Resource Allocation for Power Minimization in the Downlink of THP-Based Spatial Multiplexing MIMO-OFDMA Systems

In this paper, we deal with resource allocation in the downlink of spatial multiplexing multiple-input–multiple-output (MIMO)-orthogonal frequency-division multiple-access (OFDMA) systems. In particular, we concentrate on the problem of jointly optimizing the transmit and receive processing matrices, the channel assignment, and the power allocation with the objective of minimizing the total power consumption while satisfying different quality-of-service (QoS) requirements. A layered architecture is used in which users are first partitioned in different groups on the basis of their channel quality, and then channel assignment and transceiver design are sequentially addressed starting from the group of users with most adverse channel conditions. The multiuser interference among users belonging to different groups is removed at the base station (BS) using a Tomlinson–Harashima precoder operating at user level. Numerical results are used to highlight the effectiveness of the proposed solution and to make comparisons with existing alternatives.

I. Introduction

Dynamic resource allocation in multiple-input multiple-output (MIMO) systems based on orthogonal frequency-division multiple-access (OFDMA) technologies has gained considerable research interest [1]. In most cases, subcarriers are assigned to the active users in an exclusive manner without taking advantage of the multiuser diversity offered by the spatial domain. A possible solution to exploit the spatial dimension is to make use of space-division multiple-access (SDMA) schemes, which allow the simultaneous transmission of different users over the same frequency band. The main impairment of SDMA is represented by multiple-access interference (MAI). In downlink transmissions, MAI mitigation can only be accomplished at the base station (BS) using prefiltering techniques. The most common approach for interference mitigation is zero-forcing (ZF) linear beamforming, which relies on the idea of preinverting the channel matrix at the transmitter. Another approach is represented by the block-diagonalization ZF (BD-ZF) scheme originally proposed in [2]. Particular attention has been also devoted to dirty-paper coding (DPC) techniques [3] although their implementation is still much open. A possible solution in this direction is represented by Tomlinson–Harashima precoding (THP), which can be seen as a 1-D DPC technique [4] and has been widely used in the downlink of single- and multi-user MIMO systems [5]– [8]. In combination with prefiltering, another way to deal with interference in SDMA-OFDMA systems is user partitioning, which basically consists in properly selecting the set of users transmitting on the same subcarriers. As illustrated in [9], a common approach is to first group together users whose channels have low spatial cross correlation and then to assign the subcarriers to the various groups. In [10], the authors follow a completely different approach in which the users are first divided into groups such that the spatial cross correlations among users in different groups are low as much as possible, and then subcarriers are sequentially assigned within each group.