Neuromorphic Wireless Device-Edge Co-Inference via the Directed Information Bottleneck

An important use case of next-generation wireless systems is device-edge co-inference, where a semantic task is partitioned between a device and an edge server. The device carries out data collection and partial processing of the data, while the remote server completes the given task based on information received from the device. It is often required that processing and communication be run as efficiently as possible at the device, while more computing resources are available at the edge. To address such scenarios, we introduce a new system solution, termed neuromorphic wireless device-edge co-inference. According to it, the device runs sensing, processing, and communication units using neuromorphic hardware, while the server employs conventional radio and computing technologies. The proposed system is designed using a transmitter-centric information-theoretic criterion that targets a reduction of the communication overhead, while retaining the most relevant information for the end-to-end semantic task of interest. Numerical results on standard data sets validate the proposed architecture, and a preliminary testbed realization is reported.

Federated Learning in UAV-Enhanced Networks: Joint Coverage and Convergence Time Optimization

Federated learning (FL) involves several devices that collaboratively train a shared model without transferring their local data. FL reduces the communication overhead, making it a promising learning method in UAV-enhanced wireless networks with scarce energy resources. Despite the potential, implementing FL in UAV-enhanced networks is challenging, as conventional UAV placement methods that maximize coverage increase the FL delay significantly. Moreover, the uncertainty and lack of a priori information about crucial variables, such as channel quality, exacerbate the problem. In this paper, we first analyze the statistical characteristics of a UAV-enhanced wireless sensor network (WSN) with energy harvesting. We then develop a model and solution based on the multi-objective multi-armed bandit theory to maximize the network coverage while minimizing the FL delay. Besides, we propose another solution that is particularly useful with large action sets and strict energy constraints at the UAVs. Our proposal uses a scalarized best-arm identification algorithm to find the optimal arms that maximize the ratio of the expected reward to the expected energy cost by sequentially eliminating one or more arms in each round. Then, we derive the upper bound on the error probability of our multi-objective and cost-aware algorithm. Numerical results show the effectiveness of our approach.

D-Band RIS as a Reflect Array: Characterization and Hardware Impairments Study

Reflecting intelligent surface (RIS) has emerged as a promising technology for enhancing wireless communication performance and enabling new applications in 6G networks with potentially low energy consumption and hardware complexity thanks to their passive nature. Despite the significant growth of the literature on RIS in recent years, covering various aspects of this technology, challenges and issues regarding the practical implementation of RIS have been less addressed. This issue is even more severe at D-band frequencies due to the inherent challenges. This paper aims to connect the requirements and aspirations from the link-level side to the actually achievable RIS hardware with the focus on the various models for RIS. In order to obtain a realistic hardware scenario while maintaining a manageable parameter set, a static reflect array with similar reflection behavior as the reconfigurable one is employed. The results of the study enable an improved RIS design process due to more realistic models. The special focus of this paper is on the hardware impairments, in particular, (i) the effect of specular reflection, and (ii) the beam squint effect.

Near-Optimal LOS and Orientation Aware Intelligent Reflecting Surface Placement

Due to their passive nature and thus low energy consumption, intelligent reflecting surfaces (IRSs) have shown promise as means of extending coverage as a proxy for connection reliability. The relative locations of the base station (BS), IRS, and user equipment (UE) determine the extent of the coverage that the IRS provides which demonstrates the importance of the IRS placement problem. More specifically, locations, which determine whether BS-IRS and IRS-UE line of sight (LOS) links exist, and surface orientation, which determines whether the BS and UE are within the field of view (FoV) of the surface, play crucial roles in the quality of provided coverage. Moreover, another challenge is high computational complexity, since the IRS placement problem is a combinatorial optimization, and is NP-hard. Identifying the orientation of the surface and LOS channel as two crucial factors, we propose an efficient IRS placement algorithm that takes these two characteristics into account in order to maximize the network coverage. We prove the submodularity of the objective function which establishes near-optimal performance bounds for the algorithm. Simulation results demonstrate the performance of the proposed algorithm in a real environment.

BiSPARCs for Unsourced Random Access in Massive MIMO

This paper considers the massive MIMO unsourced random access problem in a quasi-static Rayleigh fading setting. The proposed coding scheme is based on a concatenation of a "conventional" channel code (such as, e.g., LDPC) serving as an outer code, and a sparse regression code (SPARC) serving as an inner code. The scheme combines channel estimation, single-user decoding, and successive interference cancellation in a novel way. The receiver performs joint channel estimation and SPARC decoding via an instance of a bilinear generalized approximate message passing (BiGAMP) based algorithm, which leverages the intrinsic bilinear structure that arises in the considered communication regime. The detection step is followed by a per-user soft-input-soft-output (SISO) decoding of the outer channel code in combination with a successive interference cancellation (SIC) step. We show via numerical simulation that the resulting scheme achieves stat-of-the-art performance in the massive connectivity setting, while attaining comparatively low implementation complexity.

