A multi-agent deep reinforcement learning (MADRL) is a promising approach to challenging problems in wireless environments involving multiple decision-makers (or actors) with high-dimensional continuous action space. In this paper, we present a MADRL-based approach that can jointly optimize precoders to achieve the outer-boundary, called pareto-boundary, of the achievable rate region for a multiple-input single-output (MISO) interference channel (IFC). In order to address two main challenges, namely, multiple actors (or agents) with partial observability and multi-dimensional continuous action space in MISO IFC setup, we adopt a multi-agent deep deterministic policy gradient (MA-DDPG) framework in which decentralized actors with partial observability can learn a multi-dimensional continuous policy in a centralized manner with the aid of shared critic with global information. Meanwhile, we will also address a phase ambiguity issue with the conventional complex baseband representation of signals widely used in radio communications. In order to mitigate the impact of phase ambiguity on training performance, we propose a training method, called phase ambiguity elimination (PAE), that leads to faster learning and better performance of MA-DDPG in wireless communication systems. The simulation results exhibit that MA-DDPG is capable of learning a near-optimal precoding strategy in a MISO IFC environment.

In this paper, we propose a novel deep-learning based precoding framework, called deep MIMO autoprecoder, that can be used to learn an optimal precoding policy for multiple-input multiple-output (MIMO) precoding problems, for which optimal solutions are not known. We divide the proposed framework into two steps: action value estimation and optimal action prediction. In the first step, reinforcement learning (RL) is applied to learn and estimate the state-action value function (also called Q-value function) of each discrete precoder in a codebook through interactions with environments of MIMO systems, while for the second step, a supervised learning model is used to predict an optimal precoder in a continuous precoder space that corresponds to the output from the first step. The proposed framework can be implemented by means of two deep neural networks: a deep RL algorithm, called deep Q-network (DQN), and a new neural network, that we call deep Q-value-to-policy network (DQPN). More specifically, an environment-specific DQN is trained for a MIMO scenario in order to estimate the Q-values of each precoder in environmental states. Meanwhile, a single universal DQPN is pretrained offline on the most general MIMO channel model that allows it to achieve generalization over unseen channels. The numerical results verify the effectiveness of the deep MIMO autoprecoder and show that the proposed precoding framework can outperform the conventional approximation algorithm.

A new study item termed downlink Multiuser Superposition Transmission (MUST) has been established in 3GPP Release 13 [1]. As the name suggests, MUST utilizes the concept of superposition coding for a multiuser transmission in LTE-Advanced (LTE-A) systems. While superposition coding can be applied at symbol level, the symbol-domain superposition coding not only increases network signaling overhead and implementation difficulty, but also causes performance degradation due to non-Gray mapping resulting from the symbol-domain superposition. In this paper, we propose a bit-domain multiuser superposition technique by exploiting the intrinsic properties of bit-interleaved coded modulation (BICM) which is one of key technologies for LTE systems. Simulation results show that the bit-domain MUST can increase both the cell-average and cell-edge user spectral efficiency by more than 20% compared to single-user MIMO (SU-MIMO).

The network-assisted interference cancellation and suppression (NAICS) feature has been studied by the 3rd Generation Partnership Project (3GPP) [1]. A user equipment (UE) can perform interference cancellation under the assumption that the interference parameters including traffic-to-pilot power ratio (TPR), rank indicator (RI), precoding matrix indicator (PMI) and modulation level (MOD) are known to the UE through the aid of network signaling. This assumption, however, not only leads to an excessive increase in the network-signaling load but also limits the scheduling performance. This paper presents interference cancellation techniques based on blindly-detected NAICS interference parameters. We first investigate blind detection methods for estimating a TPR being used by interfering base station (BS) and then propose low-complexity blind detection (BD) algorithms based on max-log approximation for estimating the interference parameters RI, PMI, and MOD. In particular, we introduce a compensation method for correcting the error imposed by the suboptimal approximation. The link-level simulation results are provided to verify the efficacy of the proposed interference detection algorithms and interference cancellation techniques.

