arXiv subjects

Jakob Hoydis

Publications and source records attributed to Jakob Hoydis.

At least 55 records · Page 3Linked to original sources

arXiv2020

Trainable Communication Systems: Concepts and Prototype

We consider a trainable point-to-point communication system, where both transmitter and receiver are implemented as neural networks (NNs), and demonstrate that training on the bit-wise mutual information (BMI) allows seamless integration with practical bit-metric decoding (BMD) receivers, as well as joint optimization of constellation shaping and labeling. Moreover, we present a fully differentiable neural iterative demapping and decoding (IDD) structure which achieves significant gains on additive white Gaussian noise (AWGN) channels using a standard 802.11n low-density parity-check (LDPC) code. The strength of this approach is that it can be applied to arbitrary channels without any modifications. Going one step further, we show that careful code design can lead to further performance improvements. Lastly, we show the viability of the proposed system through implementation on software-defined radios (SDRs) and training of the end-to-end system on the actual wireless channel. Experimental results reveal that the proposed method enables significant gains compared to conventional techniques.

cs.IT↗

arXiv2020

Joint Learning of Probabilistic and Geometric Shaping for Coded Modulation Systems

We introduce a trainable coded modulation scheme that enables joint optimization of the bit-wise mutual information (BMI) through probabilistic shaping, geometric shaping, bit labeling, and demapping for a specific channel model and for a wide range of signal-to-noise ratios (SNRs). Compared to probabilistic amplitude shaping (PAS), the proposed approach is not restricted to symmetric probability distributions, can be optimized for any channel model, and works with any code rate $k/m$, $m$ being the number of bits per channel use and $k$ an integer within the range from $1$ to $m-1$. The proposed scheme enables learning of a continuum of constellation geometries and probability distributions determined by the SNR. Additionally, the PAS architecture with Maxwell-Boltzmann (MB) as shaping distribution was extended with a neural network (NN) that controls the MB shaping of a quadrature amplitude modulation (QAM) constellation according to the SNR, enabling learning of a continuum of MB distributions for QAM. Simulations were performed to benchmark the performance of the proposed joint probabilistic and geometric shaping scheme on additive white Gaussian noise (AWGN) and mismatched Rayleigh block fading (RBF) channels.

cs.IT↗

arXiv2020

"Machine LLRning": Learning to Softly Demodulate

Soft demodulation, or demapping, of received symbols back into their conveyed soft bits, or bit log-likelihood ratios (LLRs), is at the very heart of any modern receiver. In this paper, a trainable universal neural network-based demodulator architecture, dubbed "LLRnet", is introduced. LLRnet facilitates an improved performance with significantly reduced overall computational complexity. For instance for the commonly used quadrature amplitude modulation (QAM), LLRnet demonstrates LLR estimates approaching the optimal log maximum a-posteriori inference with an order of magnitude less operations than that of the straightforward exact implementation. Link-level simulation examples for the application of LLRnet to 5G-NR and DVB-S.2 are provided. LLRnet is a (yet another) powerful example for the usefulness of applying machine learning to physical layer design.

cs.IT↗

arXiv2020

Deep HyperNetwork-Based MIMO Detection

Optimal symbol detection for multiple-input multiple-output (MIMO) systems is known to be an NP-hard problem. Conventional heuristic algorithms are either too complex to be practical or suffer from poor performance. Recently, several approaches tried to address those challenges by implementing the detector as a deep neural network. However, they either still achieve unsatisfying performance on practical spatially correlated channels, or are computationally demanding since they require retraining for each channel realization. In this work, we address both issues by training an additional neural network (NN), referred to as the hypernetwork, which takes as input the channel matrix and generates the weights of the neural NN-based detector. Results show that the proposed approach achieves near state-of-the-art performance without the need for re-training.

cs.IT↗

arXiv2019

Towards Massive MIMO 2.0: Understanding spatial correlation, interference suppression, and pilot contamination

