Search

Digital technology is now very sophisticated. Its use is widely applied in all areas of human life. Starting from waking up, human activities and others always use technology. In carrying out their activities, modern humans now almost all use vehicles as a mode of transportation. Today's vehicles use a variety of sensors as a sixth sense. The results of detection using sensors on the vehicle are usually displayed on the dashboard of the vehicle. Modern humans currently use sensors to complete their needs. Besides that, the internet of things technology is growing rapidly in its role and development to support the needs of modern humans. Micro-controller technology is also experiencing rapid and massive development. One of the most common and most popularly used microcontrollers is Arduino. In many streets in Indonesia, people still use vehicles not equipped with many sensors. One of them is a simple parking sensor that many old vehicles don't have. Parking sensor problems are needed at the time of parking so that the vehicle that will be parked does not hit other objects or vehicles. There are many types of ultrasonic sensors. The purpose of this research is to make a prototype ultrasonic sensor that is applied to vehicles and compare some of the most accurate ultrasonic sensors in measuring the distance between the vehicle and the object being measured.

Advances in soft sensors coupled with machine learning are enabling increasingly capable wearable systems. Since hand motion in particular can convey useful information for developing intuitive interfaces, glove-based systems can have a significant impact on many application areas. A key remaining challenge for wearables is to capture, process, and analyze data from the high-degree-of-freedom hand in real time. We propose using a commercially available conductive knit to create an unobtrusive network of resistive sensors that spans all hand joints, coupling this with an accelerometer, and deploying machine learning on a low-profile microcontroller to process and classify data. This yields a self-contained wearable device with rich sensing capabilities for hand pose and orientation, low fabrication time, and embedded activity prediction. To demonstrate its capabilities, we use it to detect static poses and dynamic gestures from American Sign Language (ASL). By pre-training a long short-term memory (LSTM) neural network and using tools to deploy it in an embedded context, the glove and an ST microcontroller can classify 12 ASL letters and 12 ASL words in real time. Using a leave-one-experiment-out cross validation methodology, networks successfully classify 96.3% of segmented examples and generate correct rolling predictions during 92.8% of real-time streaming trials. Author

The advanced data modelling capabilities of neural networks allowed deep learning to become a cornerstone of many applications of artificial intelligence (AI). AI can be brought into our environments by deploying neural models to ubiquitous Internet-of-Things (IoT), wearable and embedded devices to enhance tasks like voice assistance, home security, or health and fitness monitoring. However, such devices are typically powered by microcontroller units (MCUs), which have orders of magnitude fewer computational resources than what is typically required to execute neural networks. This resource scarcity presents a significant obstacle for on-device deep learning, forcing applications to offload the computation to a remote device or server, sacrificing user data privacy and the autonomy of the system. The challenge of bringing neural networks to microcontroller-powered devices, potentially the smallest-scale viable computing platform for deep learning, is tackled by the emerging research field called *TinyML*. In this thesis, I develop model discovery and compression methodology whose common threads are automation and holistic optimisation of network architectures and their execution software, informed by the computational limitations of microcontroller hardware. This unlocks previously-impossible applications of deep learning on MCUs and advances the predictive performance vs resource usage trade-off of neural networks compared to prior work. The novel methodology spans the domains of model compression, neural architecture search (NAS) and software-level execution optimisation. Firstly, I present μNAS, an evolutionary architecture search algorithm that uses a highly granular search space to find performant models with an extremely low (

My technical project and my STS research paper are connected primarily through the use of microcontrollers in education and the study of mechatronics. Mechatronics combines the study of mechanical engineering with electrical and computer engineering and has sprouted into a category of its own. My technical work focuses on the creation of a mechanical wall clock in the form of an abacus. The aim of this project is to demonstrate the real world practicality of implementing microcontrollers within undergraduate education. Moreover, the STS research paper aims to focus on the effects of microcontrollers on the engineering industry as well as its influence on education and student experiences. My technical work explores the idea of using microcontrollers and the principles of mechatronics to create a mechanical abacus clock. This clock will integrate mechanical and mechatronics systems that both demonstrate the practical skills of undergraduate students at the University of Virginia while paying homage to the origins of numerical methods. An abacus is a computing tool made from sliding beads on rods used for the four arithmetic operations (addition, subtraction, multiplication, and division). This machine was built entirely from scratch using a combination of 3D printed parts, motors, microcontrollers, and acrylic. The goal of the project is to promote the study of engineering and inspire students to understand the usefulness of microcontrollers in the real world. My STS research aims to analyze how microcontrollers have become an integral part of education and their effects on the student experience. My research uses the Social Construction of Technology (SCOT) method to discuss how microcontrollers have become an integral part of education as well as ways in which the student experience has changed. By analyzing the various societal factors contributing to the design of microcontrollers and use in education the paper aims to create points of discussion that will highlight effects on student experience and education. Working on these projects in tandem significantly influenced the value of both. My technical project provided me with a better understanding of mechanical and mechatronic systems as well as the practicality of using microcontrollers. Similarly, my STS research contributed to my view of the importance of using microcontrollers in education as well as ways in which the student experience can be improved. Both projects inspired me to further pursue my engineering studies and to contribute to the growth of prospective engineering students. Working on both my STS research paper and technical project this past year has allowed me to explore the use of microcontrollers in education from various perspectives and each work contributed to improving one another.