Advanced co-channel interference aware signal detection has drawn research attention during the recent development of Long Term Evolution-Advanced (LTE-A) systems and the interference-aware communications (IAC) is currently being studied by 3GPP. This paper investigates link performance abstraction for the IAC systems employing maximum- likelihood detector (MLD). The link performance of MLD can be estimated by combining two performance bounds, namely, linear receiver and genie-aided maximum-likelihood (ML) receiver. We propose a new adaptive approach where the combining parameter is adaptively adjusted according to instantaneous interference-to-signal ratio (ISR). The basic idea is to exploit the probabilistic behavior of the optimal combining ratio over the ISR. The link-level simulation results are provided to verify the prediction accuracy of the proposed link abstraction method.

An important challenge of wireless networks is to boost the cell edge performance and enable multi-stream transmissions to cell edge users. Interference mitigation techniques relying on multiple antennas and coordination among cells are nowadays heavily studied in the literature. Typical strategies in OFDMA networks include coordinated scheduling, beamforming and power control. In this paper, we propose a novel and practical type of coordination for OFDMA downlink networks relying on multiple antennas at the transmitter and the receiver. The transmission ranks, i.e. the number of transmitted streams, and the user scheduling in all cells are jointly optimized in order to maximize a network utility function accounting for fairness among users. A distributed coordinated scheduler motivated by an interference pricing mechanism and relying on a masterslave architecture is introduced. The proposed scheme is operated based on the user report of a recommended rank for the interfering cells accounting for the receiver interference suppression capability. It incurs a very low feedback and backhaul overhead and enables efficient link adaptation. It is moreover robust to channel measurement errors and applicable to both open-loop and closed-loop MIMO operations. A 20% cell edge performance gain over uncoordinated LTE-A system is shown through system level simulations.

An important challenge of wireless networks is to boost the cell edge performance and enable multi-streams transmissions to cell edge users. Interference mitigation techniques relying on multiple antennas and coordination among cells are nowadays heavily studied in the literature. In this paper, we propose a novel type of coordination for OFDMA networks relying on multiple antennas at the transmitter and the receiver where the transmission ranks, i.e. the number of transmitted streams, and the user scheduling in all cells are jointly optimized in order to maximize a network utility function accounting for fairness among users. A distributed coordinated scheduler relying on an interference pricing mechanism and a master-slave architecture is introduced. The proposed scheme is operated based on the user report of a recommended rank for the interfering cells. It incurs a very low feedback overhead and enables an efficient link adaptation. It is moreover robust to practical system impairments and is applicable to both open-loop and closed-loop MIMO operations. A cell edge performance gain of 20% over uncoordinated LTE-Advanced is shown through system level simulations.

This letter proposes a new approach to achieve the maximum diversity gain in the vertical Bell Laboratories Layered Space-Time (V-BLAST) system over multiple-input multiple-output (MIMO) Rayleigh channels. First, we will briefly review the modulation diversity scheme based on real-valued multidimensional rotated constellations that has been proposed to obtain high diversity orders for single-input single-output (SISO) systems. Then, this letter extends the modulation diversity scheme to V-BLAST MIMO systems in order to achieve the maximum diversity gain without additional power or bandwidth consumption. The proposed method is particularly suitable for MIMO applications which employ channel codes with high rate in order to support high spectral efficiency.

In this letter, we propose a new groupwise receive combiner design for multiple-input multiple-output spatial multiplexing systems. The conventional group detection (GD) suffers from a considerable performance loss since the noise components are not taken into account. The output signal-to-interference-plus-noise ratio (SINR) is defined in each subgroup in order to consider both the desired signal and noise statistics. Adopting the real-valued representation, we provide an optimal receive combiner which maximizes the SINR with a general group size. The simulation results show that the proposed scheme achieves a large performance gain over the conventional GD in coded systems. Also, when combining with near-optimal detection algorithms such as sphere decoder, the proposed GD scheme offers a comparable performance with significant reduced complexity.

