Ilario Filippini

Growing congestion in current mobile networks necessitates innovative solutions. This paper contributes a novel network planning approach for mmWave 5G networks in urban settings, focusing on Integrated Access and Backhaul (IAB) and the Smart Radio Environment (SRE). The mmWave traffic will be mainly made of short bursts to transfer large volumes of data and long idle periods where data are processed, necessitating changes in how mobile radio networks are designed. Our proposed optimization models integrate IAB with SRE technologies while leveraging the maximization of achievable peak throughput rather than conventional average throughput metrics. Results highlight the advantages of this approach during the network planning phase, providing insights into better accommodating the demands of mobile traffic without sacrificing the overall network capacity.

Smart Radio Environment (SRE) is a central paradigm in 6 G and beyond, where integrating Smart Radio Environment (SRE) components into the network planning process enables optimized performance for high-frequency Radio Access Network (RAN). This paper presents a comprehensive planning framework utilizing realistic urban scenarios and channel models to analyze diverse SRE components, including Reconfigurable Intelligent Surface (RIS), Network-Controlled Repeater (NCR), and advanced technologies like Simultaneous Transmitting and Reflecting RIS (STAR-RIS) and Trisectoral NCR (3SNCR). We propose two optimization strategies—Full Coverage Minimum Cost (FCMC) and Maximum Budget-Constrained Coverage (MBCC)—that address key cost and coverage objectives by considering both physical characteristics and scalable costs of each component, influenced by factors such as NCR amplification gain and RIS dimensions. Extensive numerical results demonstrate the significant impact of these models in enhancing network planning efficiency for high-density urban environments.

The increasing demand for mobile bandwidth is driving 5G networks toward the use of high-frequency spectrum, particularly the upper-6 GHz and mmWave bands. While these bands offer vast bandwidth potential, their propagation characteristics raise critical deployment challenges. This paper presents the first direct, on-field comparative evaluation of 5G standalone (SA) macro-cell deployments operating in these two bands, conducted in Milan, Italy. We show that the upper-6 GHz band can deliver wide-area urban coverage (up to 600 meters) with stable gigabit-level downlink throughput, even in Non-Line of Sight (NLoS) scenarios. mmWave, traditionally deemed unsuitable for NLoS, exhibits strong performance via urban reflections, achieving up to 1.3 Gbps in downlink and 250 Mbps in uplink. Furthermore, outdoor-to-indoor connectivity at mmWave frequencies proves viable through glass facades, challenging pessimistic assumptions about penetration losses. These findings, derived from synchronized deployments and extensive measurements, provide new insights into the complementary roles of these bands and the practical feasibility of their integration into future 5G networks.

MmWaves have been envisioned as a promising direction to provide Gbps wireless access. However, they are susceptible to high path losses and blockages, which can only be partially mitigated by directional antennas. That makes mmWave networks coverage-limited, thus requiring dense deployments. Integrated access and backhaul (IAB) architectures have emerged as a cost-effective solution for network densification. Resource allocation in mmWave IAB networks must face big challenges originated by heavy temporal dynamics, such as intermittent links caused by user mobility and blockages from moving obstacles. This makes it extremely difficult to find optimal and adaptive solutions. In this article, exploiting the distributed structure of the problem, we propose a Multi-Agent Reinforcement Learning (MARL) framework to optimize user throughput via flow routing and link scheduling in mmWave IAB networks characterized by mobile users and obstacles. The proposed approach implicitly captures the environment dynamics, coordinates the interference, and manages the buffer levels of IAB relay nodes. We design different MARL components, respectively for full-duplex and half-duplex networks. In addition, we propose an online training algorithm, which addresses the feasibility issues of practical systems, especially the communication and coordination among RL agents. Numerical results show the effectiveness of the proposed approach.

MmWaves have been envisioned as a promising direction to provide Gbps wireless access. However, they are susceptible to high path losses and blockages, which directional antennas can only partially mitigate. That makes mmWave networks coverage-limited, thus requiring dense deployments. Integrated access and backhaul (IAB) architectures have emerged as a cost-effective solution for network densification. Resource allocation in mmWave IAB networks must face big challenges to cope with heavy temporal dynamics, such as intermittent links caused by user mobility and blockages from moving obstacles. This makes it extremely difficult to find optimal and adaptive solutions. In this article, exploiting the distributed structure of the problem, we propose a Multi-Agent Reinforcement Learning (MARL) framework to optimize user throughput via flow routing and link scheduling in mmWave IAB networks characterized by user mobility and link outages generated by moving obstacles. The proposed approach implicitly captures the environment dynamics, coordinates the interference, and manages the buffer levels of IAB relay nodes. We design different MARL components, considering full-duplex and half-duplex IAB-nodes. In addition, we provide a communication and coordination scheme for RL agents in an online training framework, addressing the feasibility issues of practical systems. Numerical results show the effectiveness of the proposed approach.