Since the seminal paper by Marzetta from 2010, Massive MIMO has changed from being a theoretical concept with an infinite number of antennas to a practical technology. The key concepts are adopted in 5G and base stations (BSs) with $M=64$ full-digital transceivers have been commercially deployed in sub-6\,GHz bands. The fast progress was enabled by many solid research contributions of which the vast majority assume spatially uncorrelated channels and signal processing schemes developed for single-cell operation. These assumptions make the performance analysis and optimization of Massive MIMO tractable but have three major caveats: 1) practical channels are spatially correlated; 2) large performance gains can be obtained by multicell processing, without BS cooperation; 3) the interference caused by pilot contamination creates a finite capacity limit, as $M\to\infty$. There is a thin line of papers that avoided these caveats, but the results are easily missed. Hence, this tutorial article explains the importance of considering spatial channel correlation and using signal processing schemes designed for multicell networks. We present recent results on the fundamental limits of Massive MIMO, which are not determined by pilot contamination but the ability to acquire channel statistics. These results will guide the journey towards the next level of Massive MIMO, which we call ``Massive MIMO 2.0''.

eess.SP↗

arXiv2019

Learning to Communicate and Energize: Modulation, Coding and Multiple Access Designs for Wireless Information-Power Transmission

The explosion of the number of low-power devices in the next decades calls for a re-thinking of wireless network design, namely, unifying wireless transmission of information and power so as to make the best use of the RF spectrum, radiation, and infrastructure for the dual purpose of communicating and energizing. This paper provides a novel learning-based approach towards such wireless network design. To that end, a parametric model of a practical energy harvester, accounting for various sources of nonlinearities, is proposed using a nonlinear regression algorithm applied over collected real data. Relying on the proposed model, the learning problem of modulation design for Simultaneous Wireless Information-Power Transmission (SWIPT) over a point-to-point link is studied. Joint optimization of the transmitter and the receiver is implemented using Neural Network (NN)-based autoencoders. The results reveal that by increasing the receiver power demand, the baseband transmit modulation constellation converges to an On-Off keying signalling. Utilizing the observations obtained via learning, an algorithmic SWIPT modulation design is proposed. It is observed via numerical results that the performance loss of the proposed modulations are negligible compared to the ones obtained from learning. Extension of the studied problem to learning modulation design for multi-user SWIPT scenarios and coded modulation design for point-to-point SWIPT are considered. The major conclusion of this work is to utilize learning-based results to design non learning-based algorithms, which perform as well. In particular, inspired by the results obtained via learning, an algorithmic approach for coded modulation design is proposed, which performs very close to its learning counterparts, and is significantly superior due to its high real-time adaptability to new system design parameters.

cs.IT↗

arXiv2019

Learning Modulation Design for SWIPT with Nonlinear Energy Harvester: Large and Small Signal Power Regimes

Nonlinear energy harvesters (EH) behave differently depending on the range of their input power. In the literature, different models have been proposed mainly for relatively small and large input power regimes of an EH. Due to the complexity of the proposed nonlinear models, obtaining analytical optimal or well performing signal designs have been extremely challenging. Relying on the proposed models in the literature, the learning problem of modulation design for simultaneous wireless information-power transfer (SWIPT) over a point-to-point link is studied. Joint optimization of the transmitter and the receiver is implemented using neural network (NN)-based autoencoders. The results reveal that for relatively small channel input powers, as the power demand increases at the receiver, one of the symbols is shot away from the origin while the remaining symbols approach zero amplitude. In the very extreme case of merely receiver power demand, the modulations are in the form of On-Off keying signalling with a low probability of the On signal. On the other side, for relatively large channel input powers, it is observed that as the receiver power demand increases, a number of symbols approach zero amplitude, whereas the others (more than one symbol) get equally high amplitudes but with different phases. In the extreme scenario of merely receiver power demand, the modulation resembles multiple On-Off keying signalling with different phases.