This paper proposes a new preprocessing scheme for rate-one quasi-orthogonal space-time block code (QO-STBC), which requires no channel state information at the transmitter. We consider QO-STBCs in real domain so as to utilize the real-valued multidimensional constellation. First, we derive a preceding matrix which makes the effective channel matrix of QO-STBC column-wise orthogonal. Surprisingly, the proposed scheme provides the channel gains identical to the singular values of the channel matrix of STBCs. Since singular value decomposition based preceding method provides better performance than other conventional maximum-likelihood decoding method in coded systems, the proposed scheme outperforms other STBC schemes in coded systems without feedback information, while reducing the decoding complexity dramatically. Second, we employ the rotated multidimensional constellation in order to achieve full diversity in uncoded systems. Unlike the previous works where heuristic approaches should be needed to obtain solutions in QO-STBC systems with more than four transmit antennas, the proposed precoding combined with the rotated multidimensional constellation provides an optimal solution based on analysis regardless of the number of transmit antennas.

In this letter, we present a new maximum likelihood (ML) decoding algorithm for space time block codes (STBCs) that employ multidimensional constellations. We start with a lattice representation for STBCs which transforms complex channel models into real matrix equations. Based on the lattice representation, we propose a new decoding algorithm for quasiorthogonal STBCs (QO-STBC) which allows simpleML decoding with performance identical to the conventional ML decoder. Multidimensional rotated constellations are constructed for the QO-STBCs to achieve full diversity. As a consequence, for quasi-orthogonal designs with an arbitrary number of transmit antennas N (N ≤ 4), the proposed decoding scheme achieves full rate and full diversity while reducing the decoding complexity from 𝒪(McN/2) to 𝒪(McN/4) in a Mc-QAM constellation.

In this paper, we present a new precoding technique using rotation transformations for closed loop multiple-input multiple-output (MIMO) wireless systems, which does not require the singular value decomposition (SVD) operation of the channel transfer matrix and allows a simple maximum-likelihood (ML) decoding at the receiver. We divide the precoding process into two steps: orthogonalization transformation which induces orthogonality between transmitted signals and beamforming transformation which achieves diversity gain. In the proposed method, we utilize a design criterion based on the minimum Euclidean distance between the received signals and then the vector orthogonalization is connected to the vector-norm maximization. In this paper, we focus on spatial multiplexing systems transmitting two independent data streams. Compared with the SVD based schemes, the proposed approach maintains a low complexity by relying only on three different kinds of rotation matrices for both the orthogonalization and beamforming transformation. Simulation results confirm that the proposed two step precoding achieves the better performance than the conventional SVD based MIMO precodings with reduced complexity.

This paper proposes an efficient beamforming scheme which attains optimality as singular value decomposition (SVD) based systems with low complexity utilizing transmit maximum-ratio combining (TMRC) techniques. The TMRC scheme is the optimum structure for single beamforming systems in terms of received signal-to-noise ratio (SNR) in multiple-input single-output (MISO) channels. In this paper, we generalize the TMRC scheme to multiple beamforming multiple-input multiple-output (MIMO) systems which support more than one data stream in coded systems. We express each beamforming vector as a linear combination of TMRC vectors whose coefficients are optimized in a successive manner. Optimization of the beamforming vector is followed by the decorrelation process. All TMRC vectors used as a basis of the remaining beamforming vectors are made orthogonal to previously computed beamforming vectors. Exploiting the concept of the gradient ascent algorithm, we propose a simple non-iterative method of computing the precoder which obtains the near optimal performance. Also we derive a closed form expression of the output SNR distribution for the proposed scheme. Simulation results demonstrate that the proposed scheme achieves the almost identical link performance as the SVD-based system for arbitrary configurations with reduced complexity.

