1,721,065 research outputs found
BFA-Sense: Learning Beamforming Feedback Angles for Wi-Fi Sensing
No full textIn this paper, we propose BFA-Sense, a completely novel approach to implement standard-compliant Wi-Fi sensing applications. Wi-Fi sensing enables game-changing applications in remote healthcare, home entertainment, and home surveillance, among others. However, existing work leverages the manual extraction of the uncompressed channel state information (CSI) from Wi-Fi chips, which is not supported by the 802.11 standard-compliant devices and hence requires the use of specialized equipment. On the contrary, BFA-Sense leverages the compressed beamforming feedback angles (BFAs) transmitted during the standard-compliant sounding procedure to characterize the propagation environment. Conversely from the uncompressed CSI, BFAs (i) can be recorded without any firmware modification, and (ii) allows a single monitor device to simultaneously capture the channels between the access point and all the stations, thus providing much better sensitivity. We evaluate BFA-Sense through an extensive data collection campaign with three subjects performing twenty different activities in three different environments. We assess the cross-domain adaptability of BFA-Sense through embedding learning for tackling unseen environments with a few samples from the new environment. The results show that the proposed BFAs-based approach achieves about 11% more accuracy when compared to CSI-based prior work
Smartphone Identification via Passive Traffic Fingerprinting: A Sequence-to-Sequence Learning Approach
No full textPassive cyber-security attacks do not require any modification of the data stream generated by the victim, nor the creation of a false statement; in particular, those attacks based on statistical analysis aim at acquiring sensible information by just analyzing traffic patterns. Our work sits on the conjecture that the PDCCH, which is transmitted in clear text, may be effectively used to statistically characterize the traffic generated by a smartphone in standby mode. Through this statistical signature, the attacker may then infer whether an unknown traffic pattern is generated by the victim user's terminal, guessing if the victim is in a certain geographical area, and in turn gaining the ability to track the victim's movements and/or to profile their habits. In this work, we propose a data collection and processing framework that successfully obtains such signatures. User data patterns (transport block sizes and communications direction) are retrieved by analyzing the mobile network scheduling. Hence, a sequence-to-sequence learning framework to extract smartphone signatures from passive traffic is put forward, and is experimentally validated using a dataset of 40 user traces, successfully identifying up to 90 percent of the users
HIJACK: Learning-based Strategies for Sound Classification Robustness to Adversarial Noise
No full textThe effective deployment of smart service systems within homes, workspaces and cities, requires gaining context and situational awareness to take action when changes are detected. To this end, sound classification systems are widely adopted and integrated into several smart devices to continuously monitor the environment. However, sound classification algorithms are prone to adversarial attacks that pose a considerable security threat to smart service systems where they are integrated. In this paper, we devise HIJACK, a novel machine learning framework entailing five neural network strategies to enforce the robustness of sound classification systems to adversarial noise injection. The HIJACK methodologies can be applied to any neural network-based sound classifier and consist of tailored transformations of the input audio during training along with specific additional layers added to the neural network architecture. To assess the noise robustness provided by the HIJACK strategies, we design a measure based on a L-2-adversarial attack to sound classification identified as the normalized fast gradient method (NFGM) - that constructs the adversarial noise by maximizing the sound mis-classification probability. We assessed the robustness of HIJACK to the proposed NFGM attack on a publicly available dataset. The results show that the combination of the five HIJACK strategies allows reaching robustness to adversarial noise 58 times larger than state-of-the-art neural networks for sound classification, guaranteeing a classification accuracy above 83%
Mobility prediction via sequential learning for 5G mobile networks
No full textHere, we present a mobility prediction framework for 5G mobile systems. Our work stems from the intuition that mobility in vehicular networks is highly correlated, and such correlation can be captured by advanced neural network designs to anticipate the users' point of attachment. To prove this, we combine Markov chains with recurrent and convolutional neural networks, training them on mobility trajectories estimated by the received radio signal from mobile millimeter-wave devices. The proposed framework is decentralized, i.e., user trajectories are independently learned by each base station. In this paper, various problems are pragmatically tackled and solved, such as dealing with imbalanced datasets, as some trajectories are under represented, and obtaining a mobility classifier whose accuracy increases as new mobility samples are collected.The proposed technique is assessed using emulated traces obtained through the SUMO mobility simulator for the city of Cologne. Numerical results show accuracies higher than 88% in the prediction of the next serving base station from 4 seconds before the handover is performed. Mobility (next base station) predictors like the ones presented here are key for network management purposes within 5G networks, e.g., to proactively allocate communication and edge computing resources
