM 2003 - 179 - 0-7803-7974-8/03/$17.00 © 2003 IEEE