arXiv subjects

Giuseppe Caire

Publications and source records attributed to Giuseppe Caire.

At least 433 records · Page 24Linked to original sources

arXiv2011

Multi-User MIMO with outdated CSI: Training, Feedback and Scheduling

Conventional MU-MIMO techniques, e.g. Linear Zero-Forced Beamforming (LZFB), require sufficiently accurate channel state information at the transmitter (CSIT) in order to realize spectral efficient transmission (degree of freedom gains). In practical settings, however, CSIT accuracy can be limited by a number of issues including CSI estimation, CSI feedback delay between user terminals to base stations, and the time/frequency coherence of the channel. The latter aspects of CSIT-feedback delay and channel-dynamics can lead to significant challenges in the deployment of efficient MU-MIMO systems. Recently it has been shown by Maddah-Ali and Tse (MAT) that degree of freedom gains can be realized by MU-MIMO even when the knowledge of CSIT is completely outdated. Specifically, outdated CSIT, albeit perfect CSIT, is known for transmissions only after they have taken place. This aspect of insensitivity to CSIT-feedback delay is of particular interest since it allows one to reconsider MU-MIMO design in dynamic channel conditions. Indeed, as we show, with appropriate scheduling, and even in the context of CSI estimation and feedback errors, the proposed MAT scheme can have performance advantages over conventional MU-MIMO in such scenarios.

cs.IT↗

arXiv2011

Achieving "Massive MIMO" Spectral Efficiency with a Not-so-Large Number of Antennas

The main focus and contribution of this paper is a novel network-MIMO TDD architecture that achieves spectral efficiencies comparable with "Massive MIMO", with one order of magnitude fewer antennas per active user per cell. The proposed architecture is based on a family of network-MIMO schemes defined by small clusters of cooperating base stations, zero-forcing multiuser MIMO precoding with suitable inter-cluster interference constraints, uplink pilot signals reuse across cells, and frequency reuse. The key idea consists of partitioning the users population into geographically determined "bins", such that all users in the same bin are statistically equivalent, and use the optimal network-MIMO architecture in the family for each bin. A scheduler takes care of serving the different bins on the time-frequency slots, in order to maximize a desired network utility function that captures some desired notion of fairness. This results in a mixed-mode network-MIMO architecture, where different schemes, each of which is optimized for the served user bin, are multiplexed in time-frequency. In order to carry out the performance analysis and the optimization of the proposed architecture in a clean and computationally efficient way, we consider the large-system regime where the number of users, the number of antennas, and the channel coherence block length go to infinity with fixed ratios. The performance predicted by the large-system asymptotic analysis matches very well the finite-dimensional simulations. Overall, the system spectral efficiency obtained by the proposed architecture is similar to that achieved by "Massive MIMO", with a 10-fold reduction in the number of antennas at the base stations (roughly, from 500 to 50 antennas).

cs.IT↗

arXiv2011

Network MIMO with Linear Zero-Forcing Beamforming: Large System Analysis, Impact of Channel Estimation and Reduced-Complexity Scheduling

We consider the downlink of a multi-cell system with multi-antenna base stations and single-antenna user terminals, arbitrary base station cooperation clusters, distance-dependent propagation pathloss, and general "fairness" requirements. Base stations in the same cooperation cluster employ joint transmission with linear zero-forcing beamforming, subject to sum or per-base station power constraints. Inter-cluster interference is treated as noise at the user terminals. Analytic expressions for the system spectral efficiency are found in the large-system limit where both the numbers of users and antennas per base station tend to infinity with a given ratio. In particular, for the per-base station power constraint, we find new results in random matrix theory, yielding the squared Frobenius norm of submatrices of the Moore-Penrose pseudo-inverse for the structured non-i.i.d. channel matrix resulting from the cooperation cluster, user distribution, and path-loss coefficients. The analysis is extended to the case of non-ideal Channel State Information at the Transmitters (CSIT) obtained through explicit downlink channel training and uplink feedback. Specifically, our results illuminate the trade-off between the benefit of a larger number of cooperating antennas and the cost of estimating higher-dimensional channel vectors. Furthermore, our analysis leads to a new simplified downlink scheduling scheme that pre-selects the users according to probabilities obtained from the large-system results, depending on the desired fairness criterion. The proposed scheme performs close to the optimal (finite-dimensional) opportunistic user selection while requiring significantly less channel state feedback, since only a small fraction of pre-selected users must feed back their channel state information.

cs.IT↗

arXiv2011

Quantized Compute and Forward: A Low-Complexity Architecture for Distributed Antenna Systems

