Co-channel interference, arising either from neighboring cells or due to the co-channel deployment of macro and low-power base stations (BS), limit the uplink rates in 5G and beyond systems. Although interference rejection combining (IRC) equalizer is widely adopted to mitigate this issue, it suffers from higher complexity (cubic in BS antennas) compared to traditional MMSE equalizer (cubic in users). This is particularly concerning given the large number of BS antennas used in 5G and beyond systems. Further, due to inaccurate estimation of interference statistics, IRC suffers from performance loss compared to traditional MMSE at low-to-moderate interference levels. Based on these observations, the present work proposes an opportunistic IRC receiver that utilizes a small neural network at the BS to dynamically select the optimal equalizer between MMSE and IRC. This approach significantly reduces the equalizer complexity while also achieving effective interference mitigation. Our 5G link level simulations show that, with 12 layers and 64 BS antennas, the proposed AI-based opportunistic receiver achieves a 68% reduction in complexity compared to IRC, while also meeting the desired performance target.

In reconfigurable intelligent surfaces (RIS), the reflective elements increase the dimensionality of the overall channel model and, correspondingly, the training requirements. This paper analyzes how channel training and spectral efficiency guide the choice of the operational dimensionality of RIS. We derive an inner bound on the training-based capacity as a function of pilot power, data transmission power, and RIS array dimensions. We study the outcomes and implications of data/pilot power optimization in RIS-assisted systems according to the achievable rate metric. Further, this work sheds light on the tradeoff between the superior array gains of larger RIS on the one hand, and the respective training requirements on the other hand, when optimizing the end-to-end capacity. Beyond a certain critical array size, further increase of array size is not beneficial for spectral efficiency due to the training requirement. This critical size is calculated, and its dependence on the signal-to-noise ratio (SNR) and coherence interval of the channel is clarified. The benefits and drawbacks of different training schemes are analyzed in this context, and demonstrated by simulations.

QR decomposition based M-algorithm (QRDM) is a tree search method that achieves near-optimal signal detection in multiple-input multiple-output (MIMO) systems at a significantly lower complexity compared to maximum-likelihood detection (MLD). However, the complexity of QRDM still remains much higher than linear detectors typically employed in practical systems. While there exist some low-complexity variants of QRDM, they often sacrifice performance for complexity. The present work proposes a robust low-complexity variant of QRDM that leverages an initial linear solution to introduce two fundamental optimizations. First, it computes the costs for only k-bit neighbors of the linear solution in each layer, instead of all symbols of the modulation alphabet. Second, it dynamically changes the number of surviving nodes in each layer based on a threshold on the accumulated cost. These enhancements lead to a significant reduction in complexity while also achieving improved performance compared to conventional QRDM and its existing low-complexity variants. Notably, the proposed detector addresses the problem of premature pruning of reliable solutions in early layers, which is the key cause of sub-optimality in existing QRDM variants.

Modern base station (BS) employ distributed architecture, wherein the overall processing is divided between a radio unit (RU) and a distributed unit (DU), that are connected by a fronthaul link. In open-RAN (O-RAN) 7.2 split architecture, RU performs only a few simple operations, while most of the BS processing happens at the DU, providing the benefits of centralized computing and control. A key challenge in 7.2 split is to limit the overhead on the fronthaul link within acceptable limits. To address this issue, a common solution is to perform combining operation on the received signals at RU antennas, and send low-dimensional signals to DU over fronthaul. The combining matrix used at the RU is typically based on older/aged CSI received from DU, resulting in performance degradation, which is referred as channel aging problem. An alternate approach to limit the fronthaul overhead is to quantize the frequency-domain (i.e., post-FFT) samples at RU to fewer bits (say, 3 or 4 bits) and send the low-resolution signals over fronthaul. Since this approach does not use aged channels, its performance is not affected by channel aging, but is mainly limited by quantization noise. It is recently shown that, under certain conditions, quantization approach achieves superior error performance compared to combining, while also maintaining lower overhead. Motivated by this observation, the present work proposes an AI-based solution to opportunistically switch from combining to quantization under favorable conditions, thereby mitigating channel aging while also reducing fronthaul overhead.

