Search

The rapid evolution of artificial intelligence (AI) technologies holds transformative potential for the healthcare sector. In critical situations requiring immediate decision-making, healthcare professionals can leverage machine learning (ML) algorithms to prioritize and optimize treatment options, thereby reducing costs and improving patient outcomes. However, the sensitive nature of healthcare data presents significant challenges in terms of privacy and data ownership, hindering data availability and the development of robust algorithms. Federated Learning (FL) addresses these challenges by enabling collaborative training of ML models without the exchange of local data. This paper introduces a novel FL platform designed to support the configuration, monitoring, and management of FL processes. This platform operates on Platform-as-a-Service (PaaS) principles and utilizes the Message Queuing Telemetry Transport (MQTT) publish-subscribe protocol. Considering the production readiness and data sensitivity inherent in clinical environments, we emphasize the security of the proposed FL architecture, addressing potential threats and proposing mitigation strategies to enhance the platform's trustworthiness. The platform has been successfully tested in various operational environments using a publicly available dataset, highlighting its benefits and confirming its efficacy.

MAC randomization is a widely used technique implemented on most modern smartphones to protect user's privacy against tracking based on Probe Request frames capture. However, there exist weaknesses in such a methodology which may still expose distinctive information, allowing to track the device generating the Probe Requests. Such techniques, known as MAC de-randomization algorithms, generally exploit Information Elements (IEs) contained in the Probe Requests and use clustering methodologies to group together frames belonging to the same device. While effective on heterogeneous device types, such techniques are not able to differentiate among devices of identical type and running the same Operating System (OS). In this paper, we propose a MAC de-randomization technique able to overcome such a weakness. First, we propose a new dataset of Probe Requests captured from devices sharing the same characteristics. Secondly, we observe that the time-frequency pattern of Probe Request emission is unique among devices and can therefore be used as a discriminative feature. We embed such a feature in a two-stage clustering methodology and show through experiments its effectiveness compared to state-of-the-art techniques based solely on IEs fingerprinting. The original dataset used in this work is made publicly available for reproducible research., Comment: Conference pre-print

The Internet of Things (IoT) has boomed in recent years, with an ever-growing number of connected devices and a corresponding exponential increase in network traffic. As a result, IoT devices have become potential witnesses of the surrounding environment and people living in it, creating a vast new source of forensic evidence. To address this need, a new field called IoT Forensics has emerged. In this paper, we present \textit{CSI Sniffer}, a tool that integrates the collection and management of Channel State Information (CSI) in Wi-Fi Access Points. CSI is a physical layer indicator that enables human sensing, including occupancy monitoring and activity recognition. After a description of the tool architecture and implementation, we demonstrate its capabilities through two application scenarios that use binary classification techniques to classify user behavior based on CSI features extracted from IoT traffic. Our results show that the proposed tool can enhance the capabilities of forensic investigations by providing additional sources of evidence. Wi-Fi Access Points integrated with \textit{CSI Sniffer} can be used by ISP or network managers to facilitate the collection of information from IoT devices and the surrounding environment. We conclude the work by analyzing the storage requirements of CSI sample collection and discussing the impact of lossy compression techniques on classification performance., Comment: Paper accepted for publication at conference Mediterranean Communication and Computer Networking Conference (MedComNet 2023)

The Internet of Things is in constant growth, with millions of devices used every day in our homes and workplaces to ease our lives. Such a strict coexistence between humans and smart devices makes the latter digital witnesses of our every-day lives through their sensor systems. This opens up to a new area of digital investigation named IoT Forensics, where digital traces produced by smart devices (network traffic, in primis) are leveraged as evidences for forensic purposes. It is therefore important to create tools able to capture, store and possibly analyse easily such digital traces to ease the job of forensic investigators. This work presents one of such tools, named Feature-Sniffer, which is thought explicitly for Wi-Fi enabled smart devices used in Smart Building/Smart Home scenarios. Feature-Sniffer is an add-on for OpenWrt-based access points and allows to easily perform online traffic feature extraction, avoiding to store large PCAP files. We present Feature-Sniffer with an accurate description of the implementation details, and we show its possible uses with practical examples for device identification and activity classification from encrypted traffic produced by IoT cameras. We release Feature-Sniffer publicly for reproducible research., Comment: Paper accepted for publication at IEEE 8th World Forum of Internet of Things (IEEE WF-IOT 2022)