The Microcontroller (MCU) is a cutting-edge piece of technology that has both civilian and engineering applications. The STS paper will specifically look at how the inception and development of the MCU have shaped society and engineering into what it is now. The application and social impact of MCUs will be discussed in detail, proving that although MCUs may be a dark-horse technology, they are highly impactful and important in the development of the digital age. The technical project is a user-controlled ping-pong ball-launching robot that has a deterministic feedback loop with a self-correction procedure. The robot can aim at 1 of 9 target squares with immense accuracy. Using image processing, the detection of the ball will be reported back to the robot and if it has missed it will readjust and make a correction to fire again. The project uses a combination of a pan/tilt mount controlled by servo motors, a push solenoid, and a computer user interface as a control mechanism for aiming.

The subject of research in the article are algorithms for fast calculation of autoregression coefficients in digital spectral analysis and estimation of the number of arithmetic operations required for their implementation. The aim of the article – comparative analysis of the speed of different algorithms for calculating the coefficients of autoregression as part of the algorithms of spectral analysis, including analysis of the complexity of their microcontroller implementation. Tasks to be solved: selection of spectral analysis methods suitable for diagnostics of technological equipment, analysis of methods for calculating autoregression coefficients and derivation of relations for estimating software complexity of algorithms and calculation of numerical estimates of addition and multiplication for some algorithms, adaptation of developed methods and estimates to microcontrollers. spectrum Applied methods: algorithm theory, Fourier transform, natural series, microcontroller programming. The results obtained: it is shown that spectral estimation methods based on Yul-Walker equations, which require the calculation of autoaggression coefficients, combine sufficient resolution and resistance to interference with acceptable implementation complexity. Estimates of the number of additions and multiplications for the Levinson, Durbin, and Trench algorithms are obtained, and their comparative analysis is performed. The calculation times for microcontroller arithmetic with fixed and floating points were count upon. Conclusions: When constructing spectrum analyzers for the diagnosis of technological equipment, it is advisable to use the Yul-Walker method. A comparison of Levinson, Durbin, and Trench algorithms for calculating autoregression coefficients showed that the Trench method requires a minimum number of additions, and the Durbin method requires a minimum number of multiplications. At microcontroller realization of spectrum analyzers, it is necessary to consider features of the arithmetic used by the controller. The Trench method is the fastest in the case of floating-point arithmetic and small-scale modeling. In other cases, Durbin's method is more effective., Предметом исследования в статье являются алгоритмы быстрого вычисления коэффициентов авторегрессии при цифровом спектральном анализе и оценки числа арифметических операций, необходимых для их выполнения. Цель статьи - сравнительный анализ быстродействия различных алгоритмов вычисления коэффициентов авторегрессии как составной части алгоритмов спектрального анализа, в том числе анализ сложности их микроконтроллерной реализации. Решаемые задачи: выделение методов спектрального анализа, пригодных для диагностики технологического оборудования, вывод соотношений для оценивания программной сложности алгоритмов и расчет числовых оценок вычислительной сложности, адаптация разработанных методов и оценок к микроконтроллерной реализации анализаторов спектра. Применяемые методы: теория алгоритмов, преобразование Фурье, линейная алгебра, натуральные ряды, программирование микроконтроллеров. Полученные результаты: показано, что метод спектрального оценивания на основе уравнений Юла-Уолкера, требующих вычисления коэффициентов авторегрессии, сочетает достаточную разрешающую способность и помехоустойчивость с приемлемой сложностью реализации. Получены оценки числа сложений и умножений для алгоритмов Левинсона, Дурбина и Тренча и выполнен их сравнительный анализ. Рассчитаны времена вычисления для микроконтроллерной арифметики с фиксированной и плавающей запятой. Выводы: сравнение алгоритмов Левинсона, Дурбина и Тренча для вычисления коэффициентов авторегресси показало, что метод Тренча требует минимального количества сложений, а метод Дурбина – минимального количества умножений. При микроконтроллерной реализации анализаторов спектра следует учитывать особенности используемой контроллером арифметики. Метод Тренча является наиболее быстродействующим в случае применения арифметики с плавающей точкой и небольшого порядка модели. В остальных случаях более эффективен метод Дурбина., Предметом дослідження у статті є алгоритми швидкого обчислення коефіцієнтів авторегресії при цифровому спектральному аналізі та оцінки кількості арифметичних операцій, необхідних для їх реалізації. Мета статті – порівняльний аналіз швидкодії різних алгоритмів обчислення коефіцієнтів авторегресії як складової частини алгоритмів спектрального аналізу, у тому числі аналіз складності їхньої мікроконтролерної реалізації. Завдання, що вирішуються: виділення методів спектрального аналізу, придатних для діагностики технологічного обладнання, аналіз методів обчислення коефіцієнтів авторегресії та виведення співвідношень для оцінювання програмної складності алгоритмів та розрахунок числових оцінок числа додавань та множень для деяких алгоритмів, адаптація розроблених методів та оцінок до мікроконтролерної реалізації аналізаторів спектру Застосовувані методи: теорія алгоритмів, перетворення Фур’є, натуральні ряди, програмування мікроконтролерів. Отримані результати: показано, що методи спектрального оцінювання на основі рівнянь Юла-Уолкера, що вимагають обчислення коефіцієнтів авторегресії, поєднує достатню роздільну здатність і стійкість до перешкод з прийнятною складністю реалізації. Отримано оцінки числа додавань та множень для алгоритмів Левінсона, Дурбіна та Тренча, виконано їх порівняльний аналіз. Розраховані часи обчислення для мікроконтролерної арифметики з фіксованою та плаваючою точкою. Висновки: При побудові аналізаторів спектра для діагностики технологічного обладнання доцільно застосовувати метод Юла-Уолкера. Порівняння алгоритмів Левінсона, Дурбіна та Тренча для обчислення коефіцієнтів авторегресії показало, що метод Тренча потребує мінімальної кількості додавань, а метод Дурбіна – мінімальної кількості множень. При мікроконтролерній реалізації аналізаторів спектра слід враховувати особливості використовуваної контролером арифметики. Метод Тренча є найбільш швидкодіючим у разі застосування арифметики з плаваючою точкою та невеликого порядку моделі. В інших випадках ефективніший метод Дурбіна.