Sleep disturbance in Mild Cognitive Impairment and association with cognitive functioning. A case-control study. , 2018 Nov 9;10:360. doi: 10.3389/fnagi.2018.00360
Full text linkObjectives: The aims of the current study are to (1) report the frequency of specific sleep disturbance symptoms in Mild Cognitive Impairment (MCI) and cognitive healthy older persons; (2) examine whether overall poor sleep and specific sleep disturbance symptoms are more common in persons with MCI compared to cognitive healthy older controls and; (3) examine the association between sleep disturbances and performance in general and specific cognitive domains in persons with MCI and separately in cognitive healthy older persons. Methods: Data were collected at the Fondazione Ospedale San Camillo Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS), Venice, Italy as part of the European VPH-DARE@IT project. We included 69 persons with MCI (mean age 75.7; SD = 7.7) and 72 sex-matched cognitively healthy controls (mean age 71.8; SD = 7.0). Participants underwent extensive neuropsychological assessment and evaluation of subjective sleep performance with the Sleep Continuity in Alzheimer's Disease Scale (SCADS). Results: A fifth of MCI patients (21.7%, n = 15) had poor sleep compared to 15.3% (n = 11) of cognitively healthy controls. MCI patients had a 3.2 higher odds of having poor sleep compared to cognitively healthy controls after adjustment for age, education, sex, and general cognitive functioning (Odds Ratio (OR)=3.2; 95% Confidence Interval (CI)=1.1-9.2). Persons who reported waking up twice or more during the night had higher odds of being MCI compared to those who never wake or wake only once (OR = 2.6; 95% CI = 1.1-6.1). In MCI patients, poor sleep was associated with better general cognitive functioning and short-term working memory, whereas in cognitive healthy older persons poor sleep was associated with impairment in episodic memory performance and executive functioning. Discussion: Our results confirm previous studies showing that sleep disturbances are common in MCI, and this may be due to an ongoing neurodegenerative process rather than a symptom of cognitive impairment. Future research with objective sleep measurements are needed in MCI as well as interventions to improve sleep with the aim of preventing cognitive decline
The spatial representation of numerical and non-numerical sequences: Evidence from neglect
No full textProduzione di verbi e deficit fonologico in una paziente con afasia di Broca.
No full textA patient with left anterior lesion, probable Broca's aphasia, is required to generate the Italian past participle of real and fake verbs. The results are discussed in light of the two mechanism model and the model by Bird et al., (2003) and suggest an impairment in the processing of complex phonological forms
Design and evaluation of a low-cost acoustic chamber for underwater networking experiments
No full textTesting acoustic equipment before sea experiments is a necessary step, which usually requires large and expensive facilities. In this paper, we present the design guidelines, structure and details of a small-scale, low-cost acoustic chamber for in-lab testing of underwater acoustic networks. The chamber has been assembled with the objective to be of low cost and limited size: therefore, its installation fits small university laboratories that cannot afford large testing pools. The chamber was designed to mitigate the extreme multipath which, in a small chamber, makes communications unreliable. Considering this challenge, our chamber includes a phono-absorbing coating on the walls and floor, to be optionally complemented by a panel of the same coating material, to be installed at the water surface level. After providing the details of several phono-absorbing materials to motivate our specific choice, we carry out a number of transmission experiment with EvoLogics modems, proving that our design substantially reduces the severe multipath and thereby improves the communications quality
Wi-Fi Multi-Path Parameter Estimation for Sub-7 GHz Sensing: A Comparative Study
No full textThanks to the definition of the new IEEE 802.11bf standard, the development of Wi-Fi sensing applications is gaining momentum in the research community. In this regard, several studies have shown that learning-based approaches that leverage the frequency response of the Wi-Fi channel in the sub-7GHz bands can reach high accuracy in different classification tasks, such as activity recognition, or person identification. Instead, more fine-grained applications - e.g., human localization and tracking, or respiration and heartbeat monitoring - require implementing model-based approaches to estimate the Wi-Fi multi-path parameters and analyze the time evolution of the paths associated with specific targets (the human body or chest). In this paper, we investigate the performance of six super-resolution algorithms for sub-7GHz multi-path parameter estimation. Our extensive evaluation indicates that the estimation accuracy that can be achieved through commercial devices allows implementing human localization and tracking strategies but is insufficient to effectively design human vital signs monitoring applications due to the limited frequency and spatial diversity. We pledge to release our implementations for further investigations
How to differentiate hemianesthesia from left tactile neglect: A preliminary case report
No full text- …