Modern base stations (BS) utilize distributed architectures, wherein the BS functionalities are split between a radio unit (RU) and a distributed unit (DU), that are connected via a fronthaul link. Managing the overhead on this link within tolerable limits is challenging due to ever increasing uplink data traffic and limited fronthaul capacity. In the widely accepted O-RAN 7.2 split architecture, the RU performs combining and sends lower dimensional signals over the fronthaul, thereby reducing the overhead. This approach suffers from performance loss due to channel aging, since RU combining is based on older/aged CSI received from DU. An alternate, less explored approach for reducing overhead is to transmit low-resolution signals over the fronthaul, obtained via quantization of post-FFT samples at the RU. While quantization noise results in some degradation, this approach does not suffer from channel aging. To date, a realistic performance comparison between combining and quantization is not available in the literature. The present work fills this gap and explores the regimes where quantization is favorable in both performance and overhead compared with combining. Our results suggest that opportunistically switching from combining to quantization under favorable conditions achieves superior performance at reduced overhead.

Index modulation is a MIMO technology where some transmit antennas are idle during each transmission interval, but the receiver requires knowledge of all link gains at all times. Even though index modulation is particularly sensitive to the cost of estimating the channel state information (CSI), the impact of CSI cost and imperfections on the capacity of index modulation has not been adequately characterized until now. As a result, the marginal cost/benefit of each additional antenna, and the optimal antenna alphabet have remained unclear. This study computes the spectral efficiency of index modulation subject to training and determines optimal antenna alphabets. Our approach involves a comprehensive examination of the influence of pilot power and degrees of freedom on the achievable rate of index modulation. The results include 2.5dB improvement over the best previously known bound for$4\times2$spatial modulation at 6b/s/Hz. Additionally, we determine the conditions under which single-antenna transmission is superior to spatial modulation, and vice versa. Moreover, this training-based spectral efficiency analysis is extended to the multiuser uplink, identifying the number of users that can be accommodated while maximizing the uplink sum-rate under spatial modulation.

The need for higher uplink data rates in 5G and beyond systems is hindered by co-channel interference that can arise either from the neighboring cells or due to the co-channel deployment of macro and low-power base stations. To mitigate this issue, MMSE with interference rejection combining (IRC) has been widely adopted in practical systems. Unfortunately, MMSE-IRC suffers from higher complexity (cubic in number of BS antennas) compared with traditional MMSE equalizer (cubic in number of users). This is particularly concerning given the large number of BS antennas used in 5G and beyond systems. Further, it is illustrated in the present work that, under certain practical constraints, MMSE-IRC shows either the same or even worse performance compared with MMSE. Based on these observations, the present work proposes an opportunistic receiver that switches to MMSE from MMSE-IRC under favorable conditions, thereby reducing the average complexity without compromising in performance. The proposed receiver is validated through 5G link level simulations under practical channel models, demonstrating a significant reduction in complexity compared with MMSE-IRC.

In spatial modulation, even though only one antenna is active in each transmission interval, the receiver needs channel estimates with respect to all transmit antennas at all times. Thus, spatial modulation is more sensitive to the cost of the estimation of channel state information (CSI) compared with conventional single-antenna transmission. Despite its importance, an accurate, joint characterization of the impact of CSI cost and CSI imperfections on the capacity of spatial modulation has been unavailable thus far. As a result, the marginal cost/benefit of each additional antenna, and hence the optimal antenna alphabet, has been unclear. This work calculates the spectral efficiency of spatial modulation subject to training through a tight characterization of the dependence of the achievable rate on the power and degrees of freedom dedicated to pilots. Our results reliably characterize the cases when conventional single-antenna transmission is superior to spatial modulation or vice versa.

Channel training in reconfigurable intelligent surfaces (RIS) is different from MIMO; in addition to pilots, it also requires setting RIS training states. Several RIS training schemes are studied in the literature, but the effect of training overhead and accuracy on capacity has not received the deserved attention. This is necessary for guiding the choice of RIS dimensionality and the parameters of modulation and coding. The present work fills this gap and shows that the spectral efficiency of RIS with DFT training (that uses the columns of DFT matrix for the RIS training states) is higher than that with the canonical training (that uses the canonical basis for RIS training states). Specifically, with 32 RIS elements, DFT training achieves a 2 bits/s/Hz gain compared to canonical training when the coherence interval is 150 time slots. Our results also reveal that beyond a certain critical RIS array size, further increase in size is not beneficial and that the optimal array size goes down with increase in SNR.

