The transition to AI-native 6G networks embeds artificial intelligence (AI) across all network functions, enabling automation, efficiency, and advanced services. The AI-native approach is more than just a technical enhancement; it is a fundamental shift that drives the evolution toward autonomous networks, unlocking new advancements in network capabilities and user experience. A key enabler is AI as a Service (AIaaS), which allows Communication Service Providers (CSPs) to expose AI capabilities via standardized APIs. Built on Machine Learning Operations (MLOps), AIaaS streamlines AI model lifecycle management, reducing complexity for developers and optimizing AI-driven workflows.This paper investigates the integration of AIaaS into network exposure frameworks such as the 3rd Generation Partnership Project (3GPP) Network Exposure Function (NEF), Service Enabler Architecture Layer (SEAL), and the open-source CAMARA interface emphasizing the interacting and collaborating robots (cobots) use case in the context of industrial automation. We discuss how AIaaS enables CSPs to offer AI-powered services with guaranteed Quality of Service (QoS), driving innovation across industries and shaping the future of AI-native 6G networks.

Federated learning enables training a global model from data located at the client nodes, without data sharing and moving client data to a centralized server. Performance of federated learning in a multi-access edge computing (MEC) network suffers from slow convergence due to heterogeneity and stochastic fluctuations in compute power and communication link qualities across clients. We propose a novel coded computing framework, CodedFedL, that injects structured coding redundancy into federated learning for mitigating stragglers and speeding up the training procedure. CodedFedL enables coded computing for non-linear federated learning by efficiently exploiting distributed kernel embedding via random Fourier features that transforms the training task into computationally favourable distributed linear regression. Furthermore, clients generate local parity datasets by coding over their local datasets, while the server combines them to obtain the global parity dataset. Gradient from the global parity dataset compensates for straggling gradients during training, and thereby speeds up convergence. For minimizing the epoch deadline time at the MEC server, we provide a tractable approach for finding the amount of coding redundancy and the number of local data points that a client processes during training, by exploiting the statistical properties of compute as well as communication delays. We also characterize the leakage in data privacy when clients share their local parity datasets with the server. Additionally, we analyze the convergence rate and iteration complexity of CodedFedL under simplifying assumptions, by treating CodedFedL as a stochastic gradient descent algorithm. Finally, for demonstrating gains that CodedFedL can achieve in practice, we conduct numerical experiments using practical network parameters and benchmark datasets, in which CodedFedL speeds up the overall training time by up to 15× in comparison to the benchmark schemes.

Spatial correlation has been recognized by 3GPP as one of the key elements in millimeter-wave (mmWave) channel modeling. Correlated channel behavior is induced by macro objects, such as buildings, as well as by micro objects, including humans around the mmWave receivers. The 3GPP's three-dimensional (3D) spatially consistent channel model designed to capture these phenomena assumes a-priori knowledge of the correlation distance between the receivers. In this paper, we propose a novel spatially-consistent human body blockage state generation procedure, which extends the standardized 3D channel model by 3GPP to capture the correlation between the line-of-sight (LoS) links and the reflected cluster states affected by human body blockage. The proposed model is based on analytical expressions for the conditional link state probability, thus permitting the parametrization of the spatial field of receivers. It also does not require any a-priori information on the correlation distance as the latter is identified explicitly based on the environmental parameters. We compare the results for the proposed model with those obtained with the uncorrelated blockage model and conclude that in many special cases correlation manifests itself in quantitatively different propagation conditions experienced at the nearby receivers.

Millimeter wave (mmWave) communication is a key enabler for realizing the data rates, capacity and latency requirements envisioned for emerging 5G systems. This paper addresses the system performance of 73 GHz mmWave access, which is representative of the systems expected to be standardized in the second phase of 3GPP specifications. In particular, we address the performance scalability of mmWave systems and characterize the tradeoffs and challenges involved in improving their coverage and capacity for different cell density regimes. Our results, based on industry standard 3GPP methodology, show that the directional communication enabled through massive MIMO capability of mmWave systems, is key towards addressing the interference limitation on performance scaling, thus illustrating the promise of mmWave communication to provide a scalable solution capable of handling the needs of future ultra-dense networks.