Thanks to its capability of manipulating electromagnetic signals, the reconfigurable intelligent surface (RIS) is gaining momentum in alleviating the impact of blockages on mmWave signals, by providing redirected transmission paths. However, obstacles can also inevitably appear in the redirected paths. This can be solved by installing multiple RISs and switching among them. In this letter, for the first time, we adaptively switch among RISs for a mobile user in real time to optimize its achievable rate, without need for a priori knowledge on potential obstacles. We present an actor-critic based approach to learn unknown obstacles and variational spatial correlations originated by the user mobility, which is followed by the analysis on ergodic achievable rate. Experimental results have shown that the approach can achieve rates about 15% less than the optimum and 76% more than the state-of-the-art.

In a world where positioning is becoming increasingly important and emerging applications have progressively tighter requirements, the current satellite-based positioning solutions can not address their needs. We propose a new network-based positioning approach based on O-RAN 5G. The solution is easily deployable, transparent to the users, and flexible to the application needs, thanks to the programmability of the network. This is the first step towards a positioning network service targeting a flexible accuracy level, minimizing the number of parameters, and optimizing the resource and computation utilization.

In the never-ending pursuit of bandwidth, mid-bands have been recently reconsidered as the primary spectrum for 5G NR and its evolution. Laying between the crowded lower frequency bands and the propagation-unfriendly higher bands (i.e., millimeter wave), the upper 6 GHz band (6425-7125 MHz) presents a valuable compromise between capacity and coverage. Recognizing its potential, the research community has expressed interest in this spectrum and conducted several studies. Despite this enthusiasm, the deployment of upper 6 GHz testbeds remains elusive. This paper aims to address this gap by presenting the results of a comprehensive measurement campaign conducted on a 5G NR cellular system operating in the upper 6 GHz band (6425-7125 MHz), specifically deployed within the Politecnico di Milano campus. Our objective was to evaluate the system's performance in a realistic environment and provide insights supported by empirical measurements. The measurement campaign yielded positive results, showcasing a remarkable channel capacity in urban areas, which remained consistently high even at the cell edge and in challenging non-line-of-sight and outdoor-to-indoor scenarios.

Millimeter-Wave (mm-Wave) Radio Access Networks (RANs) are a promising solution to tackle the overcrowding of the sub-6 GHz spectrum, offering wider and underutilized bands. However, they are characterized by inherent technical challenges, such as a limited propagation range and blockage losses caused by obstacles. Integrated Access and Backhaul (IAB) and Reconfigurable Intelligent Surfaces (RIS) are two technologies devised to face these challenges. This work analyzes the optimal network layout of RANs equipped with IAB and RIS in real urban scenarios using MILP formulations to derive practical design guidelines. In particular, it shows how optimizing the peak user throughput of such networks improves the achievable peak throughput, compared to the traditional mean-throughput maximization approaches, without actually sacrificing mean throughputs. In addition, it indicates star-like topologies as the best network layout to achieve the highest peak throughputs.

The introduction of the mm-Wave spectrum into 5G NR promises to bring about unprecedented data throughput to future mobile wireless networks but comes with several challenges. Network densification has been proposed as a viable solution to increase RAN resilience, and the newly introduced Integrated-Access-and-Backhaul (IAB) is considered a key enabling technology with compelling cost-reducing opportunities for such dense deployments. Reconfigurable Intelligent Surfaces (RIS) have recently gained extreme popularity as they can create Smart Radio Environments by EM wave manipulation and behave as inexpensive passive relays. However, it is not yet clear what role this technology can play in a large RAN deployment. With the scope of filling this gap, we study the blockage resilience of realistic mm-Wave RAN deployments that use IAB and RIS. The RAN layouts have been optimised by means of a novel mm-Wave planning tool based on MILP formulation. Numerical results show how adding RISs to IAB deployments can provide high blockage resistance levels while significantly reducing the overall network planning cost.