Index modulation represents the transmitted information in two parts: by selecting a subset of available transmission dimensions (antennas, sub-carriers, or time-slots) whose selection index carries information, and by modulation symbols transmitted in the selected dimensions. Index modulation is motivated by reducing the transmitter hardware complexity and has attracted significant research attention in the past decade. In practice, knowing the spectral efficiency or capacity is essential for setting the parameters of modulation and coding for index modulation, but approximations and bounds thus far have not been accurate enough for that purpose. We calculate close lower and upper bounds for the spectral efficiency of index modulation. Our lower and upper bounds meet at high-SNR when the number of receive antennas is greater than or equal to the number of transmit antennas, thus the high-SNR capacity of index modulation in these cases has been fully characterized. A catalog of results is provided for spatial modulation, generalized spatial modulation, time and frequency index modulation. Extensive simulations illustrate the usefulness and accuracy of our results. For example, for spatial modulation with$4\times 2$antennas at 8 bits/s/Hz, our results are 2dB tighter than the best available bounds in the literature.

Index modulation emits information through the index of the activated component of a vector signal as well as the value of the activated component. Indexing could occur across antennas, subcarriers, or other degrees of freedom. When index modulation is applied to antennas, it is known as spatial modulation. Earlier bounds or estimates for the spectral efficiency of spatial modulation have been too loose for determining the parameters of coding and modulation, which are important in practice. Furthermore, the best bounds formerly available did not effectively elucidate the relationship of spatial modulation capacity with SIMO and MIMO capacity at low- and high-SNR. The present work develops novel, tighter bounds on the spectral efficiency of spatial modulation. Specifically, for a 4 × 2 antenna configuration at 8 bits/slHz, our results are 2dB tighter than the best bounds available in the literature.

Time-indexed media-based modulation (TI-MBM) is an index modulation scheme where time slots in a transmission frame are indexed to convey additional information bits in media-based modulation (MBM). It was shown in the literature that, in frequency selective fading channels with inter-symbol interference, TI-MBM with cyclic-prefixed single-carrier (CPSC) scheme can achieve better transmission rates, bit error performance, and multipath diversity gain compared to conventional MBM. In this paper, we analyze the capacity of TI-MBM with CPSC, which has not been reported before. We show that, for a given channel matrix, selecting the index of the activated time slots and RF mirrors according to uniform distribution is suboptimal. We derive the probabilities with which time slots and RF mirrors in TI-MBM can be activated such that the achievable transmission rate is maximized. Further, we prove that the transmit symbols from a Gaussian mixture distribution, whose mixture weights are chosen to be the product of these probabilities, can achieve capacity.

Efficient constellation design is important for improving performance in communication systems. The problem of multidimensional constellation design has been studied extensively in the literature in the context of multidimensional coded modulation and space-time coded MIMO systems. Such constellations are formally called as lattice codes, where a finite set of points from a certain high dimensional lattice is chosen based on some criteria. In this paper, we consider the problem of constellation/signal set design for media-based modulation (MBM), a recent MIMO channel modulation scheme with promising theoretical and practical benefits. Constellation design for MBM is fundamentally different from those for multidimensional coded modulation and conventional MIMO systems mainly because of the inherent sparse structure of the MBM signal vectors. Specifically, we need a structured sparse lattice code with good distance properties. In this work, we show that using an (N,K) non-binary block code in conjunction with the lattice based multilevel squaring construction, it is possible to systematically construct a signal set for MBM with certain guaranteed minimum distance. The MBM signal set obtained using the proposed construction is shown to achieve significantly improved bit error performance compared to conventional MBM signal set. In particular, the proposed signal set is found to achieve higher diversity slopes in the low-to-moderate SNR regime.