Dense deployments of millimeter-wave (mmWave) base stations (BSs) are being considered as the most feasible solution to meet the steadily growing data rate demands of mobile users. Accordingly, the achievable performance gains of mmWave-based dense networks in real deployments have to be studied carefully, since mmWave radio technology features specific transceiver, antenna, and propagation properties. In this paper, we contribute an accurate performance evaluation of single- versus dual-polarized MIMO systems operating over the mmWave channel in typical urban scenarios as well as address the impact of device- and network-centric parameters on the performance gains enabled by MIMO in dense to ultra-dense BS deployments. This study relies on our in-house ray-based modeler and takes into account the key mmWave system effects, such as multi-path propagation, utilization of dual-polarized antennas, and characteristic interference models. Our results show that the benefit of using mmWave- MIMO grows with increasing BS density, thus encouraging a further study of this technology especially for (ultra-)dense setups. We also demonstrate that non-coherent non-polarized diffuse scattering component may reduce the capacity gain of dual-polarized vs. single- polarized MIMO.

The increasing demand of data, along with the spectrum scarcity, are motivating a urgent shift towards exploiting new bands. This is the main reason behind identifying mmWaves as the key disruptive enabling technology for 5G cellular networks. Indeed, utilizing new bands means facing new challenges; in this context, they are mainly related to the radio propagation, which is shorter in range and more sensitive to obstacles. The resulting key aspects that need to be taken into account when designing mmWave cellular systems are directionality and link intermittency. The lack of network level results motivated this work, which aims at providing the first of a kind open source mmWave framework, based on the network simulator ns-3. The main focus of this work is the modeling of customizable channel, physical (PHY) and medium access control (MAC) layers for mmWave systems. The overall design and architecture of the model are discussed in details. Finally, the validity of our proposed framework is corroborated through the simulation of a simple scenario.

With the shortage of spectrum in conventional cellular frequencies, millimeter-wave (mmWave) bands are being widely considered for use in next-generation networks. Multihop relaying is likely to play a significant role in mmWave cellular systems for self backhauling, range extension and improved robustness from path diversity. However, designing scheduling policies for these systems is challenging due to the need to account for both adaptive directional transmissions and dynamic time-division duplexing schedules, which are key enabling features of mmWave systems. This paper considers the problem of joint scheduling and congestion control in a multihop mmWave network using a Network Utility Maximization (NUM) framework. Interference is modeled with an exact model and two auxiliar simplified models: actual interference (AI), with a graph-based calculation of the Signal to Interference plus Noise Ratio (SINR) depending on dynamic link activity and directivity, as well as upper and lower bounds computed from worst-case interference (WI) and interference free (IF) approximations. Throughput and utility optimal policies are derived for all interference models (AI, WI and IF) with both deterministic Maximum Weighted and randomized Pick and Compare scheduling algorithms, jointly with decentralized Dual Congestion Control. Results are evaluated with numerical simulations, using accurate mmWave channel and beamforming gain approximations based on measurement campaigns.

With the severe spectrum shortage in conventional cellular bands, millimeter wave (mmW) frequencies between 30 and 300 GHz have been attracting growing attention as a possible candidate for next-generation micro- and picocellular wireless networks. The mmW bands offer orders of magnitude greater spectrum than current cellular allocations and enable very high-dimensional antenna arrays for further gains via beamforming and spatial multiplexing. This paper uses recent real-world measurements at 28 and 73 GHz in New York, NY, USA, to derive detailed spatial statistical models of the channels and uses these models to provide a realistic assessment of mmW micro- and picocellular networks in a dense urban deployment. Statistical models are derived for key channel parameters, including the path loss, number of spatial clusters, angular dispersion, and outage. It is found that, even in highly non-line-of-sight environments, strong signals can be detected 100-200 m from potential cell sites, potentially with multiple clusters to support spatial multiplexing. Moreover, a system simulation based on the models predicts that mmW systems can offer an order of magnitude increase in capacity over current state-of-the-art 4G cellular networks with no increase in cell density from current urban deployments.

Interference cancelation (IC) can provide significant gains in wireless networks with strong interference, that arise, for example, in emerging femto- and picocellular deployments. This paper consider the problem of optimal downlink rate selection in networks where each mobile can perform IC on up to one interferer. When mobiles are capable of IC, it is argued that rate selection can play an analogous roles as power control by permitting a tradeoff between rates on the desired link with “cancellability” on interfering links. A utility maximizing scheduler based on loopy belief propagation is presented that enables computationally-efficient local processing and low communication overhead. It is shown that the fixed points of the method are provably globally optimal for arbitrary (potentially non-convex) rate and utility functions. In addition, the result applies to an arbitrary networks where the interference is determined by a single dominant interferer, for which the IC problem is a special case. Simulations are presented in industry standard femtocellular network models.