We consider a low-complexity version of the Compute and Forward scheme that involves only scaling, offset (dithering removal) and scalar quantization at the relays. The proposed scheme is suited for the uplink of a distributed antenna system where the antenna elements must be very simple and are connected to a oint processor via orthogonal perfect links of given rate R0. We consider the design of non-binary LDPC codes naturally matched to the proposed scheme. Each antenna element performs individual (decentralized) Belief Propagation decoding of its own quantized signal, and sends a linear combination of the users' information messages via the noiseless link to the joint processor, which retrieves the users' messages by Gaussian elimination. The complexity of this scheme is linear in the coding block length and polynomial in the system size (number of relays).

cs.IT↗

arXiv2011

A Nonlinear Approach to Interference Alignment

Cadambe and Jafar (CJ) alignment strategy for the K-user scalar frequency-selective fading Gaussian channel, with encoding over blocks of 2n+1 random channel coefficients (subcarriers) is considered. The linear zero-forcing (LZF) strategy is compared with a novel approach based on lattice alignment and lattice decoding (LD). Despite both LZF and LD achieve the same degrees of freedom, it is shown that LD can achieve very significant improvements in terms of error rates at practical SNRs with respect to the conventional LZF proposed in the literature. We also show that these gains are realized provided that channel gains are controlled to be near constant, for example, by means of power control and opportunistic carrier and user selection strategies. In presence of relatively-small variations in the normalized channel coefficient amplitudes, CJ alignment strategy yields very disappointing results at finite SNRs, and the gain of LD over ZLF significantly reduces. In light of these results, the practical applicability of CJ alignment scheme remains questionable, in particular for Rayleigh fading channels, where channel inversion power control yields to unbounded average transmit power.

cs.IT↗

arXiv2011

On Achievability of Gaussian Interference Channel Capacity to within One Bit

In the earlier version of this paper, it was wrongly claimed that time-sharing is required to achieve the capacity region of the Gaussian interference channel to within one bit, especially at corner points. The flaw in the argument of the earlier version lies in fixing the decoding paradigm for a fixed common/private message splitting encoding strategy. More specifically, the additional constraints (7b) and (7d) in the earlier version arise if we force the common messages to be always decoded at unintended receivers. However, (7b) and (7d) can be eliminated by allowing the decoders to ignore unintended common messages, particularly at corner points of the rate region, without resorting to time-sharing at the transmit side, as suggested in the earlier version. For these reasons, our earlier claim is invalid.

cs.IT↗

arXiv2010

Relay Strategies Based on Cross-Determinism for the Broadcast Relay Channel

We consider a two-user Gaussian multiple-input multiple-output (MIMO) broadcast channel with a common multiple-antenna relay, and a shared digital (noiseless) link between the relay and the two destinations. For this channel, this paper introduces an asymptotically sum-capacity-achieving quantize-and-forward (QF) relay strategy. Our technique to design an asymptotically optimal relay quantizer is based on identifying a cross-deterministic relation between the relay observation, the source signal, and the destination observation. In a relay channel, an approximate cross deterministic relation corresponds to an approximately deterministic relation, where the relay observation is to some extent a deterministic function of the source and destination signals. We show that cross determinism can serve as a measure for quantization penalty. By identifying an analogy between a deterministic broadcast relay channel and a Gaussian MIMO relay channel, we propose a three-stage dirty paper coding strategy, along with receiver beamforming and quantization at the relay, to asymptotically achieve an extended achievable rate region for the MIMO broadcast channel with a common multiple-antenna relay.

cs.IT↗

arXiv2010

Coarse Network Coding: A Simple Relay Strategy to Resolve Interference

Reminiscent of the parity function in network coding for the butterfly network, it is shown that forwarding an even/odd indicator bit for a scalar quantization of a relay observation recovers 1 bit of information at the two destinations in a noiseless interference channel where interference is treated as noise. Based on this observation, a coding strategy is proposed to improve the rate of both users at the same time using a relay node in an interference channel. In this strategy, the relay observes a linear combination of the two source signals, and broadcasts a common message to the two destinations over a shared out-of-band link of rate R0 bits per channel use. The relay message consists of the bin index of a structured binning scheme obtained from a 2^R0-way partition of the squared lattice in the complex plane. We show that such scalar quantization-binning relay strategy asymptotically achieves the cut-set bound in an interference channel with a common out-of-band relay link of limited rate, improving the sum rate by two bits for every bit relayed, asymptotically at high signal to noise ratios (SNR) and when interference is treated as noise. We then use low-density parity-check (LDPC) codes along with bit-interleaved coded-modulation (BICM) as a practical coding scheme for the proposed strategy. We consider matched and mismatched scenarios, depending on whether the input alphabet of the interference signal is known or unknown to the decoder, respectively. For the matched scenario, we show the proposed strategy results in significant gains in SNR. For the mismatched scenario, we show that the proposed strategy results in rate improvements that, without the relay, cannot be achieved by merely increasing transmit powers. Finally, we use generalized mutual information analysis to characterize the theoretical performance of the mismatched scenario and validate our simulation results.