Optical flow estimation is crucial for autonomous navigation and localization of unmanned aerial vehicles (UAV). On micro and nano UAVs, real-time calculation of the optical flow is run on low power and resource-constrained microcontroller units (MCUs). Thus, lightweight algorithms for optical flow have been proposed targeting real-time execution on traditional single-core MCUs. This paper introduces an efficient parallelization strategy for optical flow computation targeting new-generation multicore low power RISC-V based microcontroller units. Our approach enables higher frame rates at lower clock speeds. It has been implemented and evaluated on the eight-core cluster of a commercial octa-core MCU (GAP8) reaching a parallelization speedup factor of 7.21 allowing for a frame rate of 500 frames per second when running on a 50MHz clock frequency. The proposed parallel algorithm significantly boosts the camera frame rate on micro unmanned aerial vehicles, which enables higher flight speeds: the maximum flight speed can be doubled, while using less than a third of the clock frequency of previous singlecore implementations., 2022 IEEE International Symposium on Circuits and Systems (ISCAS), ISBN:978-1-6654-8485-5, ISBN:978-1-6654-8484-8, ISBN:978-1-6654-8486-2

©2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. In this paper a wearable system with a sensor embroidered on a textile substrate to detect urine leaks is presented. The system consists of a moisture textile capacitive sensor together with the signal conditioning and its wireless transmission to the cloud. The proposed system has been integrated on underwear and hospital sheet to detect the urine leakage on the diaper users and critical ill patient, respectively. The methodology used by the microcontroller to measure the sensor value is a charge/discharge method. The information is visualised through a computer or smartphone, where can be seen the current state of the sensor. The system has a warning set up to communicate any urine leakage. The experimental results show the functionality of the proposed system which could supply a new tool to hospitals, nursing homes or other institutions to detect when the patient diaper or sheet need to be removed. This tool can optimise the hospital protocol and improve the patient quality of life. This work was supported by Spanish Government-MINECO under Project TEC2016-79465-R and AGAUR-UPC(2020 FIB 00028).

The authors gratefully acknowledge funding from European Union’s Horizon 2020 Research and Innovation programme under the Marie Sklodowska Curie grant agreement No. 813278 (A-WEAR, http://www.a-wear.eu/).