In this paper, we present a low-complexity method to generate a transmit beamforming vector for multiple-input-multiple-output (MIMO) systems. We begin by introducing new definitions regarding orthogonality between two complex valued vectors and then present new expressions of complex rotation matrices for the complex vector orthogonalization. The rotation matrices are utilized to derive the weight vector for the maximum-norm combining (MNC) process of two complex vectors, which provides a constructive basis for a new beamforming method. The proposed transmit beamforming method uses successive column combining of MIMO channel matrices based on MNC, and as a result, an approximate solution to the optimum beamforming vector is obtained. The proposed method offers a good tradeoff between complexity and performance. Simulation results demonstrate that the proposed beamforming method achieves the near-optimal performance with much reduced computational complexity, compared to the optimal beamforming scheme using singular-value decomposition (SVD) of the channel matrix.

This paper proposes a new precoding algorithm for orthogonalized spatial multiplexing (OSM) systems over flat-fading multiple-input multiple-output (MIMO) channels. The OSM scheme was recently introduced for closed-loop MIMO systems which allows single symbol decodability for maximum likelihood detection. To further improve the performance in OSM systems, we propose a new precoding method. For identifying the parameters of a precoder, we introduce a partitioning approach on the minimum Euclidean distance between constellation points in the effective channel. Also, it is shown that two real value parameters and one bit are required for feedback information in 4-QAM systems. Simulation results demonstrate that our precoding algorithm allows us to significantly improve the system performance with small increase of feedback values. We also confirm through simulations that the performance of the proposed scheme is the same as the optimum closed-loop MIMO systems.

Channel inversion is one of the simplest techniques for multiuser downlink systems with single-antenna users. In this paper, we extend the regularized channel inversion technique developed for the single-antenna user case to multiuser multiple- input multiple-output (MIMO) channels with multiple-antenna users. We first employ the multiuser preprocessing to project the multiuser signals near the null space of the unintended users based on the MMSE criterion, and then the single-user preprocessing is applied to the decomposed MIMO interference channels. In order to reduce the complexity, we focus on non- iterative solutions for the multiuser transmit beamforming and use a linear receiver based on an MMSE criterion. Simulation results show that the proposed scheme outperforms existing joint iterative algorithms in most multiuser configurations.

This paper introduces a new efficient design scheme for spatial multiplexing (SM) systems over closed loop multiple- input multiple-output (MIMO) wireless channels. Extending the orthogonalized spatial multiplexing (OSM) scheme which was developed recently for transmitting two data streams, we propose a new SM scheme where a larger number of data streams can be supported. To achieve this goal, we partition the data streams into several subblocks and execute the block-diagonalization process at the receiver. The proposed scheme still guarantees single- symbol maximum likelihood (ML) detection with small feedback information. Simulation results verify that the proposed scheme achieves a huge performance gain at a bit error rate (BER) of 10-4over conventional closed-loop schemes based on minimum mean-square error (MSE) or bit error rate (BER) criterion. We also show that an additional 2.5 dB gain can be obtained by optimizing the group selection with extra feedback information.

This paper proposes a low complexity unitary precoding scheme for multiple-input multiple-output (MIMO) systems. The singular-value decomposition (SVD) based transmission is capable of maximizing the system throughput when combined with power allocation and bit loading, and is known to be optimum in terms of capacity. This paper focuses on a system which attains the same optimality as the SVD-based system with low complexity utilizing transmit maximum-ratio combining (TMRC) techniques. The TMRC scheme is the optimum structure for single beamforming systems in terms of received signal-to-noise ratio (SNR) in multiple-input single-output (MISO) channels. In this paper, we generalize the TMRC scheme to multiple beamforming MIMO systems which supports more than one data stream in coded systems. Simulation results demonstrate that the proposed scheme achieves the almost identical link performance as the SVD precoding system for arbitrary configurations with reduced complexity.