In this paper, we consider media-based modulation (MBM) in a cyclic-prefixed single-carrier setting (CPSC-MBM) in inter-symbol interference channels and focus on low-complexity signal detection at the receiver. For this, we exploit the structured sparsity that is inherently present in CPSC-MBM signals. We formulate the sparse vector detection problem of interest as an equivalent sparse matrix detection problem, whose reconstruction complexity can be significantly less without incurring much loss in recovery performance. This technique of recovering sparse matrices, termed as {\em sparse matrix sketching}, applied to CPSC-MBM signal recovery is shown to achieve good bit error performance with significant complexity gains compared to the sparse vector detection approach.

Time-indexed media-based modulation (TI-MBM) is a block transmission scheme, where a block of N time-slots forms a frame of which only K time-slots are used for data transmission (active slots) and the remaining N - K time-slots are left vacant (inactive slots). The choice of the active timeslots conveys additional information bits along with the bits conveyed by the MBM signals transmitted in the active timeslots. It has been shown that TI-MBM can achieve improved bit error performance compared to the conventional MBM scheme. However, a mathematical diversity analysis of TI-MBM has not been reported in the literature so far. In this paper, we view the time-slot indexing as coding across time (since the choice of active and inactive time-slots are dependent on each other) and carry out a mathematical analysis of the diversity performance of TIMBM. Our analytical results show that, while conventional MBM can not extract the multipath diversity of the channel, TI-MBM can achieve higher diversity orders by extracting the channel multipath diversity for suitable choice of system parameters. We provide simulation results that support the analytically predicted diversity orders.

In direct conversion radio frequency (RF) front-end architecture, RF impairments such as IQ imbalance (IQI), DC offset, flicker noise, and oscillator leakage can affect the performance. In this work, we study the performance of media-based modulation (MBM), an attractive channel modulation scheme, in the presence of transmit and receiver side IQI and propose efficient compensation techniques. MBM is shown to be more resilient to IQI compared to conventional modulation. A scheme which jointly estimates and compensates the channel and IQI parameters using widely linear least squares estimation is proposed. The proposed scheme is shown to alleviate the transmit and receive IQI-induced BER degradation in MBM.

Media-based modulation (MBM) is an attractive channel modulation scheme with rate, performance, and hardware advantages. In this paper, we are concerned with two important aspects of MBM. The first one is on the capacity of MBM and the other is on low-complexity detection of high-rate MBM signals using structured spare recovery techniques. We derive closed-form expression for the capacity of generalized MBM (GMBM). The main idea in the capacity analysis is to recognize that MBM uses two alphabets to convey information, namely, the source alphabet (e.g., QAM/PSK) and the channel alphabet (fade coefficients). This observation allows us to show that the capacity is achieved when the mutual information is maximized over the source-channel product alphabet. We then propose a greedy structured sparse recovery algorithm for the detection of GMBM signal vectors. The proposed algorithm is a two-stage algorithm in which support recovery is done in greedy manner first, followed by data detection in the nonzero positions. Simulation results show that the proposed algorithm achieves good performance at low complexity even when the system is highly underdetermined.

Media-based modulation (MBM) is a promising modulation scheme which is attracting recent research attention. In MBM, radio frequency (RF) mirrors are used to create a channel modulation alphabet based on the ON/OFF (i.e., transparent/opaque) status of these mirrors. The index of the mirror activation pattern in a channel use conveys information bits in addition to the bits conveyed through conventional modulation symbols. MBM has been shown to achieve improved performance compared to conventional modulation schemes. In this paper, we introduce time-slot indexing to MBM,which further improves the performance. The proposed time-indexed MBM (TI-MBM) is a block transmission scheme, where a time slot in a given frame can be used or unused, and the choice of the slots used for transmission conveys time-index bits. We study the proposed TI-MBM scheme in frequency-selective channels and show that TI-MBM achieves better performance compared to conventional MBM. Further, recognizing that the TI-MBM signal structure promotes sparsity, we exploit the use of sparse recovery algorithms for the detection of TI-MBM signals. We show that compressive sampling matching pursuit (CoSaMP) and subspace pursuit (SP) based TI-MBM signal detection can achieve significantly improved performance compared to conventional minimum mean square (MMSE) detection.