Machine learning is moving to the smallest computing devices. Today machine learning is applied even in tiny IoT microcontroller boards. In the IoT, LoRa is a popular communication technology to connect remote devices with gateways. Still, the confluence of machine learning in microcontrollers and networked LoRa connectivity is not yet fully exploited. In this paper we design a new LoRa connectivity for the Arduino Portenta H7, a recent microcontroller board equipped with embedded sensors suitable for diverse machine learning tasks. With the solution that we found the Arduino Portenta H7 is able to become part of a LoRa mesh network. This capacity increases the Portenta's range of applications. For the vision of distributed machine learning at the tiny edge, we can add with the Portenta an important board to become a smart compute node within a LoRa mesh network. This work was partially supported by the Spanish Government under contracts PID2019-106774RB-C21, PCI2019-111851-2 (LeadingEdge CHIST-ERA), PCI2019-111850-2 (DiPET CHIST-ERA).

Tento článek se zaměřuje na představení vyvinutého kalicího stroje založeného na indukčním ohřevu. Tento stroj se od běžných řešení výrazně liší především rozsáhlým měřicím systémem, díky němuž máme dokonalý přehled o tom, co se během procesu děje. Dále je popsán vyvinutý řídicí systém využívající PLC. V článku je zmíněn i návrh vlastního řízení frekvenčního měniče pomocí mikrokontrolérů. This article focuses on introducing the developed hardening machine based on induction heating. This machine differs significantly from conventional solutions mainly due to its extensive measuring system, which gives us a complete overview of what is happening during the process. Furthermore, the developed control system using PLC is described. This paper also mentions the design of own control of the frequency converter using microcontrollers.

Machine learning applications on resource-constrained devices such as microcontroller units often use models trained externally on more powerful devices. This approach, however, limits a later adaptation of the machine learning model in the device to changing data. Differently, on-device training allows the model to be updated for new datasets, but the training process needs to take into account the resource limitations of the device. This paper compares on-device training performance for a keyword spotting task using two popular microcontroller boards, Arduino Nano 33 BLE Sense and Arduino Portenta H7, in terms of inference accuracy, training latency, and current consumption. We use feedforward neural networks having a single hidden layer for models. The inference accuracy has been significantly improved using the Portenta H7 board by employing more neurons fitted to its memory budget, compared to the Nano board. With a neural network having 25 neurons for a hidden layer, the 5.0 x inference and 4.2 x training speedups are achieved using the Arduino Portenta H7 board, compared to the Arduino Nano 33 BLE Sense. While the memory of the Arduino Nano 33 BLE Sense is capable to train a neural network for the keyword spotting task, the Arduino Portenta H7 gives new possibilities for exploring more complex models for more complex problems thanks to a larger memory budget and adapting a model to new data in lower latency. This work was partially supported by the Spanish Government under contracts PID2019-106774RB-C21, PCI2019-111850-2 (DiPET CHIST-ERA CHIST-ERA-SDCDN-002), PCI2019-111851-2 (LeadingEdge CHIST-ERA), and UK Engineering and Physical Sciences Research Council (EP/T022345/1).

The use of information technology and technological medical devices has contributed significantly to the transformation of healthcare. Despite that, many problems have arisen in diagnosing or predicting diseases, either as a result of human errors or lack of accuracy of measurements. Therefore, this paper aims to provide an integrated health monitoring system to measure vital parameters and diagnose or predict disease. Through this work, the percentage of various gases in the blood through breathing is determined, vital parameters are measured and their effect on feelings is analyzed. A supervised learning model is configured to predict and diagnose based on biometric measurements. All results were compared with the results of the Omron device as a reference device. The results proved that the proposed design overcame many problems as it contributed to expanding the database of vital parameters and providing analysis on the effect of emotions on vital indicators. The accuracy of the measurements also reached 98.8% and the accuracy of diagnosing COVID-19 was 64%. The work also presents a user interface model for clinicians as well as for smartphones using the Internet of things.

The Internet of Things (IoT) era is mainly dependent on the word “Smart”, such as smart cities, smart homes, and smart cars. This aspect can be achieved through the merging of machine learning algorithms with IoT computing models. By adding the Artificial Intelligence (AI) algorithms to IoT, the result is the Cognitive IoT (CIoT). In the automotive industry, many researchers worked on self-diagnosis systems using deep learning, but most of them performed this process on the cloud due to the hardware limitations of the end-devices, and the devices obtain the decision via the cloud servers. Others worked with simple traditional algorithms of machine learning to solve these limitations of the processing capabilities of the end-devices. In this paper, a self-diagnosis smart device is introduced with fast responses and little overhead using the Multi-Layer Perceptron Neural Network (MLP-NN) as a deep learning technique. The MLP-NN learning stage is performed using a Tensorflow framework to generate an MLP model’s parameters. Then, the MLP-NN model is implemented using these model’s parameters on a low cost end-device such as ARM Cortex-M Series architecture. After implementing the MLP-NN model, the IoT implementation is built to publish the decision results. With the proposed implemented method for the smart device, the output decision based on sensors values can be taken by the IoT node itself without returning to the cloud. For comparison, another solution is proposed for the cloud-based architecture, where the MLP-NN model is implemented on Cloud. The results clarify a successful implemented MLP-NN model for little capabilities end-devices, where the smart device solution has a lower traffic and latency than the cloud-based solution.

