The Terahertz (THz) band (0.1–10 THz) is projected to enable broadband wireless communications of the future, and many envision deep learning as a solution to improve the performance of THz communication systems and networks. However, there are few available datasets of true THz signals that could enable testing and training of deep learning algorithms for the research community. In this paper, we provide an extensive dataset of 120,000 data frames for the research community. All signals were transmitted at 165 GHz but with varying bandwidths (5 GHz, 10 GHz, and 20 GHz), modulations (4PSK, 8PSK, 16QAM, and 64QAM), and transmit amplitudes (75 mV and 600 mV), resulting in twenty-four distinct bandwidth-modulation-power combinations each with 5,000 unique captures. The signals were captured after down conversion at an intermediate frequency of 10 GHz. This dataset enables the research community to experimentally explore solutions relating to ultrabroadband deep and machine learning applications.

The high mobility, density and multi-path evident in modern wireless systems makes the channel highly non-stationary. This causes temporal variation in the channel distribution that leads to the existence of time-varying joint interference across multiple degrees of freedom (DoF, e.g., users, antennas, frequency and symbols), which renders conventional precoding sub-optimal in practice. In this work, we derive a High-Order Generalization of Mercer’s Theorem (HOGMT), which decomposes the multi-user non-stationary channel into two (dual) sets of jointly orthogonal subchannels (eigenfunctions), that result in the other set when one set is transmitted through the channel. This duality and joint orthogonality of eigenfuntions ensure transmission over independently flat-fading subchannels. Consequently, transmitting these eigenfunctions with optimally derived coefficients eventually mitigates any interference across its degrees of freedoms and forms the foundation of the proposed joint spatio-temporal precoding. The transferred dual eigenfuntions and coefficients directly reconstruct the data symbols at the receiver upon demodulation, thereby significantly reducing its computational burden, by alleviating the need for any complementary post-coding. Additionally, the eigenfunctions decomposed from the time-frequency delay-Doppler channel kernel are paramount to extracting the second-order channel statistics, and therefore completely characterize the underlying channel. We evaluate this using a realistic non-stationary channel framework built in Matlab and show that our precoding achieves${\geqslant }4$orders of reduction in BER at SNR${\geqslant }15$dB in OFDM systems for higher-order modulations and less complexity compared to the state-of-the-art precoding.

Terahertz (THz)-band (0.1–10 THz) communication will be key in enabling high speed wireless links due to the wide available bandwidths. At THz frequencies, the path-loss is governed by high spreading loss due to small antenna apertures and by molecular absorption loss due to water vapor. The latter also determines the available transmission bandwidth, which shrinks with distance. Modulations that consider the high propagation loss and the distance-dependent bandwidth are needed to fully exploit the THz channel’s bandwidth. Using a hierarchical constellation to simultaneously service users at symbol rates, Hierarchical Bandwidth Modulation (HBM) leverages molecular absorption to increase aggregate data rates in a broadcast system while offering flexibility to receivers experiencing high path loss. This paper introduces HBM and evaluates its performance. The symbol error rate performance for a 4/M-QAM HBM system is derived and verified using simulations. These results are used to define the design constraints for an HBM system: the HBM functional region and transition region. The functional region is verified using an experimental testbed for ultrabroadband communications. The results show that with proper design HBM successfully achieves its goal to exploit the distance-dependent characteristics of the THz channel, to spatially multiplex users, and to increase the system capacity.

Ultra-broadband communication in emerging spectrum, like Terahertz (THz) band, is the frontier to meet the data rate requirements of future wireless communication systems. Existing signal processing based methods are developed for sub-6 GHz band, which cannot capture the intricacies in ultra-broad THz bandwidth and non-linearities arising from hardware. To overcome these limitations, we develop neural network (NN) models for OFDM receiver, where expert knowledge of wireless communication is infused in different stages and parameters of the model to create a practical receiver that can adapt to different wireless environments. The parameters of the NN are derived from underlying theory and can be adapted to different wireless environments. Our model is designed to capture the correlation between real and imaginary components of wireless signals, that can be trained with limited data. The models are trained with over-the-air captured OFDM signals, transmitted in THz band with 10 GHz bandwidth. Our results show significant improvement in bit error rate (BER) for different modulation orders (upto 6 dB in BPSK and 1.2 dB in QAM 64) compared to existing signal processing based receiver designs.

Modern wireless channels are increasingly dense and mobile making the channel highly non-stationary. The time-varying distribution and the existence of joint interference across multiple degrees of freedom (e.g., users, antennas, frequency and symbols) in such channels render conventional precoding sub-optimal in practice, and have led to historically poor characterization of their statistics. The core of our work is the derivation of a high-order generalization of Mercer’s Theorem to decompose the non-stationary channel into constituent fading sub-channels (2-D eigenfunctions) that are jointly orthogonal across its degrees of freedom. Consequently, transmitting these eigenfunctions with optimally derived coefficients eventually mitigates any interference across these dimensions and forms the foundation of the proposed joint spatio-temporal precoding. The precoded symbols directly reconstruct the data symbols at the receiver upon demodulation, thereby significantly reducing its computational burden, by alleviating the need for any complementary decoding. These eigenfunctions are paramount to extracting the second-order channel statistics, and therefore completely characterize the underlying channel. Theory and simulations show that such precoding leads to >104× BER improvement (at 20dB) over existing methods for non-stationary channels.

