Gianmarco Baldini

The knowledge of the Line of Sight (LOS) or Not Line of Sight (NLOS) propagation condition is useful information in wireless communication system. Such knowledge can be inferred by the analysis of the signal, by using specific signal structures (e.g., preambles) or by the application of machine learning (ML) algorithms. In recent times, deep learning (DL) has been applied with success to the classification of LOS/NLOS conditions but with a significant computational time, which can be a practical issue in computing constrained devices. On the other hand, ML relies on the identification of key discriminating features, which can enhance the classification performance. This paper explores the application of entropy metrics to this classification problem. Beyond Shannon entropy, researchers have developed various entropy metrics in recent years in various domains (e.g., healthcare), but they have been scarcely applied to UWB LOS/NLOS classification to the best of the author’s knowledge. This paper addresses this gap by applying entropy metrics in combination with ML classifiers to the public eWINE dataset, characterised by seven different propagation environments where UWB signals were transmitted and recorded in LOS and NLOS conditions. The results presented in this paper show that entropy metrics can significantly enhance the LOS/NLOS classification accuracy and can produce an overall competitive performance. In addition, this paper presents a novel instance selection approach based on the use of entropy metrics, which is demonstrated to significantly outperform even the direct application of some DL algorithms on the basis of the results presented in the literature on the same eWine data set. To summarise the novelty aspects of this study, for the first time in the literature, this study presents an extensive analysis of the discriminative advantage (discrimination index) of entropy measures introduced in the research literature in other domains (e.g., mechanical problems, analysis of physiological signals) in UWB multipath environments for UWB LOS/NLOS classification. In addition, this study presents for the first time the application of an ensemble instance selection algorithm based on entropy measures to the problem of UWB LOS/NLOS classification to handle “noise” or “boundary” samples in the data set, thereby improving model generalisation.

Global Navigation Satellites Systems (GNSSs) have become a predominant feature in the digital life of citizens, and they provide positioning services in various applications including pedestrian and vehicular navigation. In urban environments with the presence of buildings and another obstacles, GNSS positioning may be unreliable because of non-line-of-sight (NLOS) conditions, and the classification of observed satellite visibility between LOS and NLOS may improve GNSS receivers to improve their performance to provide the positioning services. In this context, machine learning algorithms using features like signal noise ratio, pseudorange, elevation angle, and others have been applied to this problem both in supervised and unsupervised mode. Because the ground truth information on LOS/NLOS conditions may not always be available, unclustering algorithms have been applied for unsupervised classification, but the classification performance is still limited. This paper proposes an ensemble approach where different clustering algorithms, both historical and recently introduced in the literature, are combined to improve the LOS/NLOS classification accuracy. Even if the ensemble approach manages to achieve a significant improvement, a novel and more sophisticated approach is proposed in this paper, where the contributions of each clustering algorithm are weighted. The optimal values of the weights are estimated using various Meta-Heuristics Algorithms (MHA) on a subset of GNSS data where the ground-truth information is available (i.e., labeled data set). In a subsequent step, the performance of the optimal weighted clustering ensemble is evaluated. The approach is applied to a recent public data set with 57 satellites, where it is shown to outperform the specific clustering approaches by a large margin (more than 7%). The Meta Heuristics Algorithm (MHA)s have similar performance, with the Dynamic Opposition Grey Wolf Optimization (DOLGWO) having a minor advantage against the other MHAs.

Radio frequency fingerprinting (RFF) is a technique for identifying wireless devices based on their intrinsic physical characteristics. In recent years, a variety of techniques have been proposed for this problem, including the use of convolutional neural networks (CNN) applied to the images created by applying Time‐Frequency (TF) transforms to the signals emitted by the wireless devices. The main contribution of this letter is the novel application of CNN in combination with the Singular Superlet Transform (SSLT), which has not yet been introduced in the RFF literature. The SSLT is a recently introduced TF superresolution representation that consists of increasingly bandwidth‐constrained wavelets and is particularly suitable for transient signals. In the proposed approach, the original RFF signal is converted into images that are input to a CNN with a multi‐head attention layer. The letter evaluates the SSLT‐based approach on two different public RFF datasets of wireless devices used in the Internet of Things (IoT) for different values of Signal‐to‐Noise Ratio (SNR). The results show that the proposed approach is able to outperform other combinations of CNN and TF transforms used in the literature for RFF, and it is particularly robust in the presence of noise.

The presence of Internet of Things (IoT) devices in modern working and living environments is growing rapidly. The data collected in such environments enable us to model users’ behaviour and consequently identify and authenticate them. However, these data may contain information about the user’s current activity, emotional state, or other aspects that are not relevant for authentication. In this work, we employ adversarial deep learning techniques to remove privacy-revealing information from the data while keeping the authentication performance levels almost intact. Furthermore, we develop and apply various techniques to offload the computationally weak edge devices that are part of the machine learning pipeline at training and inference time. Our experiments, conducted on two multimodal IoT datasets, show that P²ESA can be efficiently deployed and trained, and with user identification rates of between 75.85% and 93.31% (c.f. 6.67% baseline), can represent a promising support solution for authentication, while simultaneously fully obfuscating sensitive information.