Introduction. Walking machines have been interesting for decades. Modern technologies make it possible to create new designs with digital control. Creating software that allows a walking machine to move independently is a difficult task. Walking machine onboard computer needs to process data from sensors in real time. The article demonstrates design and algorithms used to control the motion of an experimental walking machine. Materials and Methods. To simulate the motion of a walking machine and experimental studies, a stand replicating all the electronic systems of the machine was made. The order of rearrangement of the supports during the motion and the trajectory of the support movement are shown. The design of sensors and their principle of operation are considered. The simulation bench with a description of its electronic components is demonstrated. Results. The optimal parameters of the support motion are determined. A cyclic algorithm for specifying the motion of a support along a trajectory consisting of rectilinear segments is described. The problem of synchronization of motion of a set of supports using multithreaded asynchronous programming adapted for multidimensional processors has been solved. The process of lowering the support to the surface and the response of the cyclic algorithm to changes in the shock and load sensor readings are simulated. Discussion and Conclusion. An algorithm for propulsion with reaction to changes in sensor readings has been developed. The conducted research allowed us to obtain an optimal algorithmic model of motion, to which it is easy to add new reactions of the automatic motion control system based on sensor readings.

In this article, we present a novel autocorrelation-based frequency estimation algorithm for single-tone sinusoidal signals. In comparison to other state-of-the-art frequency estimation methods, the proposed one provides a better tradeoff between accuracy, complexity, and estimation range. In particular, the algorithm is able to achieve the Cramer–Rao lower bound for moderate and high signal-to-noise ratio, and its implementation is feasible even in resource-constrained microcontrollers, as those commonly used in the Industrial Internet-of-Things (IIoT) applications and low-cost instrumentation. Finally, we investigate the performance of the algorithm in the case of a practical IIoT application, i.e., frequency estimation of unbalanced three-phase power systems, showing that it outperforms several other autocorrelation-based estimators.

Precise motor control requires high accuracy of the rotor position through the incremental encoder. The speed and accuracy of the acquisition equipment (microcontroller) play an important element in terms of cost and efficiency. In this paper, the author presents alternative methods for speed acquisition from an incremental encoder. In the first stage of research, the main performances of the STM32 microcontroller, connected with an incremental encoder, will be analyzed and compared with two different acquisition systems, i.e., ELVIS II and a Unidrive M701 power inverter. Using the LabVIEW graphical programming language, a user-friendly, convenient, and flexible human–machine interface is designed. Due to the advantages provided by the STM32 microcontroller in terms of processing power, cost, and programming interface, the obtained results are accurate and consistent. Through experimental testing and analysis, the speed acquisition is stable for both developed software algorithms used for ELVIS II and STM32 platforms. It is the aim of the paper to propose a useful speed acquisition tool in low-cost, high-accuracy prototyping applications.

We address the secure pairing of mobile devices based on accelerometer data under various transportation environments, e.g., train, tram, car, bike, walking, etc. As users commonly commute by several transportation modes, extracting session keys from various scenarios to secure the private network of user's devices or even the public network formed by devices belonging to distinct users that share the same location is crucial. The main goal of our work is to establish the amount of entropy that can be collected from these environments in order to determine concrete security bounds for each environment. We test several signal processing techniques on the extracted data, e.g., low-pass and high-pass filters, then apply sigma-delta modulation in order to expand the size of the feature vectors and increase both the pairing success rate and security level. Further, we bootstrap secure session keys by the use of existing cryptographic building blocks EKE (Encrypted Key Exchange) and SPEKE (Simple Password Exponential Key Exchange). We implement our proof-of-concept application on Android smart-phones and take benefit from numerical processing environments for the off-line analysis of the collected datasets.