In this paper, we propose a new groupwise receiver combiner design method for spatial multiplexing multiple- input multiple-output (MIMO) systems. The conventional group detection (GD) suffers from a considerable performance loss since the noise components are not taken into account. The proposed method further enhances the detection performance than the conventional GD by considering the noise correlation. To prevent the noise enhancement problem, the proposed combiner design utilizes a signal-to-interference-plus-noise ratio (SINR) maximization criterion for each desired subgroup. In addition, we investigate an effect of group selection based on the received SINR. The simulation results show that the proposed scheme achieves a performance gain of 4 dB at a frame error rate (FER) of 10-2over the conventional GD with four transmit and four receive antennas.

In this paper, we propose a two-step precoding approach for multiple-input multiple-output (MIMO) systems by dividing the precoding process into two steps: spatial multiplexing transformation which establishes orthogonalized subchannels with the same quality and beamforming transformation which achieves diversity gain. In the proposed precoding method, we improve the system performance by maximizing the minimum Euclidean distance between the received signals. Compared with the optimal singular value decomposition based precoding methods, the proposed approach significantly reduces the processing complexity as well as the feedback overhead since precoding techniques are based on simple rotation transformations.

In this paper, we propose a new precoding algorithm for orthogonalized spatial multiplexing (OSM) systems over flat-fading multiple-input multiple-output (MIMO) channels. The OSM scheme was recently introduced for closed-loop MIMO systems which allows single symbol decodable maximum likelihood detection. To further improve the performance of the OSM system, we propose a new precoding method by maximizing the minimum Euclidean distance between constellation points in the effective channel. In order to efficiently identify the parameters of a precoder which maximizes the minimum distance, we introduce a partitioning approach. Through analysis, it is shown that one real value parameter and two bits are required for feedback information for precoding in 16-QAM systems. Simulation results demonstrate that our algorithm provides 9 dB and 7.5 dB gains at a bit error rate (BER) of 10-4over the conventional OSM systems for 4-QAM and 16-QAM, respectively. We also confirm that the performance of the proposed scheme is the same as that of the optimum closed-loop MIMO systems in terms of the minimum distance. Consequently, our precoding algorithm significantly improves the system performance with a small increase of feedback amount.

In this paper, we present a new transmission technique for the downlink of a multiuser multiple-input multiple- output (MIMO) system. Recently, the block-diagonalization (BD) algorithm has been proposed for the multiuser MIMO downlink where each user has multiple antennas. The key idea of the BD scheme is to eliminate all multiuser interference by transmitting each user's data along the nullspace of the other users' channel matrix. However, the BD algorithm based on the zero- interference condition is not a proper choice when the interference is known at the transmitter noncausally. In this paper, we propose a block-echelonization (BE) algorithm for utilizing the noncausally known interference at the transmitter. The sum rate analysis and simulation results show the effectiveness of the proposed multiuser scheme.

We present a new hybrid scheme which combines space-time block codes (STBC) and spatial multiplexing (SM) to achieve both diversity and spatial multiplexing gain. We design a new precoding scheme based on the givens rotation by utilizing the phase feedback information from the receiver. Exploiting the STBC structure embedded in the hybrid scheme, the proposed scheme enables a simple maximum likelihood (ML) detection such that the receiver complexity is reduced by a factor of 1/16 for the case of 4bps/Hz. Practical precoder designs are also provided to minimize the feedback information, while maintaining near optimal performance. Also, unlike conventional schemes which require at least two receive antennas, the proposed hybrid scheme works well with a single receive antenna. Simulation results show that the proposed hybrid scheme achieves a performance gain of 7dB at a bit error rate (BER) of 10-4in 2 bps/Hz spectral efficiency over other closed- loop hybrid schemes.