Characterization of the weak Pareto boundary of resource allocation problems in wireless networks -- Implications to cell-less systems

We establish necessary and sufficient conditions for a network configuration to provide utilities that are both fair and efficient in a well-defined sense. To cover as many applications as possible with a unified framework, we consider utilities defined in an axiomatic way, and the constraints imposed on the feasible network configurations are expressed with a single inequality involving a monotone norm. In this setting, we prove that a necessary and sufficient condition to obtain network configurations that are efficient in the weak Pareto sense is to select configurations attaining equality in the monotone norm constraint. Furthermore, for a given configuration satisfying this equality, we characterize a criterion for which the configuration can be considered fair for the active links. We illustrate potential implications of the theoretical findings by presenting, for the first time, a simple parametrization based on power vectors of achievable rate regions in modern cell-less systems subject to practical impairments.

The Story of QoS Prediction in Vehicular Communication: From Radio Environment Statistics to Network-Access Throughput Prediction

As cellular networks evolve towards the 6th Generation (6G), Machine Learning (ML) is seen as a key enabling technology to improve the capabilities of the network. ML provides a methodology for predictive systems, which, in turn, can make networks become proactive. This proactive behavior of the network can be leveraged to sustain, for example, a specific Quality of Service (QoS) requirement. With predictive Quality of Service (pQoS), a wide variety of new use cases, both safety- and entertainment-related, are emerging, especially in the automotive sector. Therefore, in this work, we consider maximum throughput prediction enhancing, for example, streaming or HD mapping applications. We discuss the entire ML workflow highlighting less regarded aspects such as the detailed sampling procedures, the in-depth analysis of the dataset characteristics, the effects of splits in the provided results, and the data availability. Reliable ML models need to face a lot of challenges during their lifecycle. We highlight how confidence can be built on ML technologies by better understanding the underlying characteristics of the collected data. We discuss feature engineering and the effects of different splits for the training processes, showcasing that random splits might overestimate performance by more than twofold. Moreover, we investigate diverse sets of input features, where network information proved to be most effective, cutting the error by half. Part of our contribution is the validation of multiple ML models within diverse scenarios. We also use Explainable AI (XAI) to show that ML can learn underlying principles of wireless networks without being explicitly programmed. Our data is collected from a deployed network that was under full control of the measurement team and covered different vehicular scenarios and radio environments.

Machine Learning-based Methods for Reconfigurable Antenna Mode Selection in MIMO Systems

MIMO technology has enabled spatial multiple access and has provided a higher system spectral efficiency (SE). However, this technology has some drawbacks, such as the high number of RF chains that increases complexity in the system. One of the solutions to this problem can be to employ reconfigurable antennas (RAs) that can support different radiation patterns during transmission to provide similar performance with fewer RF chains. In this regard, the system aims to maximize the SE with respect to optimum beamforming design and RA mode selection. Due to the non-convexity of this problem, we propose machine learning-based methods for RA antenna mode selection in both dynamic and static scenarios. In the static scenario, we present how to solve the RA mode selection problem, an integer optimization problem in nature, via deep convolutional neural networks (DCNN). A Multi-Armed-bandit (MAB) consisting of offline and online training is employed for the dynamic RA state selection. For the proposed MAB, the computational complexity of the optimization problem is reduced. Finally, the proposed methods in both dynamic and static scenarios are compared with exhaustive search and random selection methods.

MIMO Systems with Reconfigurable Antennas: Joint Channel Estimation and Mode Selection

Reconfigurable antennas (RAs) are a promising technology to enhance the capacity and coverage of wireless communication systems. However, RA systems have two major challenges: (i) High computational complexity of mode selection, and (ii) High overhead of channel estimation for all modes. In this paper, we develop a low-complexity iterative mode selection algorithm for data transmission in an RA-MIMO system. Furthermore, we study channel estimation of an RA multi-user MIMO system. However, given the coherence time, it is challenging to estimate channels of all modes. We propose a mode selection scheme to select a subset of modes, train channels for the selected subset, and predict channels for the remaining modes. In addition, we propose a prediction scheme based on pattern correlation between modes. Representative simulation results demonstrate the system's channel estimation error and achievable sum-rate for various selected modes and different signal-to-noise ratios (SNRs).