Francesco Marcuzzi’s research while affiliated with Alpen-Adria-Universität Klagenfurt and other places

What is this page?

This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (11)

FIGURE 5. This collection of graphs show the influence of the Modified Distribution Approach on the distribution of the number of nodes per sector. This intuitive and heuristic approach to the generation of a network allows to reduce considerably the variance of the final result. The "empirical" label on the legend of each graph identifies the curve of the original distribution where network generation is carried out with the parameters shown in Table 1.
FIGURE 6. This collection of graphs show the influence of the Modified Distribution Approach on the generated networks. The underlying grid in each graph shows 200 meter divides. In this representation of the network, the nodes were removed for simplicity of visualization, although the LPs are still present to account for varying density.
FIGURE 13. This graph shows the performance of our DFN trained on inferring the SNR of a link on a single, specific PLC network. Density was described with 10 quantiles. This configuration shows an improved performance w.r.t. to the unseen networks case -an unchanging topology reduces the degrees of freedom with which the topological parameters can change (e.g., number of nodes is always the same). By limiting the topology variability, we obtain more precise results.
Values for the processes used to create a random LV power line network.
Values for the processes used to create a random LV power line network in the modified distributions approach: size.

+1

Topology-Based Machine Learning: Predicting Power Line Communication Quality in Smart Grids
  • Article
  • Full-text available

January 2023

·

80 Reads

·

11 Citations

IEEE Access

Francesco Marcuzzi

·

Smart Grids (SG) envision the exchange of both power and data, enabling system and customers to generate and transfer energy in a more efficient and balanced way. Among the relevant communication technologies, we find Power-Line Communications (PLC), which allow for data transmission on the electrical cables used for power delivery. Despite the hostile medium, PLC offer reliability and data rates to support exchange of control traffic, smart metering and polling applications. The distribution portion of the power delivery network, which we focus on in this work, is the most topologically complex, which makes channel prediction complicated. We show how random realistic topologies can be generated and then used to train a Machine Learning (ML) algorithm to infer PLC link quality (based on channel response) based solely on topology descriptors.We eventually show how precisely the communication quality can be inferred from the SG topology through ML. In doing so, we also discuss how the ML approach offers the common ground between top-down and bottom-up approaches for network characterization and how it enables smart decision making in the SG.

Download
FIGURE 1. Taxonomy of learning models.
FIGURE 2. Structure of a fully connected neural network with 2 hidden layers.
FIGURE 3. Decision boundaries and maximum separation for SVM framework.
FIGURE 4. Autoencoder block with neural networks. A latent representation z is learned by the encoder and used as input of the decoder for recovering x.
FIGURE 5. GAN framework in which the generator and discriminator are learned during the training process.

+8

Machine Learning Tips and Tricks for Power Line Communications

June 2019

·

3,544 Reads

·

42 Citations

IEEE Access

·

·

Davide Righini

·

Francesco Marcuzzi

A great deal of attention has been recently given to Machine Learning (ML) techniques in many different application fields. This paper provides a vision of what ML can do in Power Line Communications (PLC). We first and briefly describe classical formulations of the ML, and distinguish deterministic from statistical learning models with relevance to communications. We then discuss ML applications in PLC for each layer, namely, for characterization and modeling, for the development of physical layer algorithms, for media access control and networking. Finally, other applications of the PLC that can benefit from the usage of ML, as grid diagnostics, are analyzed. Illustrative numerical examples are reported to serve the purpose of validating the ideas and motivate future research endeavors in this stimulating signal/data processing field.

Download
Machine Learning Tips and Tricks for Power Line Communications

April 2019

·

138 Reads

·

·

Davide Righini

·

Francesco Marcuzzi

A great deal of attention has been recently given to Machine Learning (ML) techniques in many different application fields. This paper provides a vision of what ML can do in Power Line Communications (PLC). We firstly and briefly describe classical formulations of ML, and distinguish deterministic problems from statistical problems with relevance to communications. We then discuss ML applications in PLC for each layer, namely, for characterization and modeling, for physical layer algorithms, for media access control and networking algorithms. Finally, other applications of PLC that can benefit from the usage of ML, as grid diagnostics, are analyzed. Illustrative numerical examples are reported to serve the purpose of validating the ideas and motivate future research endeavors in this stimulating signal/data processing field.

Smart Routing for Improved PLC Backhauling of the Radio Access Network

February 2018

·

23 Reads

·

1 Citation

Francesco Marcuzzi

·

Transmission line theory enables the bottom up study of networks based on wireline infrastructures. This technique is here applied to a simulator that brings together powerline communication networks with small radio cells ones in a hybrid paradigm: this allows to implement a study of channel capacity and communication quality based on the geometrical properties of the network and the type of cables employed to enable connectivity. Results are shown regarding how deterministic properties related to performance in the network can be retrieved and used to enable a smart routing algorithm.