We propose a new transmit beam- forming technique for multiple-input-multiple-output (MIMO) systems to improve the link level performance. We present a method for orthogonalization of two complex-valued vectors by introducing a variation of the Jacobi rotations. We will show that based on the proposed rotation transformations, the orthogonality can be established among different complex-valued column vectors in the channel response matrix. Utilizing the orthogonality, we can achieve the channel gain comparable to the maximum singular value of the channel matrix. Simulation results demonstrate that the proposed beamforming scheme achieves the near-optimum performance with much reduced complexity and feedback overhead. Especially, for the two transmit antenna case, we show that the proposed beamforming scheme provides the optimal beamforming vector for MIMO systems.

In this paper, we propose a new transmit diversity technique for multiple-input multiple-output (MIMO) systems to improve the link level performance of open-loop systems over flat fading channels. By cyclically applying a predetermined set of precoding weight vectors, artificially induced fluctuation is created to achieve additional diversity gain in flat fading channels. To design the set of the precoding vectors, we exploit the knowledge on the distribution of near optimum precoding vectors observed in a beamforming scheme based on the rotation transformations. Simulation results demonstrate that the proposed open-loop diversity scheme with an arbitrary number of transmit antennas achieves a full diversity gain with computational complexity comparable to a single-input single-output (SISO) system.

This paper proposes a simple downlink transmit beamforming scheme for multiple-input multiple-output systems with low complexity. We first introduce the beamforming system where the base station (BS) utilizes only a fraction of the full channel response. The mobile station (MS) of the proposed system selects a single antenna at the receive side based on which the BS adjusts beamforming weights without any overhead to the structure of the BS. Without requiring the singular value decomposition operation, the proposed scheme achieves a performance very close to that of the optimum beamforming system. An exact analysis of the received signal-to-noise ratio for the proposed scheme with 2 receive antennas is also presented. Simulation results show that with much reduced complexity the performance of the proposed beamforming technique is only a few tenth of a dB away from the optimal beamforming system regardless of antenna configuration.

In this paper, we investigate coded layered space-time architectures for frequency-selective fading multiple-input multiple-output orthogonal frequency-division multiplexing (OFDM) channels. By computing outage capacity formulas, we will show that the capacity of the vertical Bell Labs layered space-time (V-BLAST) architecture can closely approach the Shannon capacity in the frequency-selective OFDM environment. Motivated by the capacity analysis, we propose pragmatic approaches which preserve the optimality of the layered space-time concept. We present methods to prevent the error propagation from catastrophically affecting the signal detection in subsequent layers. First, we start with a comprehensive signal modeling which includes error propagation. We derive an improved signal detector and describe the optimal soft-bit log-likelihood ratio value-computation method by taking decision errors into account for soft-input channel decoding. Then, to further enhance the V-BLAST performance, we show that cancellation using decoded decisions from previous layers makes the decision errors almost completely disappear, so that the layered space-time architecture can approach the attainable channel capacity. Finally, simulations confirm that the proposed schemes show a significant performance improvement over the conventional methods

In this paper, we propose a novel spatial multiplexing (SM) scheme for transmission over flat-fading multiple-input multiple-output (MIMO) channels, which allows simple maximum-likelihood decoding at the receiver with small feedback information. We begin with a real-valued representation of the complex-valued system model, and show that we can achieve orthogonality between transmitted signals by applying a proper rotation to transmitted symbols. Compared with other closed-loop methods, the proposed scheme significantly reduces the processing complexity at both transmitter and receiver as well as the feedback overhead. We also combine the proposed orthogonalization method with antenna-selection techniques, and then discuss a criterion based on the minimum Euclidean distance between received vectors for selecting the optimal subset of multiple transmit antennas. Using geometrical properties of the receive constellations, we also present a simple antenna-selection metric for the proposed SM systems. Simulation results demonstrate that our SM system performs close to the optimum closed-loop system with much-reduced complexity

In this paper, we present a novel maximum likelihood (ML) decoding algorithm for space-time block codes (STBC) over fading channels. Using a lattice representation for space-time codes by transforming complex channel models into real matrix equations, we propose a new efficient ML decoding algorithm with performance identical to the conventional ML decoder. We show that the complex orthogonal space-time codes, in fact, allow a separate ML decoding for each inphase and quadrature-phase component. For rate one quasiorthogonal designs with N transmit antennas (N > 4), the proposed decoding scheme reduces the decoding complexity from O(McN/2) to O(McN/4) in a Mc-QAM constellation. Moreover, multidimensional rotated constellations are constructed by which the quasi-orthogonal codes can achieve full diversity, while the proposed ML decoding method offers significant computational savings as compared with previous approaches.