cs.IT↗

arXiv2019

Joint Learning of Geometric and Probabilistic Constellation Shaping

The choice of constellations largely affects the performance of communication systems. When designing constellations, both the locations and probability of occurrence of the points can be optimized. These approaches are referred to as geometric and probabilistic shaping, respectively. Usually, the geometry of the constellation is fixed, e.g., quadrature amplitude modulation (QAM) is used. In such cases, the achievable information rate can still be improved by probabilistic shaping. In this work, we show how autoencoders can be leveraged to perform probabilistic shaping of constellations. We devise an information-theoretical description of autoencoders, which allows learning of capacity-achieving symbol distributions and constellations. Recently, machine learning techniques to perform geometric shaping were proposed. However, probabilistic shaping is more challenging as it requires the optimization of discrete distributions. Furthermore, the proposed method enables joint probabilistic and geometric shaping of constellations over any channel model. Simulation results show that the learned constellations achieve information rates very close to capacity on an additive white Gaussian noise (AWGN) channel and outperform existing approaches on both AWGN and fading channels.

cs.IT↗

arXiv2019

Model-free Training of End-to-end Communication Systems

The idea of end-to-end learning of communication systems through neural network-based autoencoders has the shortcoming that it requires a differentiable channel model. We present in this paper a novel learning algorithm which alleviates this problem. The algorithm enables training of communication systems with an unknown channel model or with non-differentiable components. It iterates between training of the receiver using the true gradient, and training of the transmitter using an approximation of the gradient. We show that this approach works as well as model-based training for a variety of channels and tasks. Moreover, we demonstrate the algorithm's practical viability through hardware implementation on software-defined radios where it achieves state-of-the-art performance over a coaxial cable and wireless channel.

cs.IT↗

arXiv2019

Deep Reinforcement Learning Autoencoder with Noisy Feedback

End-to-end learning of communication systems enables joint optimization of transmitter and receiver, implemented as deep neural network-based autoencoders, over any type of channel and for an arbitrary performance metric. Recently, an alternating training procedure was proposed which eliminates the need for an explicit channel model. However, this approach requires feedback of real-valued losses from the receiver to the transmitter during training. In this paper, we first show that alternating training works even with a noisy feedback channel. Then, we design a system that learns to transmit real numbers over an unknown channel without a preexisting feedback link. Once trained, this feedback system can be used to communicate losses during alternating training of autoencoders. Evaluations over additive white Gaussian noise and Rayleigh block-fading channels show that end-to-end communication systems trained using the proposed feedback system achieve the same performance as when trained with a perfect feedback link.

cs.IT↗

arXiv2019

Massive MIMO is a Reality -- What is Next? Five Promising Research Directions for Antenna Arrays

Massive MIMO (multiple-input multiple-output) is no longer a "wild" or "promising" concept for future cellular networks - in 2018 it became a reality. Base stations (BSs) with 64 fully digital transceiver chains were commercially deployed in several countries, the key ingredients of Massive MIMO have made it into the 5G standard, the signal processing methods required to achieve unprecedented spectral efficiency have been developed, and the limitation due to pilot contamination has been resolved. Even the development of fully digital Massive MIMO arrays for mmWave frequencies - once viewed prohibitively complicated and costly - is well underway. In a few years, Massive MIMO with fully digital transceivers will be a mainstream feature at both sub-6 GHz and mmWave frequencies. In this paper, we explain how the first chapter of the Massive MIMO research saga has come to an end, while the story has just begun. The coming wide-scale deployment of BSs with massive antenna arrays opens the door to a brand new world where spatial processing capabilities are omnipresent. In addition to mobile broadband services, the antennas can be used for other communication applications, such as low-power machine-type or ultra-reliable communications, as well as non-communication applications such as radar, sensing and positioning. We outline five new Massive MIMO related research directions: Extremely large aperture arrays, Holographic Massive MIMO, Six-dimensional positioning, Large-scale MIMO radar, and Intelligent Massive MIMO.