In recent times, neural networks have been applied to the classification of different types of anomalies in optical networks. On the other side, the evaluation of the severity of a specific anomaly has been scarcely investigated as this may be a more challenging problem due to the presence of noise in the data collected by the sensors. This letter proposes a hybrid machine learning/CNN approach, where the presence of noise is mitigated by a pre-processing step based on an adaptive smoothing algorithm and an ensemble of transforms to generate a feature space given in input to a CNN. The approach is applied to a recent public data set with sensor data collected from a real fiber optical network for the fiber bending anomaly, where it is shown to outperform the direct application of CNN on the original sensor data in the time domain.

The performance of the services provided by Global Navigation Satellite Systems (GNSSs) can be seriously degraded by the presence of wireless interferences, and Machine Learning (ML) has been applied to address this problem using the digital artifacts generated by the GNSS receiver. While such an application is not novel in the literature, the analysis of the impact of the bit-depth at which the GNSS signal is recorded has not received significant attention. The type and power level of the wireless interference are also important factors to investigate in this context. This paper addresses this gap by performing an extensive analysis of the impact of these factors on a data set of GNSS signals subject to three different types of wireless interferences with ML and DL algorithms. The analysis is a combination of a pre-processing phase where the Carrier-to-Noise Ratio (CNR) values of different satellites are evaluated, the extraction of relevant features for ML, and the application of a Convolutional Neural Network (CNN) with a multi-head attention layer. The results show that the proposed approach is able to detect the presence of interference with great accuracy (e.g., 99%) but the type of interference and bit-depth can decrease the performance.

The classification of the wireless propagation channel between Line-of-Sight (LOS) or Non-Line-of-Sight (NLOS) is useful in the operation of wireless communication systems. The research community has increasingly investigated the application of machine learning (ML) to LOS/NLOS classification and this paper is part of this trend, but not all the different aspects of ML have been analyzed. In the general ML domain, poisoning and adversarial attacks and the related mitigation techniques are an active area of research. Such attacks aim to hamper the ML classification process by poisoning the data set. Mitigation techniques are designed to counter this threat using different approaches. Poisoning attacks in LOS/NLOS classification have not received significant attention by the wireless communication community and this paper aims to address this gap by proposing the application of a specific mitigation technique based on outlier detection algorithms. The rationale is that poisoned samples can be identified as outliers from legitimate samples. In particular, the study described in this paper proposes a recent outlier detection algorithm, which has low computing complexity: the sparse data observers (SDOs) algorithm. The study proposes a comprehensive analysis of both conventional and novel types of attacks and related mitigation techniques based on outlier detection algorithms for UltraWideBand (UWB) channel classification. The proposed techniques are applied to two data sets: the public eWINE data set with seven different UWB LOS/NLOS different environments and a radar data set with the LOS/NLOS condition. The results show that the SDO algorithm outperforms other outlier detection algorithms for attack detection like the isolation forest (iForest) algorithm and the one-class support vector machine (OCSVM) in most of the scenarios and attacks, and it is quite competitive in the task of increasing the UWB LOS/NLOS classification accuracy through sanitation in comparison to the poisoned model.

5G delivers limitless mobile broadband capabilities, enabling extensive connectivity for various devices, including smartphones, sensors, and machines. Its most notable feature lies in its ability to facilitate instantaneous and highly reliable machine communications. Because of the rapid evolution of communication networks, innovative architectures such as Open RAN (RAN) have emerged, promising increased flexibility, efficiency, and sustainability. To realize the full potential of ORAN networks, however, it is critical to understand their power consumption implications. In this paper, in order to understand the way O-RAN consumes energy, a comprehensive power consumption model is used. The study not only provides theoretical analysis but also incorporates a practical implementation of the architecture. For this, an algorithm is employed to determine a base monolithic architecture of O-RAN by the placement of Distributed Unit (DU) and Centralized Unit (CU) using available data of Radio Unit (RU) in France and Ireland. The research further explores the impact of different parameter variations in various Evaluation Scenarios (ESs) on the architecture. By conducting this thorough evaluation, this research provides valuable insights for researchers and industry practitioners to gain a deeper understanding of energy consumption patterns and derive ideas on how to effectively reduce power consumption in these networks.

Intrusion Detection Systems (IDSs) are an important tool to mitigate cybersecurity threats in the ICT infrastructures. Preferable properties of the IDSs are the optimization of the attack detection accuracy and the minimization of the computing resources and time. A signification portion of IDSs presented in the research literature is based on Machine Learning (ML) and Deep Learning (DL) elements, but they may be prone to adversarial attacks, which may undermine the overall performance of the IDS algorithm. This paper proposes a novel IDS focused on the detection of cybersecurity attacks in 5G networks, which addresses in a simple but effective way two specific adversarial attacks: (1) tampering of the labeled set used to train the ML algorithm, (2) modification of the features in the training data set. The approach is based on the combination of two algorithms, which have been introduced recently in the research literature. The first algorithm is the Extremely Randomized Tree (ERT) algorithm, which enhances the capability of Decision Tree (DT) and Random Forest (RF) algorithms to perform classification in data sets, which are unbalanced and of large size as IDS data sets usually are (legitimate traffic messages are more numerous than attack related messages). The second algorithm is the recently introduced Infinite Feature Selection algorithm, which is used to optimize the choice of the hyper-parameter defined in the approach and improve the overall computing efficiency. The result of the application of the proposed approach on a recently published 5G IDS data set proves its robustness against adversarial attacks with different degrees of severity calculated as the percentage of the tampered data set samples.