This paper proposes an improved vertical Bell Labs Layered Space-Time (V-BLAST) with iterative detection and decoding (IDD) scheme for coded layered space-time architectures in MIMO-OFDM systems. For the iterative process, a low-complexity demapper is developed by making use of both nonlinear interference cancellation and linear filtering. Also a simple cancellation method based on hard decision is presented to reduce the overall complexity. Simulation results demonstrate that the proposed V-BLAST with IDD scheme offers the performance close to the optimal turbo-MIMO approach, while providing tremendous savings in computational complexity.

In this paper, we present an asymptotical analysis of channel capacity of Bell labs layered space-time (BLAST) architectures based on a zero-forcing(ZF) criterion in the sense of signal-to-noise ratio (SNR). We begin by introducing a new relationship related to multi-input multi-output (MIMO) channel capacity. We prove that Diagonal Bell Labs Space-Time (D-BLAST) attains the lower bound for MIMO channels when interference nulling is carried out based on the ZF-criterion. An exact closed-form expression for the probability density function of the channel capacity is analyzed. Based on the asymptotic behavior of the channel capacity of each layer, closed-form expressions for the asymptotic ergodic capacity are derived for BLAST. Based on the analysis presented in this paper, we gain an insight on the channel capacity behavior for a MIMO channel. Computer simulation results have verified the validity and accuracy of the proposed analysis for a wide range of antenna array sizes.

In this paper, we propose a new spatial multiplexing scheme for transmission over flat-fading multiple-input multiple-output (MIMO) channels, which allows a simple maximum-likelihood decoding at the receiver with small feedback information. We begin with a real-valued representation of the complex-valued system model and show that we can achieve orthogonality between transmitted signals by applying a proper rotation to transmitted symbols. Based on the minimum Euclidean distance between received vectors, we also present a simple antenna selection metric for the proposed spatial multiplexing systems. Simulation results demonstrate that our spatial multiplexing system performs close to the optimum closed loop system with much reduced complexity and feedback overhead.

Multiple-input-multiple-output (MIMO) systems provide a very promising means to increase the spectral efficiency for wireless systems. By using orthogonal frequency-division multiplexing (OFDM), wideband transmission can be achieved over frequency-selective fading radio channels. First, in this paper, we introduce an improved vertical Bell Labs layered space-time (V-BLAST) receiver which takes the decision errors into account. Second, we propose an iterative detection and decoding (IDD) scheme for coded layered space-time architectures in MIMO-OFDM systems. For the iterative process, a low-complexity demapper is developed by making use of both nonlinear interference cancellation and linear minimum mean-square error filtering. Also, a simple cancellation method based on hard decision is presented to reduce the overall complexity. Simulation results demonstrate that the proposed IDD scheme combined with the improved V-BLAST performs almost as well as the optimal turbo-MIMO approach, while providing tremendous savings in computational complexity.

In this paper, we investigate coded layered space-time architectures for frequency-selective fading MIMO-OFDM channels. We start with a comprehensive signal model including error propagations. Based on the signal model, we derive an improved signal detector in the mean-square-error criterion. We also describe the optimal soft bit log-likelihood ratio (LLR) value computation method by taking the decision errors into account for soft-input Viterbi decoder. We compare the proposed scheme with the conventional V-BLAST system and show that significant performance improvement at high data rates can be attained.