eess.SP↗

arXiv2019

Adaptive Neural Signal Detection for Massive MIMO

Symbol detection for Massive Multiple-Input Multiple-Output (MIMO) is a challenging problem for which traditional algorithms are either impractical or suffer from performance limitations. Several recently proposed learning-based approaches achieve promising results on simple channel models (e.g., i.i.d. Gaussian). However, their performance degrades significantly on real-world channels with spatial correlation. We propose MMNet, a deep learning MIMO detection scheme that significantly outperforms existing approaches on realistic channels with the same or lower computational complexity. MMNet's design builds on the theory of iterative soft-thresholding algorithms and uses a novel training algorithm that leverages temporal and spectral correlation to accelerate training. Together, these innovations allow MMNet to train online for every realization of the channel. On i.i.d. Gaussian channels, MMNet requires two orders of magnitude fewer operations than existing deep learning schemes but achieves near-optimal performance. On spatially-correlated channels, it achieves the same error rate as the next-best learning scheme (OAMPNet) at 2.5dB lower SNR and with at least 10x less computational complexity. MMNet is also 4--8dB better overall than a classic linear scheme like the minimum mean square error (MMSE) detector.

eess.SP↗

arXiv2019

Transmitter Classification With Supervised Deep Learning

Hardware imperfections in RF transmitters introduce features that can be used to identify a specific transmitter amongst others. Supervised deep learning has shown good performance in this task but using datasets not applicable to real world situations where topologies evolve over time. To remedy this, the work rests on a series of datasets gathered in the Future Internet of Things / Cognitive Radio Testbed [4] (FIT/CorteXlab) to train a convolutional neural network (CNN), where focus has been given to reduce channel bias that has plagued previous works and constrained them to a constant environment or to simulations. The most challenging scenarios provide the trained neural network with resilience and show insight on the best signal type to use for identification , namely packet preamble. The generated datasets are published on the Machine Learning For Communications Emerging Technologies Initiatives web site 4 in the hope that they serve as stepping stones for future progress in the area. The community is also invited to reproduce the studied scenarios and results by generating new datasets in FIT/CorteXlab.

eess.SP↗

arXiv2019

Smart Radio Environments Empowered by AI Reconfigurable Meta-Surfaces: An Idea Whose Time Has Come

Future wireless networks are expected to constitute a distributed intelligent wireless communications, sensing, and computing platform, which will have the challenging requirement of interconnecting the physical and digital worlds in a seamless and sustainable manner. Currently, two main factors prevent wireless network operators from building such networks: 1) the lack of control of the wireless environment, whose impact on the radio waves cannot be customized, and 2) the current operation of wireless radios, which consume a lot of power because new signals are generated whenever data has to be transmitted. In this paper, we challenge the usual "more data needs more power and emission of radio waves" status quo, and motivate that future wireless networks necessitate a smart radio environment: A transformative wireless concept, where the environmental objects are coated with artificial thin films of electromagnetic and reconfigurable material (that are referred to as intelligent reconfigurable meta-surfaces), which are capable of sensing the environment and of applying customized transformations to the radio waves. Smart radio environments have the potential to provide future wireless networks with uninterrupted wireless connectivity, and with the capability of transmitting data without generating new signals but recycling existing radio waves. This paper overviews the current research efforts on smart radio environments, the enabling technologies to realize them in practice, the need of new communication-theoretic models for their analysis and design, and the long-term and open research issues to be solved towards their massive deployment. In a nutshell, this paper is focused on discussing how the availability of intelligent reconfigurable meta-surfaces will allow wireless network operators to redesign common and well-known network communication paradigms.

