214 research outputs found
Optimal policies for two-user energy harvesting device networks with imperfect state-of-charge knowledge2014 Information Theory and Applications Workshop (ITA)
No full textEnergy Harvesting Devices (EHDs) are enjoying continuously increasing popularity in Wireless Sensor Network research, due to their ability to “harvest” energy from the environment, thus allowing long-term and autonomous operation. Traditional approaches generally assume that the exact energy value of the State-of-Charge (SOC) of an EHD is known. In reality, batteries are practically composed by electro-chemical rechargeable elements or super-capacitors, where the estimation of the energy levels is a complex task. In this paper, we analyze operation policies able to maximize the long-term reward for a network consisting of a pair of EHDs and a central controller (CC), under imperfect knowledge of the SOC. More precisely, we suppose that the CC only knows whether each EHD is “LOW” or ”HIGH,”and has to determine the amount of energy devoted to the transmission over a shared wireless channel. We show that the performance degradation due to the imperfect knowledge of the SOC decreases with the battery capacity of the two nodes, and is practically negligible when this value is sufficiently high
Stochastic Optimization and Machine Learning Modeling for Wireless Networking
No full textIn the last years, the telecommunications industry has seen an increasing interest in the development of advanced solutions that enable communicating nodes to exchange large amounts of data. Indeed, well-known applications such as VoIP, audio streaming, video on demand, real-time surveillance systems, safety vehicular requirements, and remote computing have increased the demand for the efficient generation, utilization, management and communication of larger and larger data quantities. New transmission technologies have been developed to permit more efficient and faster data exchanges, including multiple input multiple output architectures or software defined networking: as an example, the next generation of mobile communication, known as 5G, is expected to provide data rates of tens of megabits per second for tens of thousands of users and only 1 ms latency. In order to achieve such demanding performance, these systems need to effectively model the considerable level of uncertainty related to fading transmission channels, interference, or the presence of noise in the data.
In this thesis, we will present how different approaches can be adopted to model these kinds of scenarios, focusing on wireless networking applications. In particular, the first part of this work will show how stochastic optimization models can be exploited to design energy management policies for wireless sensor networks. Traditionally, transmission policies are designed to reduce the total amount of energy drawn from the batteries of the devices; here, we consider energy harvesting wireless sensor networks, in which each device is able to scavenge energy from the environment and charge its battery with it. In this case, the goal of the optimal transmission policies is to efficiently manage the energy harvested from the environment, avoiding both energy outage (i.e., no residual energy in a battery) and energy overflow (i.e., the impossibility to store scavenged energy when the battery is already full).
In the second part of this work, we will explore the adoption of machine learning techniques to tackle a number of common wireless networking problems. These algorithms are able to learn from and make predictions on data, avoiding the need to follow limited static program instructions: models are built from sample inputs, thus allowing for data-driven predictions and decisions. In particular, we will first design an on-the-fly prediction algorithm for the expected time of arrival related to WiFi transmissions. This predictor only exploits those network parameters available at each receiving node and does not require additional knowledge from the transmitter, hence it can be deployed without modifying existing standard transmission protocols. Secondly, we will investigate the usage of particular neural network instances known as autoencoders for the compression of biosignals, such as electrocardiography and photo plethysmographic sequences. A lightweight lossy compressor will be designed, able to be deployed in wearable battery-equipped devices with limited computational power. Thirdly, we will propose a predictor for the long-term channel gain in a wireless network. Differently from other works in the literature, such predictor will only exploit past channel samples, without resorting to additional information such as GPS data. An accurate estimation of this gain would enable to, e.g., efficiently allocate resources and foretell future handover procedures. Finally, although not strictly related to wireless networking scenarios, we will show how deep learning techniques can be applied to the field of autonomous driving. This final section will deal with state-of-the-art machine learning solutions, proving how these techniques are able to considerably overcome the performance given by traditional approaches
Lightweight Lossy Compression of Biometric Patterns via Denoising Autoencoders
No full textWearable Internet of Things (IoT) devices permit the massive collection of biosignals (e.g., heart-rate, oxygen level, respiration, blood pressure, photo-plethysmographic signal, etc.) at low cost. These, can be used to help address the individual fitness needs of the users and could be exploited within personalized healthcare plans. In this letter, we are concerned with the design of lightweight and efficient algorithms for the lossy compression of these signals. In fact, we underline that compression is a key functionality to improve the lifetime of IoT devices, which are often energy constrained, allowing the optimization of their internal memory space and the efficient transmission of data over their wireless interface. To this end, we advocate the use of autoencoders as an efficient and computationally lightweight means to compress biometric signals. While the presented techniques can be used with any signal showing a certain degree of periodicity, in this letter we apply them to ECG traces, showing quantitative results in terms of compression ratio, reconstruction error and computational complexity. State of the art solutions are also compared with our approach
Considerazioni costituzionali a margine del processo promosso dai rider di Torino contro Foodora: il lavoro di fronte alla gig economy
No full textAgli esordi del 2020 la Corte di Cassazione si è pronunciata sul contenzioso che vedeva contrapposti l’azienda Foodora e i rider cui non stato rinnovato il contratto in seguito agli scioperi del 2016. La vicenda giudiziaria stata il casus belli che ha portato definitivamente alla luce anche in Italia il fenomeno della gig economy, consistente nelle più precarie forme di organizzazione del lavoro nel settore terziario mediante l’uso di piattaforme digitali. Esaminata la gig economy dalla prospettiva dei lavoratori coinvolti, una lettura costituzionale dei rapporti economici porta a chiedersi come si configuri il valore fondativo e antropologico che riconosciuto al lavoro dalla Costituzione italiana nelle nuove forme del lavoro digitale, e come debba intervenire l’azione del legislatore e dei poteri pubblici, finalisticamente orientata al perseguimento dei principi fondamentali