A wrist-worn PPG sensor coupled with a lightweight algorithm can run on a MCU to enable non-invasive and comfortable monitoring, but ensuring robust PPG-based heart-rate monitoring in the presence of motion artifacts is still an open challenge. Recent state-of-the-art algorithms combine PPG and inertial signals to mitigate the effect of motion artifacts. However, these approaches suffer from limited generality. Moreover, their deployment on MCU-based edge nodes has not been investigated. In this work, we tackle both the aforementioned problems by proposing the use of hardware-friendly Temporal Convolutional Networks (TCN) for PPG-based heart estimation. Starting from a single "seed" TCN, we leverage an automatic Neural Architecture Search (NAS) approach to derive a rich family of models. Among them, we obtain a TCN that outperforms the previous state-of-the- art on the largest PPG dataset available (PPGDalia), achieving a Mean Absolute Error (MAE) of just 3.84 Beats Per Minute (BPM). Furthermore, we tested also a set of smaller yet still accurate (MAE of 5.64 - 6.29 BPM) networks that can be deployed on a commercial MCU (STM32L4) which require as few as 5k parameters and reach a latency of 17.1 ms consuming just 0.21 mJ per inference.

In the aerospace industry, commercial-off-the-shelf (COTS) static random access memories (SRAMs) are a cost-effective solution for obtaining high performance at the system level, which is difficult to obtain using space-qualified components. In addition, the usage of dynamic voltage scaling (DVS) is commonly used in space environments, where low power consumption is a critical issue. This article presents an analysis of the sensitivity against single-event upsets (SEUs) of various COTS bulk SRAMs and advanced low-power SRAMs (A-LPSRAMs) against proton radiation when using DVS to save power. Experimental results will show clear evidence that the sensitivity to SEUs increases when the power is lowered. Two sets of successive technologies (130-, 90-, and 65-nm bulk SRAMs, and 150- and 110-nm A-LPSRAMs) are evaluated against 15-MeV protons and compared with results of 14-MeV neutrons presented in a previous work. Experimental data are finally compared with analytical simulations obtained by using the multiscales single-event phenomena predictive platform (MUSCA-SEP3) Monte Carlo tool to predict the effect of DVS on the SEE sensitivity on more modern technologies in the ITRS/IRDS roadmap.

Введение. Статья посвящена исследованию автоматизации технологического процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков в биогазовой установке. Материалы и методы. Рассматриваются экспериментальные исследования автоматизации технологического процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков, в том числе сформулированные российскими учеными, анализируются общие требования, предъявляемые к методике и составу оборудования по автоматизации биогазовой установки. Приведена методика проведения экспериментальных исследований автоматизации процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков в биогазовой установке. Приведен состав оборудования, средства измерения и автоматизации. Разработан алгоритм работы биогазовой установки при проведении эксперимента и модернизация системы управления. Результаты. В лаборатории «Биотехнологии» проведены исследования автоматизации технологического процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков и их влияние на выработку биогаза. Разработана функциональная схема автоматизированного технологического комплекса биогазовой установки. При разработке автоматической системы управления был выбран однокристальный микроконтроллер семейства Mega, обладающий наилучшими характеристиками с точки зрения применения в управляющих системах. Представлены основные характеристики, а также описаны структура и архитектура микроконтроллера. Выбрана аппаратно-программная платформа Arduino Mega, которая обеспечивает удобство при программировании и подключения периферийных устройств к микроконтроллеру для управления биогазовой установкой в автоматическом режиме. Получены экспериментальные зависимости выработки метана в зависимости от времени выдержки в мезофильном и термофильном режиме анаэробного сбраживания биомассы. Обсуждение. Недостаточная теоретическая и практическая изученность автоматизации технологического процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков не может решить проблему энергосбережения и охраны окружающей среды. Предложена структурная схема автоматизации технологического процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков. Заключение. Разработана методика модернизации системы управления и автоматизации процесса анаэробного сбраживания осадков сточных вод и животноводческих стоков в биогазовой установке. Обоснована методика выбора аппаратно-программной платформы Arduino Mega 2560., Introduction. The article is devoted to the study of the automation of the technological process of anaerobic digestion of sewage sludge and livestock effluents in a biogas plant. Materials and methods. Experimental studies of the automation of the technological process of anaerobic digestion of sewage sludge and livestock effluents, including those formulated by Russian scientists, are considered, the general requirements for the methodology and composition of equipment for automation of a biogas plant are analyzed. The method of conducting experimental studies of automation of the process of anaerobic digestion of sewage sludge and livestock effluents in a biogas plant is given. The composition of equipment, measuring instruments and automation is given. The algorithm of operation of the biogas plant during the experiment was developed, and the modernization of the control system. Results. In the laboratory of Biotechnologies, studies have been conducted on the automation of the technological process of anaerobic digestion of sewage sludge and livestock effluents and their impact on biogas production. A block diagram of an automated technological complex of a biogas plant has been developed. To implement an automatic control system for a biogas plant, a microcontroller of the Mega family has been selected, manufactured using low-consumption CMOS technology, which, in combination with an improved RISC architecture, allows achieving the best ratio of performance and energy consumption. The main characteristics of the microcontroller processor are presented. The Arduino Mega 2560 hardware and software platform has been selected, which provides convenience when programming and connecting peripheral devices to a microcontroller for controlling a biogas plant in automatic mode. Experimental dependences of methane production depending on the exposure time in mesophilic and thermophilic modes of anaerobic fermentation of biomass are obtained. Discussion. Insufficient theoretical and practical knowledge of the automation of the technological process of anaerobic digestion of sewage sludge and livestock effluents cannot solve the problem of energy conservation and environmental protection. The structural scheme of automation of the technological process of anaerobic digestion of sewage sludge and livestock effluents is proposed. Conclusion. A method of modernization of the control system and automation of the process of anaerobic digestion of sewage sludge and livestock effluents in a biogas plant has been developed. The method of choosing the Arduino Mega 2560 hardware and software platform is substantiated.