cs.IT↗

arXiv2010

Multi-Cell MIMO Downlink with Cell Cooperation and Fair Scheduling: a Large-System Limit Analysis

We consider the downlink of a cellular network with multiple cells and multi-antenna base stations, including a realistic distance-dependent pathloss model, clusters of cooperating cells, and general "fairness" requirements. Beyond Monte Carlo simulation, no efficient computation method to evaluate the ergodic throughput of such systems has been presented so far. We propose an analytic solution based on the combination of large random matrix results and convex optimization. The proposed method is computationally much more efficient than Monte Carlo simulation and provides surprisingly accurate approximations for the actual finite-dimensional systems, even for a small number of users and base station antennas. Numerical examples include 2-cell linear and three-sectored 7-cell planar layouts, with no inter-cell cooperation, sector cooperation, or full inter-cell cooperation.

cs.IT↗

arXiv2010

Multi-cell MIMO Downlink with Fairness Criteria: the Large System Limit

We consider the downlink of a cellular network with multiple cells and multi-antenna base stations including arbitrary inter-cell cooperation, realistic distance-dependent pathloss and general "fairness" requirements. Beyond Monte Carlo simulation, no efficient computation method to evaluate the ergodic throughput of such systems has been provided so far. We propose an analytic method based on the combination of the large random matrix theory with Lagrangian optimization. The proposed method is computationally much more efficient than Monte Carlo simulation and provides a very accurate approximation (almost indistinguishable) for the actual finite-dimensional case, even for of a small number of users and base station antennas. Numerical examples include linear 2-cell and planar three-sectored 7-cell layouts, with no inter-cell cooperation, sector cooperation, and full cooperation.

cs.IT↗

arXiv2010

Channel State Feedback over the MIMO-MAC

We consider the problem of designing low latency and low complexity schemes for channel state feedback over the MIMO-MAC (multiple-input multiple-output multiple access channel). We develop a framework for analyzing this problem in terms of minimizing the MSE distortion, and come up with separated source-channel schemes and joint source-channel schemes that perform better than analog feedback. We also develop a strikingly simple code design based on scalar quantization and uncoded QAM modulation that achieves the theoretical asymptotic performance limit of the separated approach with very low complexity and latency, in the case of single-antenna users.

cs.IT↗

arXiv2010

Multiuser MISO Transmitter Optimization for Inter-Cell Interference Mitigation

The transmitter optimization (i.e., steering vectors and power allocation) for a MISO Broadcast Channel (MISO-BC) subject to general linear constraints is considered. Such constraints include, as special cases, the sum power, the per-antenna or per-group-of-antennas power, and "forbidden interference direction" constraints. We consider both the optimal dirty-paper coding and the simple suboptimal linear zero-forcing beamforming strategies, and provide numerically efficient algorithms that solve the problem in its most general form. As an application, we consider a multi-cell scenario with partial cell cooperation, where each cell optimizes its precoder by taking into account interference constraints on specific users in adjacent cells. The effectiveness of the proposed methods is evaluated in a simple system scenario including two adjacent cells, under different fairness criteria that emphasize the bottleneck role of users near the cell "boundary". Our results show that "active" Inter-Cell Interference (ICI) mitigation outperforms the conventional "static" ICI mitigation based on fractional frequency reuse.

cs.IT↗

arXiv2010

Linear Finite-Field Deterministic Networks With Many Sources and One Destination

We find the capacity region of linear finite-field deterministic networks with many sources and one destination. Nodes in the network are subject to interference and broadcast constraints, specified by the linear finite-field deterministic model. Each node can inject its own information as well as relay other nodes' information. We show that the capacity region coincides with the cut-set region. Also, for a specific case of correlated sources we provide necessary and sufficient conditions for the sources transmissibility. Given the "deterministic model" approximation for the corresponding Gaussian network model, our results may be relevant to wireless sensor networks where the sensing nodes multiplex the relayed data from the other nodes with their own data, and where the goal is to decode all data at a single "collector" node.

cs.IT↗

arXiv2010

Joint Scheduling and ARQ for MU-MIMO Downlink in the Presence of Inter-Cell Interference