Real-time measurements on the occupancy status of indoor and outdoor spaces can be exploited in many scenarios (HVAC and lighting system control, building energy optimization, allocation and reservation of spaces, etc.). Traditional systems for occupancy estimation rely on environmental sensors (CO2, temperature, humidity) or video cameras. In this paper, we depart from such traditional approaches and propose a novel occupancy estimation system which is based on the capture of Wi-Fi management packets from users' devices. The system, implemented on a low-cost ESP8266 microcontroller, leverages a supervised learning model to adapt to different spaces and transmits occupancy information through the MQTT protocol to a web-based dashboard. Experimental results demonstrate the validity of the proposed solution in four different indoor university spaces., Comment: Submitted to Balkancom 2018

The vision of the Internet of Thing is becoming a reality and novel communications technologies such as the upcoming 5G network architecture are designed to support its full deployment. In this scenario, we discuss the benefits that a publish/subscribe protocol such as MQTT or its recently proposed enhancement MQTT+ could bring into the picture. However, deploying pub/sub brokers with advanced caching and aggregation functionalities in a distributed fashion poses challenges in protocol design and management of communication resources. In this paper, we identify the main research challenges and possible solutions to scale up a pub/sub architecture for upcoming IoT applications in 5G networks, and we present our perspective on systems design, optimisation, and working implementations.

In the last few years, the Message Queueing Telemetry Transport (MQTT) publish/subscribe protocol emerged as the de facto standard communication protocol for IoT, M2M and wireless sensor networks applications. Such popularity is mainly due to the extreme simplicity of the protocol at the client side, appropriate for low-cost and resource-constrained edge devices. Other nice features include a very low protocol overhead, ideal for limited bandwidth scenarios, the support of different Quality of Services (QoS) and many others. However, when an edge device is interested in performing processing operations over the data published by multiple clients, the use of MQTT may result in high network bandwidth usage and high energy consumption for the end devices, which is unacceptable in resource constrained scenarios. To overcome these issues, we propose in this paper MQTT+, which provides an enhanced protocol syntax and enrich the pub/sub broker with data filtering, processing and aggregation functionalities. MQTT+ is implemented starting from an open source MQTT broker and evaluated in different application scenarios.

We show that subsets of interacting oscillators may synchronize in different ways within a single network. This diversity of synchronization patterns is promoted by increasing the heterogeneous distribution of coupling weights and/or asymmetries in small networks. We also analyze consistency, defined as the persistence of coexistent synchronization patterns regardless of the initial conditions. Our results show that complex weighted networks display richer consistency than regular networks, suggesting why certain functional network topologies are often constructed when experimental data are analyzed., Comment: 9 pages, 6 figures

We introduce new algebro-topological invariants of directed networks, based on the topological construction of the directed clique complex. The shape of the underlying directed graph is encoded in a way that can be studied mathematically to obtain network invariants such as the Euler characteristic and the Betti numbers. Two different cases illustrate the application of the Euler characteristic. We investigate how the evolution of a Boolean recurrent artificial neural network is influenced by its topology in a dynamics involving pruning and strengthening of the connections, and to show that the topological features of the directed clique complex influence the dynamical evolution of the network. The second application considers the directed clique complex in a broader framework, to define an invariant of directed networks, the network degree invariant, which is constructed by computing the topological invariant on a sequence of sub-networks filtered by the minimum in- or out-degree of the nodes. The application of the Euler characteristic presented here can be extended to any directed network and provides a new method for the assessment of specific functional features associated with the network topology., Comment: 13 pages, 3 figures. Figures and captions improved, typographical changes

Ongoing research in buffet loads alleviation has provided an application for recently developed piezoelectric actuators capable of higher force output than previously existing actuators could provide and that can be embedded within the vehicle s structure. These new actuators, having interdigitated electrodes, promise increased performance over previous piezoelectric actuators that were tested on the fin of an F/A-18 aircraft. Two new actuators being considered by the United States Air Force to reduce buffet loads on high performance aircraft were embedded into the fins of an F/A-18 wind-tunnel model and tested in the Transonic Dynamics Tunnel at the NASA Langley Research Center. The purpose of this test program, called ENABLE (Evaluation of New Actuators in a Buffet Loads Environment), was to examine the performance of the new actuators in alleviating fin buffeting, leading to a systems -level study of a fin buffet loads alleviation system architecture being considered by the USAF, Boeing, and NASA for implementation on high performance aircraft. During this windtunnel test, the two actuators performed superbly in alleviating fin buffeting. Peak values of the power spectral density functions for tip acceleration were reduced by as much as 85%. RMS values of tip acceleration were reduced by as much as 40% while using less than 50% of the actuators capacity. Details of the wind-tunnel model and results of the wind-tunnel test are provided herein.