Enabling highly-mobile millimeter wave (mmWave) and terahertz (THz) wireless communication applications requires overcoming the critical challenges associated with the large antenna arrays deployed at these systems. In particular, adjusting the narrow beams of these antenna arrays typically incurs high beam training overhead that scales with the number of antennas. To address these challenges, this paper proposes a multi-modal machine learning based approach that leverages positional and visual (camera) data collected from the wireless communication environment for fast beam prediction. The developed framework has been tested on a real-world vehicular dataset comprising practical GPS, camera, and mmWave beam training data. The results show the proposed approach achieves more than 75% top-1 beam prediction accuracy and close to 100% top-3≈beam prediction accuracy in realistic communication scenarios.

Neural Network (NN) based receivers have seen limited adoption in practical systems due to a lack of explainability and performance guarantees, despite their efficacy as a data-driven tool for physical layer signal processing. In order to bridge this gap in explainability, we present an equivalent NN-based receiver that performs the same optimizations used by classical receivers for symbol detection. Achieving equivalence is crucial to explaining how a NN-based receiver classifies symbols in high-dimensional channels and determining its structure that is robust to the underlying channel with minimum training. We realize this by deriving the risk function that guarantees equivalence, which also provides a measure of the disparity between NN-based and classical receivers. Consequently, this information allows us to derive mathematically tight data-dependent bounds on the bit error rate of NN-based receivers, and empirically determine its structure that achieves minimum error rate. Extensive simulation results show the efficacy of the derived bounds and structure of NN-based receivers for single and multi-antenna systems over a variety of channels.

Terahertz (THz)-band (0.1 - 10 THz) communication is envisioned as a key wireless technology to satisfy the need for higher wireless data rates in denser networks. Several ongoing approaches are being considered to overcome the grand challenge of the THz band, i.e., the limited communication distance. Among others, the use of new 2D nanomaterials such as graphene to create novel plasmonic devices that operate directly in the THz range and can be densely packed has been proposed. This paper presents a novel THz plasmonic array architecture which leverages the properties of graphene to greatly simplify its design and operation. Each element of the plasmonic array is an independent front-end, consisting of an on-chip plasmonic source, modulator and antenna. The advantages of this array architecture over conventional array architectures are discussed. The trade-offs in the design of the front-end and the array are exhaustively studied in transmission. The ability to perform continuous dynamic beamforming is presented. A new tailored algorithm is developed for beamforming weight selection. Extensive numerical results are provided to demonstrate the functionality of the array for dynamic beamforming and increased power output.

Reservoir Computing (RC) is a recent neurologically inspired concept for processing time dependent data that lends itself particularly well to hardware implementation by using the device physics to conduct information processing. In this paper, we apply RC to channel estimation in Orthogonal Frequency Division Multiplexing (OFDM) systems. Due to the multipath propagation environment between a transmitter and receiver, the received signal undergoes attenuation, time delay and phase shift. For mitigating these random effects and decoding the transmitted signal at the receiver, accurate channel estimation is vital. Statistical approaches for channel estimation assume that accurate channel information is available at the receiver. However, the time-variance of the channel complicates the channel estimation process by making the current estimation outdated. Recurrent Neural Networks (RNNs), which are analogous to the functioning of the human brain, are therefore utilized for channel prediction. Training algorithms for RNNs are categorized as gradient-descent methods, which often results in high computational complexity and leads to non-convergence due to the presence of bifurcations. In this paper, an Echo State Network (ESN), which is a class of RC approach, has been used for training a RNN to estimate the channel state information. Using this approach, the training and hence, the implementation complexity is significantly reduced. Simulation results show significant improvement in channel estimation accuracy for the proposed method.

We investigate the problem of embedding data in raw video sequences with minimum video mean-square distortion for any required data recovery error rate. In particular, for any given video frame sequence and any (block) transform domain of interest, we find the optimal carrier and scalar parametrized linear operator on the video data that maximize the output signal-to-interference-plus-noise ratio (SINR) of the maximum-SINR data receiver filter or, equivalently, minimize the average embedding distortion for any target message extraction error rate. The procedure is extended from single-carrier to multi-carrier (multiple messages) embedding. As a practical consideration, a sub-optimal computationally efficient embedding algorithm is also proposed. Extensive experimental results demonstrate that sub-optimal embedding as described has video distortion versus data extraction error rate performance comparable to optimal embedding. Our studies also demonstrate the robustness of the optimal (and sub-optimal) embedding schemes to H.264 compliant encoding.