196,060 research outputs found
A Machine Learning Model for Microcontrollers Enabling Low Power Indoor Positioning Systems via Visible Light Communication
No full textThis paper presents a low-power Visible Light Localisation (VLL) Artificial Intelligence (AI)-enabled system for Indoor Positioning (IP) purposes. Compared to other IP techniques, VLL offers a similar positioning accuracy, but with the extremely desirable feature of low energy consumption, an aspect of primary relevance in the framework of Wireless Sensor Networks (WSN), self-sufficient sensing systems, Industry 4.0 and Internet of Things (IoT). The proposed system is composed of three modulated optical sources (i.e. LEDs) and a photodiode receiver mounted on the target to be localised. The localisation task is performed by processing the received light intensities through Machine Learning (ML) regression models trained with a set of data gathered during a calibration phase. The regressors are designed to be executed on a low-power microcontroller present in the target, hence establishing an embedded ML paradigm also preserving reduced power consumption features. The proposed models are trained exploiting datasets with different sizes, searching for a trade-off between the training set size, i.e. the duration and complexity of the calibration phase, and the maximum tolerable root mean square error (RMSE). In both cases, some localisation tests show that a satisfactory accuracy can be reached even with a limited complexity of the calibration procedure and that the obtained results fulfil the error constraint used for model design
Low-cost accurate inductive system for thickness measurement of industrial ferromagnetic plates
No full textThis paper deals with profile thickness measurement of ferromagnetic slabs for industrial applications, by means of inductive proximity sensors. A low-cost accurate measurement system was realized, based on an embedded microcontroller and exploiting two inductive probes facing each other. Such system was designed to operate with moving plates, passing through the probe assembly mounted in a fixed position. The developed system was properly calibrated, following a procedure implemented ad hoc, and its performance was assessed using sample ferromagnetic test objects. The achieved accuracy of the thickness estimation is about 20 mu m at 2 readings/s measurement speed. The accuracy of the system does not depend on the type of ferromagnetic material (i.e on the value of the magnetic permeability, provided it is large). Afterwards, an experimental campaign was conducted, characterizing the thickness of solid and laser beam machined ferromagnetic slabs
A Machine Learning Model for Microcontrollers Enabling Low Power Indoor Positioning Systems via Visible Light Communication
No full textThis paper presents a low-power Visible Light Localisation (VLL) Artificial Intelligence (AI)-enabled system for Indoor Positioning (IP) purposes. Compared to other IP techniques, VLL offers a similar positioning accuracy, but with the extremely desirable feature of low energy consumption, an aspect of primary relevance in the framework of Wireless Sensor Networks (WSN), self-sufficient sensing systems, Industry 4.0 and Internet of Things (IoT). The proposed system is composed of three modulated optical sources (i.e. LEDs) and a photodiode receiver mounted on the target to be localised. The localisation task is performed by processing the received light intensities through Machine Learning (ML) regression models trained with a set of data gathered during a calibration phase. The regressors are designed to be executed on a low-power microcontroller present in the target, hence establishing an embedded ML paradigm also preserving reduced power consumption features. The proposed models are trained exploiting datasets with different sizes, searching for a trade-off between the training set size, i.e. the duration and complexity of the calibration phase, and the maximum tolerable root mean square error (RMSE). In both cases, some localisation tests show that a satisfactory accuracy can be reached even with a limited complexity of the calibration procedure and that the obtained results fulfil the error constraint used for model design
An overview on building-integrated photovoltaics: Technological solutions, modeling, and control
No full textThe advancement of renewable and sustainable energy generation technologies has been driven by environment-related issues, energy independence, and high costs of fossil fuels. Building-integrated photovoltaic systems have been demonstrated to be a viable technology for the generation of renewable power, with the potential to assist buildings in meeting their energy demands. This work reviews the current status of novel PV technologies, including bifacial solar cells and semi-transparent solar cells. This review discusses the various constructions of PV technologies, recent advances in these products, the influence of key design factors on electrical and thermal performance, and their potential in the design of energy-efficient smart buildings. The attention is focused on bifacial and semi-transparent PV systems, given the high level of interest of the scientific community in their current and potential applications.