This study proposes a heterogeneous hardware-based framework for network intrusion detection using lightweight artificial neural network models. With the increase in the volume of exchanged data, IoT networks’ security has become a crucial issue. Anomaly-based intrusion detection systems (IDS) using machine learning have recently gained increased popularity due to their generation’s ability to detect unseen attacks. However, the deployment of anomaly-based AI-assisted IDS for IoT devices is computationally expensive. A high-performance and ultra-low power consumption anomaly-based IDS framework is proposed and evaluated in this paper. The framework has achieved the highest accuracy of 98.57% and 99.66% on the UNSW-NB15 and IoT-23 datasets, respectively. The inference engine on the MAX78000EVKIT AI-microcontroller is 11.3 times faster than the Intel Core i7-9750H 2.6 GHz and 21.3 times faster than NVIDIA GeForce GTX 1650 graphics cards, when the power drawn was 18mW. In addition, the pipelined design on the PYNQ-Z2 SoC FPGA board with the Xilinx Zynq xc7z020-1clg400c device is optimised to run at the on-chip frequency (100 MHz), which shows a speedup of 53.5 times compared to the MAX78000EVKIT.

Wireless Sensor Networks (WSNs) are prone to highly constrained resources, as a result ensuring the proper functioning of the network is a requirement. Therefore, an effective WSN management system has to be integrated for the network efficiency. Our objective is to model, design, and propose a homogeneous WSN hybrid architecture. This work features a dedicated power utilization optimization strategy specifically for WSNs application. It is entitled Hybrid Energy-Efficient Power manager Scheduling (HEEPS). The pillars of this strategy are based on the one hand on time-out Dynamic Power Management (DPM) Intertask and on the other hand on Dynamic Voltage and Frequency Scaling (DVFS). All tasks are scheduled under Global Earliest Deadline First (GEDF) with new scheduling tests to overcome the Dhall effect. To minimize the energy consumption, the HEEPS predicts, defines and models the behavior adapted to each sensor node, as well as the associated energy management mechanism. HEEPS’s performance evaluation and analysis are performed using the STORM simulator. A comparison to the results obtained with the various state of the art approaches is presented. Results show that the power manager proposed effectively schedules tasks to use dynamically the available energy estimated gain up to 50%.

Patient monitoring requires the acquisition of increasingly larger amounts of biosignal data that needs to be managed and transferred with minimum energy consumption. As huge quantities of data are often very redundant, it is possible to reduce their size directly on the edge of the system, right after the acquisition. To do so, subspace analysis can be considered a fundamental tool that can be used to significantly reduce the size of high-dimensional data, thus minimizing the requirements for data transfer. The problem of these methods is that they often come with big memory and computation requirements, as they are ultimately equivalent to the expensive eigenspace evaluation. In order to use subspace analysis methods with the minimum requirements in terms of cost and energy consumption, we here rely on two specialized streaming algorithms for the estimation of the subspace of electroencephalogram (EEG) signals directly after the acquisition on edge devices. The implementation of these state-of-the-art algorithms is tested on a common low-end microcontroller unit (MCU), which is an ideal candidate as edge computing digital hardware platform. The functional performance of these methods is evaluated along with the requirements in terms of computational time, energy consumption and memory footprint.

The key for decentralized battery systems is a robust and communication-less control strategy for autonomous power sharing of parallel-connected DC-DC converters. Battery systems improve the reliability and quality of power supply in renewable energy systems and enable power supply for off-grid, mobile applications, including islanded grids, home storage, and electric vehicles. In many cases, components with different elec trical properties require different voltage levels. An adaptation is consequently essential and is normally implemented in DC grids for the batteries via bidirectional DC-DC converters. The power flow in both directions can thus be ensured. To achieve a power distribution in parallel connected DC-DC converters, a droop control in the form of a virtual internal resistor can be used. This paper presents a novel approach of a DC DC converter with a digitally parameterizable droop resistor, whose voltage regulation is based on an analog operational amplifier circuit to ensure low delays and robustness. The droop resistor is adjusted with a microcontroller, which offers the possibility to apply a higher-level control for load sharing via an interface. Mathematical correlations are used to clearly define the parameters of the control. Furthermore, the circuit was completely simulated and tested in the hardware setup. The shown results verify the functionality and indicate only minor deviations. Therefore, this circuit is important for future use in distributed battery systems.