Il David con la testa di Golia di Caravaggio dalla collezione al museo
No full textStoria collezionistica, dal Seicento all'attuale istituzione museale, del Davide con la testa di Golia di Caravaggio della Galleria Borghese
Applying Machine Learning Techniques to a Real Cognitive Network: File Transfer ETAs Prediction
No full textCognitive Network (CN) paradigm and Machine Learning (ML) techniques are increasingly becoming popular in Wireless network design. CN allows to efficiently arrange the network stack parameters among the nodes involved in the data transmission. It can be divided into two modules: the former permits to retrieve the ISO/OSI protocol stack parameters and the latter optimizes the transmission through a Cognitive Engine entity. This engine usually adopts ML algorithms to extract important features regarding the network, with the objective of either maximizing the global throughput or predicting relevant network behaviors. This paper analyzes how common ML algorithms are able to model transmissions occurring in a typical wireless network scenario. In particular, we test the algorithms with our CN testbed to predict transmission estimated time of arrivals (ETAs) in different network setups. We compare these results with those obtained via scp, which is the typical Unix/Linux shell program used to exchange files. We show that all learning techniques significantly improve the goodness of ETA prediction, thus suggesting to embed such algorithms in future scp revisions
Low-complexity policies for Wireless Sensor Networks with two energy harvesting devices2014 13th Annual Mediterranean Ad Hoc Networking Workshop (MED-HOC-NET)
No full textRecently, Wireless Sensor Networks (WSNs) with energy harvesting capabilities are experiencing increasing interest with respect to traditional ones, due to their promise of extended lifetime. We consider the case of a pair of Energy Harvesting Devices (EHDs), with the main goal of maximizing the long-term aggregate average importance associated with the transmitted data. The devices, at each time instant, have data of different importance levels to be transmitted, as well as different battery energy levels. In order to avoid collisions, a Central Controller (CC) allows at most the transmission of a single EHD per time slot. Assuming a negligible processing cost in terms of energy, our objective is to identify low-complexity transmission policies, that achieve good performance with respect to the optimal one. We numerically show that two policies, namely the Balanced Policy (BP) and the Heuristic Constrained Energy Independent Policy (HCEIP), despite being independent of the battery energy levels, achieve near optimal performance in most cases of interest, and can be easily found with an adaptation to the ambient energy supply. Moreover, we derive analytically an approximation of the BP and show that this policy can be considered a good lower bound for the performance of the Optimal Policy
Optimal Transmission Policies for Two-User Energy Harvesting Device Networks With Limited State-of-Charge Knowledge
No full textThis paper considers a wireless network composed of a pair of sensors powered by energy harvesting devices (EHDs), which transmit data to a receiver over a shared wireless channel. At any given time, based on the energy levels of the two rechargeable batteries of the sensors, a central controller (CC) decides on the amount of energy to be drawn from the two batteries and used for transmission. The problem considered is the maximization of the long-term average reward associated with data transmission, by optimizing the transmission strategy of the two nodes, in the case of a collision channel model and both i.i.d. and correlated energy arrivals. In addition, contrary to the traditional assumption that the amount of energy available to the sensors can be easily estimated, we derive the optimal policy in the cases where the state of charge (SOC) may not be perfectly known by the central controller, analyzing the performance degradation caused by this imperfect knowledge of the SOC. For this second scenario, supposing that the CC is only aware that each SOC is “LOW” or “HIGH,” we show that the impact of imperfect knowledge decreases with the two battery capacities and is negligible in most cases of practical interest
A machine learning-based ETA estimator for Wi-Fi transmissions
No full textRecent advancements related to device to device (D2D) communication make it possible for a transmitting node to dynamically select the interface to be used for data transfers locally, without traversing any network infrastructure. In this scenario, a controller is identified, whose goal is to manage the D2D connection after its establishment. The software defined networking paradigm makes it possible to select this controller node via software: a device becomes the master node of a Wi-Fi-direct network, whereas the remaining units, i.e., the clients, can exchange data with other devices through the master. This paper develops a machine learning-based prediction algorithm for the aforementioned scenario, in which multiple elements, while receiving data from the controller, require an accurate on-the-fly estimation of the remaining transmission time, i.e., the estimated time of arrival. Different machine learning approaches are considered for this task, with the goal of exploiting only the information available at each client, without modifying any standard communication protocol. This information is critical when, for instance, a mobile user needs to decide whether or not to delay a data transfer, based on the load of the network and on the residual time under radio coverage from an access point
On Optimal Transmission Policies for Energy Harvesting Devices: the case of two users
No full textWe consider a pair of energy harvesting devices, whose state at any given time is determined by the energy level and an importance value, associated to the transmission of a data packet to the receiver at that particular time. Under the assumption of a central controller with information on the energy level and packet importance of both nodes, we consider the problem of optimizing the transmission strategy of the two nodes over a shared wireless channel, with the goal of maximizing the long-term average importance of the transmitted data. Under the assumption of i.i.d. Bernoulli energy arrivals and a collision channel model, we show the optimality of threshold policies with respect to the data importance level, i.e., the sensor nodes should report only data with an importance value above a given threshold. We provide numerical results to evaluate the impact on the performance of factors such as the battery capacity size and the energy harvesting rate, as well as the interaction between the two nodes
- …