Focus is also devoted to the analysis of the electrical, optical, and thermal modeling procedures developed for sizing, designing, and integrating photovoltaics into larger building simulations. The development of these models has a positive impact on the implementation of next-generation smart buildings. The latest innovative developments and key issues in the application of bifacial PV solutions in buildings are also summarized and analyzed. Special attention is paid to rear side electrical performance, which can be evaluated by means of illuminance/optical backside modeling. Finally, energy management and control of PV-equipped buildings via both model-based and data-driven approaches are discussed, as well as the integration of electric storage systems in a multi-building context
Enhanced Visible Light Localization Based on Machine Learning and Optimized Fingerprinting in Wireless Sensor Networks
No full textThis article presents a robust visible light localization (VLL) technique for wireless sensor networks, with 2-D indoor positioning (IP) capabilities, based on embedded machine learning (ML) running on low-cost low-power microcontrollers. The implemented VLL technique uses four optical sources (i.e., LEDs), modulated at different frequencies. In particular, the received signal strengths (RSSs) of optical signals are evaluated by a microcontroller on board the sensor nodes via fast Fourier transform (FFT). RSSs are fed to four embedded ML regressors, aiming at estimating the target position within the workspace. The four neural networks (NNs), one per each possible triplet of LEDs, are trained by exploiting a novel technique to generate the training datasets. This method, called optimized fingerprinting (OF), allows for creating arbitrarily ample datasets by performing only few measurements in the field, avoiding time-consuming steps for collecting experimental data. The NNs are devised to be accurate yet lightweight facilitating their implementation and execution by the microcontroller. Furthermore, due to the presence of four NNs, four position estimates are obtained. This redundancy is exploited to detect and effectively manage situations of total or partial shading of one light source and to enhance the positioning accuracy under normal operating conditions (i.e., no obstacles), by averaging the four positions. Test results performed in a m workspace show an overall mean accuracy of about 1 cm with standard deviation below the centimeter and maximum error around 3 cm
Solar energy harvesting for LoRaWAN-based pervasive environmental monitoring
No full textThe aim of this paper is to discuss the characterisation of a solar energy harvesting system to be integrated in a wireless sensor node, to be deployed on means of transport to pervasively collect measurements of Particulate Matter (PM) concentration in urban areas. The sensor node is based on the use of low-cost PM sensors and exploits LoRaWAN connectivity to remotely transfer the collected data. The node also integrates GPS localisation features, that allow to associate the measured values with the geographical coordinates of the sampling site. In particular, the system is provided with an innovative, small-scale, solar-based powering solution that allows its energy self-sufficiency and then its functioning without the need for a connection to the power grid. Tests concerning the energy production of the solar cell were performed in order to optimise the functioning of the sensor node: Satisfactory results were achieved in terms of number of samplings per hour. Finally, field tests were carried out with the integrated environmental monitoring device proving its effectiveness
Monitoring of BIPV by Means of a Low Cost Wireless Sensor Network
No full textDecarbonisation and Nearly Zero Energy Buildings are interlinked concepts that are fundamental to addressing the goal of reducing carbon emissions and mitigate climate change. In this scenario, Building Integrated Photovoltaic systems represents a key strategy. On the other hand, PV devices in this way are dislocated in different parts of the building and this makes difficult the monitoring and set-up of energy conversion chains. In order to overcome this limitations, we propose an efficient and reliable sensor networks, which can be achieved by exploiting IoT approach. In particular, a low-cost wireless sensor, based on the microcontroller ESP32, was developed with the aim to measure voltage, current and temperature of photovoltaic integrated system and, from these measurements, irradiation or shading can be calculated, in order to evaluate energetic performance. Thanks to this simple sensor any PV device can become an irradiance sensor, but also the basic component of a system for the evaluation of environmental and energetic parameters of the building
Relative stopping power measurements and prosthesis artifacts reduction in proton CT
No full textWe present a set-up for proton computed tomography (pCT), composed of a microstrip silicon tracker and a YAG:Ce calorimeter, able to directly measure the relative stopping power (RSP) maps to be used in hadron therapy. The system, tested with an electron density phantom at the Trento proton Therapy Center, is able to correlate measured and expected RSP with discrepancies of the order of 1% or less. Furthermore, pCT tomographies of an anthropomorphous head phantom taken with our device, when compared with X-ray CT images of the same object, evidence a significant reduction of artifacts induced by titanium spinal bone prosthesis and tungsten dental filling
Pervasive environmental monitoring by means of self-powered particulate matter LoRaWAN sensor nodes
No full textThe aim of this paper is to propose the architecture of a self-powered sensor node, to be deployed on means of transport to pervasively collect measurements of Particulate Matter (PM) concentration in urban areas. The sensor node is based on the use of low cost PM sensors and exploits LoRaWAN connectivity to remotely transfer the collected data. It is also provided with GPS localization features that allow to associate the measured values with the geographical coordinates of the sampling site. The system is also provided with an innovative, small-scale, solar-based powering solution that allows its energy-self sufficiency and then its functioning without the need for a connection to the power grid
Working principle and performance of a scalable gravimetric system for the monitoring of access to public places
No full textHere, we propose a novel application of a low-cost robust gravimetric system for public place access monitoring purposes. The proposed solution is intended to be exploited in a multi-sensor scenario, where heterogeneous information, coming from different sources (e.g., metal detectors and surveillance cameras), are collected in a central data fusion unit to obtain a more detailed and accurate evaluation of notable events. Specifically, the word “notable” refers essentially to two event categories: the first category is represented by irregular events, corresponding typically to multiple people passing together through a security gate; the second category includes some event subsets, whose notification can be interesting for assistance provision (in the case of people with disabilities), or for statistical analysis. The employed gravimetric sensor, compared to other devices existing in the literature, exhibits a simple scalable robust structure, made up of an array of rigid steel plates, each laid on four load cells. We developed a tailored hardware and software to individually acquire the load cell signals, and to post-process the data to formulate a classification of the notable events. The results are encouraging, showing a remarkable detectability of irregularities (95.3% of all the test cases) and a satisfactory identification of the other event types
- …
