In this paper we study fully asynchronous random access (RA) multiple antenna receiver in a Rayleigh block-fading AWGN channel with pure path delays. Although our approach can be used to detect users in a grant-free random access system, where each user sporadically and without waiting for a permission from a base station (BS) transmits short messages, we in particular consider the pilot-based unsourced random access (U-RA), where those messages are from a common codebook. The first and arguably the most important task of the pilotbased U-RA receiver is to detect the list of transmitted pilots. Due to the propagation through the considered channel, those pilots are received as a superposition at the BS with delays, due to the lack of perfect timing synchronization. We show that the output of the chip matched filter at the BS receiver can be seen as the superposition of two zero-padded versions of each transmitted pilot sequence. We include this observation in the compressed sensing (CS) formulation of the activity detection (AD) problem, and solve it using the multiple measurement vectors approximate message passing (MMV-AMP) algorithm and a dedicated parametrized 2-level hierarchical sparsity (P2-LHS) denoiser. Our numerical experiments show that the proposed scheme can accurately detect U-RA messages and shows excellent robustness to timing asynchronism. The proposed denoiser vastly outperforms the standard Bernoulli-Gaussian (BG) denoiser for a large range of signal-to-noise ratio (SNR) values. Furthermore, using outage rate analysis, we investigate the performance of the asynchronous U-RA receiver using a Gaussian codebook and minimum distance decoding in the data phase. Our results show that a significant gain in throughput can be achieved when users adopt rate control (RC).

Semantic communication, rather than on a bit-by-bit recovery of the transmitted messages, focuses on the meaning and the goal of the communication itself. In this paper, we propose a novel semantic image coding scheme that preserves the semantic content of an image, while ensuring a good trade-off between coding rate and image quality. The proposed Semantic-Preserving Image Coding based on Conditional Diffusion Models (SPIC) transmitter encodes a Semantic Segmentation Map (SSM) and a low-resolution version of the image to be transmitted. The receiver then reconstructs a high-resolution image using a Denoising Diffusion Probabilistic Models (DDPM) doubly conditioned to the SSM and the low-resolution image. As shown by the numerical examples, compared to state-of-the-art (SOTA) approaches, the proposed SPIC exhibits a better balance between the conventional rate-distortion trade-off and the preservation of semantically-relevant features. Code available at https://github.com/frapez1/SPIC

To efficiently utilize the scarce wireless resource, the random access scheme has been attaining renewed interest primarily in supporting the sporadic traffic of a large number of devices encountered in the Internet of Things (IoT). In this paper we investigate the performance of slotted ALOHA—a simple and practical random access scheme—in connection with the grant-free random access protocol applied for user-centric cell-free massive MIMO. More specifically, we provide the expression of the sum-throughput under the assumptions of the capture capability owned by the centralized detector in the uplink. Further, a comparative study of user-centric cell-free massive MIMO with other types of networks is provided, which allows us to identify its potential and possible limitation. Our numerical simulations show that the user-centric cell-free massive MIMO has a good trade-off between performance and fronthaul load, especially at low activation probability regime.

In this paper we study the message detection (MD) problem for a chip-asynchronous random access (RA) multiple antennas receiver and a Rayleigh block-fading AWGN channel with a pure path delay. Although our approach can be used to detect users in a grant-free random access system, where each user sporadically and without waiting for a permission from a base station (BS) transmits a short message, we in particular consider unsourced random access (U-RA), where those messages are from a common codebook. The first and arguably the most important task of the U-RA receiver is to detect the list of transmitted messages. Due to the propagation through the considered channel, those messages are received as a superposition at the BS with delays. In order to reduce the overhead for timing synchronisation, which would be wasteful for short messages, in this work we provide an asynchronous operating mode for MD in U-RA. To do this, we show that when using a chip matched filter at the BS, the contribution of each transmitted message in the sampled received signal can be described as a superposition of two zero-padded versions of that message. We include this observation in the compressed sensing (CS) formulation of the MD problem, and solve it using the multiple measurement vectors approximate message passing (MMV-AMP) algorithm and a dedicated 2-level hierarchical sparsity (2-LHS) denoiser. Our numerical experiments show that the proposed scheme can accurately detect U-RA messages in the considered channel without any overhead for timing synchronization.

We consider the unsourced random access problem on a Rayleigh block-fading AWGN channel with multiple receive antennas. Specifically, we treat the slow fading scenario where the coherence blocklength is large compared to the number of active users and a message can be transmitted in a single fading coherence block. Unsourced random access refers to a form of grant-free random access where users are constrained to use the same codebook and therefore are a priori indistinguishable. The receiver must recover the list of transmitted messages up to permutations. In this paper, we propose an approach based on splitting the user messages into two parts. First, a small block of bits selects a relatively short codeword from a common “pilot” codebook. Then the remaining message bits are encoded by a standard block code for the Gaussian channel. The receiver makes use of a multiple measurement vector approximate message passing (MMV-AMP) algorithm to estimate the active user channels from the “pilot” part, and then uses the estimated channels to perform coherent maximum ratio combining (MRC) to decode the second part. We provide an accurate closed-form approximated analysis of the proposed scheme. Furthermore, we analyze the MRC decoding when successive interference cancellation is performed over groups of users, striking an attractive tradeoff between complexity and performance. Finally, we investigate the impact of power control policies, taking into account the unique nature of massive random access. As a byproduct, we also present an extension of the MMV-AMP algorithm which allows pathloss coefficients to be treated as deterministic unknowns by performing maximum likelihood estimation in each step of the MMV-AMP algorithm.

Deep unfolding showed to be a very successful approach for accelerating and tuning classical signal processing algorithms. In this paper, we propose learned Gaussian-mixture AMP (L-GM-AMP) - a plug-and-play compressed sensing (CS) recovery algorithm suitable for any i.i.d. source prior. Our algorithm builds upon Borgerding’s learned AMP (LAMP), yet significantly improves it by adopting a universal denoising function within the algorithm. The robust and flexible denoiser is a byproduct of modelling source prior with a Gaussian-mixture (GM), which can well approximate continuous, discrete, as well as mixture distributions. Its parameters are learned using standard backpropagation algorithm. To demonstrate robustness of the proposed algorithm, we conduct Monte-Carlo (MC) simulations for both mixture and discrete distributions. Numerical evaluation shows that the L-GM-AMP algorithm achieves state-of-the-art performance without any knowledge of the source prior.