cs.IT↗

arXiv2019

Towards Hardware Implementation of Neural Network-based Communication Algorithms

There is a recent interest in neural network (NN)-based communication algorithms which have shown to achieve (beyond) state-of-the-art performance for a variety of problems or lead to reduced implementation complexity. However, most work on this topic is simulation based and implementation on specialized hardware for fast inference, such as field-programmable gate arrays (FPGAs), is widely ignored. In particular for practical uses, NN weights should be quantized and inference carried out by a fixed-point instead of floating-point system, widely used in consumer class computers and graphics processing units (GPUs). Moving to such representations enables higher inference rates and complexity reductions, at the cost of precision loss. We demonstrate that it is possible to implement NN-based algorithms in fixed-point arithmetic with quantized weights at negligible performance loss and with hardware complexity compatible with practical systems, such as FPGAs and application-specific integrated circuits (ASICs).

cs.IT↗

arXiv2019

Novel Massive MIMO Channel Sounding Data Applied to Deep Learning-based Indoor Positioning

With a significant increase in area throughput, Massive MIMO has become an enabling technology for fifth generation (5G) wireless mobile communication systems. Although prototypes were built, an openly available dataset for channel impulse responses to verify assumptions, e.g. regarding channel sparsity, is not yet available. In this paper, we introduce a novel channel sounder architecture, capable of measuring multiantenna and multi-subcarrier channel state information (CSI) at different frequency bands, antenna geometries and propagation environments. The channel sounder has been verified by evaluation of channel data from first measurements. Such datasets can be used to study various deep-learning (DL) techniques in different applications, e.g., for indoor user positioning in three dimensions, as is done in this paper. Not only we do achieve an accuracy better than 75 cm for line of sight (LoS), as is comparable to state-of-the-art conventional positioning techniques, but also obtain similar precision for the more challenging case of non-line of sight (NLoS). Further extensive indoor/outdoor measurement campaigns will provide a more comprehensive open CSI dataset, tagged with positions, for the scientific community to further test various algorithms.

eess.SP↗

arXiv2019

Enabling FDD Massive MIMO through Deep Learning-based Channel Prediction

A major obstacle for widespread deployment of frequency division duplex (FDD)-based Massive multiple-input multiple-output (MIMO) communications is the large signaling overhead for reporting full downlink (DL) channel state information (CSI) back to the basestation (BS), in order to enable closed-loop precoding. We completely remove this overhead by a deep-learning based channel extrapolation (or "prediction") approach and demonstrate that a neural network (NN) at the BS can infer the DL CSI centered around a frequency $f_\text{DL}$ by solely observing uplink (UL) CSI on a different, yet adjacent frequency band around $f_\text{UL}$; no more pilot/reporting overhead is needed than with a genuine time division duplex (TDD)-based system. The rationale is that scatterers and the large-scale propagation environment are sufficiently similar to allow a NN to learn about the physical connections and constraints between two neighboring frequency bands, and thus provide a well-operating system even when classic extrapolation methods, like the Wiener filter (used as a baseline for comparison throughout) fails. We study its performance for various state-of-the-art Massive MIMO channel models, and, even more so, evaluate the scheme using actual Massive MIMO channel measurements, rendering it to be practically feasible at negligible loss in spectral efficiency when compared to a genuine TDD-based system.

cs.IT↗

arXiv2018

End-to-End Learning of Communications Systems Without a Channel Model

The idea of end-to-end learning of communications systems through neural network -based autoencoders has the shortcoming that it requires a differentiable channel model. We present in this paper a novel learning algorithm which alleviates this problem. The algorithm iterates between supervised training of the receiver and reinforcement learning -based training of the transmitter. We demonstrate that this approach works as well as fully supervised methods on additive white Gaussian noise (AWGN) and Rayleigh block-fading (RBF) channels. Surprisingly, while our method converges slower on AWGN channels than supervised training, it converges faster on RBF channels. Our results are a first step towards learning of communications systems over any type of channel without prior assumptions.

cs.IT↗