User scheduling and multiuser multi-antenna (MU-MIMO) transmission are at the core of high rate data-oriented downlink schemes of the next-generation of cellular systems (e.g., LTE-Advanced). Scheduling selects groups of users according to their channels vector directions and SINR levels. However, when scheduling is applied independently in each cell, the inter-cell interference (ICI) power at each user receiver is not known in advance since it changes at each new scheduling slot depending on the scheduling decisions of all interfering base stations. In order to cope with this uncertainty, we consider the joint operation of scheduling, MU-MIMO beamforming and Automatic Repeat reQuest (ARQ). We develop a game-theoretic framework for this problem and build on stochastic optimization techniques in order to find optimal scheduling and ARQ schemes. Particularizing our framework to the case of "outage service rates", we obtain a scheme based on adaptive variable-rate coding at the physical layer, combined with ARQ at the Logical Link Control (ARQ-LLC). Then, we present a novel scheme based on incremental redundancy Hybrid ARQ (HARQ) that is able to achieve a throughput performance arbitrarily close to the "genie-aided service rates", with no need for a genie that provides non-causally the ICI power levels. The novel HARQ scheme is both easier to implement and superior in performance with respect to the conventional combination of adaptive variable-rate coding and ARQ-LLC.

cs.IT↗

arXiv2009

Training and Feedback Optimization for Multiuser MIMO Downlink

We consider a MIMO fading broadcast channel where the fading channel coefficients are constant over time-frequency blocks that span a coherent time $\times$ a coherence bandwidth. In closed-loop systems, channel state information at transmitter (CSIT) is acquired by the downlink training sent by the base station and an explicit feedback from each user terminal. In open-loop systems, CSIT is obtained by exploiting uplink training and channel reciprocity. We use a tight closed-form lower bound on the ergodic achievable rate in the presence of CSIT errors in order to optimize the overall system throughput, by taking explicitly into account the overhead due to channel estimation and channel state feedback. Based on three time-frequency block models inspired by actual systems, we provide some useful guidelines for the overall system optimization. In particular, digital (quantized) feedback is found to offer a substantial advantage over analog (unquantized) feedback.

cs.IT↗

arXiv2009

MIMO Downlink Scheduling with Non-Perfect Channel State Knowledge

Downlink scheduling schemes are well-known and widely investigated under the assumption that the channel state is perfectly known to the scheduler. In the multiuser MIMO (broadcast) case, downlink scheduling in the presence of non-perfect channel state information (CSI) is only scantly treated. In this paper we provide a general framework that addresses the problem systematically. Also, we illuminate the key role played by the channel state prediction error: our scheme treats in a fundamentally different way users with small channel prediction error ("predictable" users) and users with large channel prediction error ("non-predictable" users), and can be interpreted as a near-optimal opportunistic time-sharing strategy between MIMO downlink beamforming to predictable users and space-time coding to nonpredictable users. Our results, based on a realistic MIMO channel model used in 3GPP standardization, show that the proposed algorithms can significantly outperform a conventional "mismatched" scheduling scheme that treats the available CSI as if it was perfect.

cs.IT↗

arXiv2009

MIMO Broadcast Channel Optimization under General Linear Constraints

The optimization of the transmit parameters (power allocation and steering vectors) for the MIMO BC under general linear constraints is treated under the optimal DPC coding strategy and the simple suboptimal linear zero-forcing beamforming strategy. In the case of DPC, we show that "SINR duality" and "min-max duality" yield the same dual MAC problem, and compare two alternatives for its efficient solution. In the case of zero-forcing beamforming, we provide a new efficient algorithm based on the direct optimization of a generalized inverse matrix. In both cases, the algorithms presented here address the problems in the most general form and can be applied to special cases previously considered, such as per-antenna and per-group of antennas power constraints, "forbidden interference direction" constraints, or any combination thereof.

cs.IT↗

arXiv2009

Optimized Training and Feedback for MIMO Downlink Channels

We consider a MIMO fading broadcast channel where channel state information is acquired at user terminals via downlink training and channel feedback is used to provide transmitter channel state information (CSIT) to the base station. The feedback channel (the corresponding uplink) is modeled as an AWGN channel, orthogonal across users. The total bandwidth consumed is the sum of the bandwidth/resources used for downlink training, channel feedback, and data transmission. Assuming that the channel follows a block fading model and that zeroforcing beamforming is used, we optimize the net achievable rate for unquantized (analog) and quantized (digital) channel feedback. The optimal number of downlink training pilots is seen to be essentially the same for both feedback techniques, but digital feedback is shown to provide a larger net rate than analog feedback.

cs.IT↗