Citations (8)

... Actually, NNTGM, which is the new member of the neural network identification methodology (NNIM) -based family products; say, neural network identification methodology for the branch number identification (NNIM-BNI) [2] and neural network identification methodology for the line length approximation (NNIM-LLA) [3][4][5], enhances the speed of populating OV LV BPL topology classes, thus facilitating better planning and management of OV LV BPL networks. Exploiting artificial intelligence (AI), machine learning (ML) and neural network (NN) features, the ability of NNTGM to create virtual topologies that imitate the characteristics of real-world topologies may allow communications engineers to predict and mitigate potential issues, leading to more robust and resilient communication systems [6][7][8]. Therefore, the adoption of NNTGM for populating OV LV BPL topology classes offers a significant advancement in the field of BPL networks, ensuring efficient and reliable network operation. ...

Reference:

Virtual Topologies for Populating Overhead Low-Voltage Broadband over Powerlines Topology Classes by Exploiting Neural Network Topology Generator Methodology (NNTGM) - Part 2: Numerical Results
Topology-Based Machine Learning: Predicting Power Line Communication Quality in Smart Grids

IEEE Access

Francesco Marcuzzi

·

... Initially, the aim of this work was to use the datasets of real PLC access network deployment provided by an utility to further the knowledge about PLC networks and to analyze the relation between topology and network performance. The non-trivial task of analyzing this dependency brought us to use the common characteristics of different networks to define a unified topology model (main characteristics are introduced in [23]). This general model encompasses all the observed real deployments and allows for both physical and higher layer modeling. ...

Reference:

Topology-Based Machine Learning: Predicting Power Line Communication Quality in Smart Grids
Discovering Routing Anomalies in Large PLC Metering Deployments from Field Data
  • Citing Conference Paper

  • May 2020

Francesco Marcuzzi

·

·

... The load density has been shown to strongly influence the signal attenuation due to high interference because of presence of reflecting elements [9]. A fundamental requirement VOLUME 4, 2016 Table 1. in designing a measure of load density is that it needs to represent effectively networks of different size -even though it would be possible to specify density as a number of nodes per area unit, area of a network is not easily definable, as the represented topology takes into consideration cable lengths, but not real, geographical node location. ...

Reference:

Topology-Based Machine Learning: Predicting Power Line Communication Quality in Smart Grids
Artificial-Intelligence-Based Performance Enhancement of the G3-PLC LOADng Routing Protocol for Sensor Networks
  • Citing Conference Paper

  • April 2019

Francesco Marcuzzi

·

... However, installation of fiber optic based front-hauls for indoor services could be expensive if not impossible, due to the requirement for infrastructural alteration. On the other hand, power line communication (PLC) has been deployed for broadband data transmissions for various applications such as home entertainment and internet services [4]. The ubiquity of the residential power lines has made PLC as a promising candidate for any indoor wireless system. ...

Reference:

On CPRI based Front-hauling over Residential Power Lines
Radio Access Network Backhauling Using Power Line Communications
Francesco Marcuzzi

·

... Bu yaklaşım; istatistiksel yaklaşımda gereken çok sayıda ölçüm alma gerekliliğini ortadan kaldırdığı gibi deterministik yaklaşımda ki hesap karmaşasının da önüne geçer. Daha açık ifadeyle yinelemeli bir yaklaşım metodu olduğundan çok yollu bileşenlerin ayrı ayrı hesaplanma gereğini de ortadan kaldırır [18,20]. Şekil 6'da grafik tabanlı yaklaşımın işlem adımları verilmiştir. ...

Reference:

GÜÇ HATTI HABERLEŞME KANALININ KARAKTERİZASYONU VE MODELLENMESİ CHANNEL CHARACTERIZATION AND MODELING OF POWERLINE COMMUNICATION
Artificial intelligence based routing in PLC networks
  • Citing Conference Paper

  • April 2018

Francesco Marcuzzi

·

... To mitigate instability of LoRa in complex indoor environments and the building-bound limitations of PLC, some researchers have explored hybrid networks. In 2017, Alireza Ghiasimonfared et al. [8] proposed a dual-hop PLC/WLC network that integrates LoRa and G3-PLC technologies. In their approach, the PLC link facilitates communication through obstacles such as walls, while the wireless link aggregates multiple PLC sub-networks across physically isolated areas. ...

Reference:

Efficient Large-Scale IoT Network: Integrating Asynchronous Communication and Huffman Coding in LoRa/PLC Systems
Development of a hybrid LoRa/G3-PLC IoT sensing network: An application oriented approach
  • Citing Conference Paper

  • October 2017

Alireza Ghiasimonfared

·

Davide Righini

·

Francesco Marcuzzi

·