A wireless battery charger with inductive power transfer (IPT) was proposed in this paper. The commonly used constant-current constant-voltage (CC-CV) charging method is accomplished by a closed-loop controlled IPT with a hybrid resonant circuit on the secondary side. A smooth transition between the CC stage and the CV stage can be made simply by swapping exactly the associated switches on resonant capacitors. The required charging voltage and current are regulated by controlling the phase-shifted angle of the high-frequency inverter on the primary side. To stabilize the charging current and voltage, a closed-loop digital controller was introduced with infrared (IR) transmission feedback. Precise regulation of the resonant inverter on a relative small ranged phase-shifted angle can be realized by two 16-bit microcontroller units (MCUs) with compact encoding and decoding techniques. A hybrid resonant inverter was designed for a 600 W prototype of the proposed IPT battery charger. Experimental results from exemplar cases have demonstrated that the battery charger can provide a stable charging current at the CC stage and then transit smoothly into the CV stage.

This article presents a deep theoretical analysis of the Active Disturbance Rejection Control (ADRC) regulator for the control of first-order systems, directly compared to a “traditional” Proportional Integral (PI) regulator. To complete the theoretical study, ADRC and PI are implemented into the model of a single-phase Dual Active Bridge (DAB) converter to regulate the voltage of a Direct Current (DC) network. Facing different types of disturbances and DC network parameters variations, strengths and weaknesses of the two controllers are highlighted. ADRC and PI controls are discretized and implemented in Control Hardware In the Loop (CHIL) simulations of a single-phase DAB converter to regulate the voltage of a node of multi-terminal and multi-level DC network. By changing the DAB connection points along the MVDC network, the controlled system is stressed with different disturbances, extending the result of single-terminal network software simulations.

Orientador: Fabiano Fruett Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação Resumo: Existem diversos sistemas e equipamentos, com finalidades específicas, envolvendo o monitoramento das mudanças de clima e de tempo. Entretanto, a maioria desses sistemas e equipamentos disponíveis no mercado possuem alto custo, o que normalmente inviabiliza sua utilização em larga escala. Além disso, a maioria dos sistemas e equipamentos para este fim são importados, o que dificulta ainda mais a sua inserção em aplicações corriqueiras. Este trabalho descreve o desenvolvimento, projeto, testes e aplicação de uma rede de sensores sem fio de baixo custo, para monitoramento em tempo real da temperatura ambiente, umidade relativa do ar e pressão atmosférica, como uma alternativa nacional para pesquisas que necessitem de equipamentos para monitoramento ambiental. A rede é composta de quatro nós sensores e um nó coordenador. Os nós sensores possuem um sensor de temperatura e umidade e um sensor de pressão barométrica ¿ ambos com interface digital ¿ além de um microcontrolador de baixo consumo de energia e baixo custo, e um módulo de comunicação para uma rede operando na frequência de 2,4GHz, para conexão numa rede padrão IEEE 802.15.4. O nó coordenador possui outro módulo de comunicação para a conexão na rede sem fio, um microcontrolador, um cartão de memória para armazenamento das informações de cada nó sensor e interface Ethernet para disponibilizar as medidas do sensor em uma página web, acessada através de um IP estático em um browser de internet comum. Testes em laboratório e em campo atestaram autonomia do nó sensor (em torno de 90 dias para uma bateria de 500mAh) e alcance (de aproximadamente 150m em visada direta) e comprovaram a conformidade dos valores medidos por cada nó, através da comparação com um sistema de referência Abstract: There are several systems and equipments with specific applications related to climate and weather change monitoring. However, most of these systems and equipments available in the market are too expensive ¿ which usually prevents their use in large scale. Furthermore, most of them used for such purpose are imported, which further complicates their insertion in everyday applications. This work describes the development, implementation and application of a low-cost wireless sensor network for temperature, relative humidity and atmospheric pressure monitoring, in real-time as a national alternative for researches which demand equipments for environmental monitoring. The network consists of four sensor nodes and a coordinator node. Sensor nodes have a temperature and humidity sensor and a barometric sensor ¿ both with digital interface ¿ plus a low cost and low power microcontroller and a communication module for the wireless network operating at 2.4GHz for establishing connection to an IEEE 802.15.4 network. Coordinator node has another communication module to the wireless network connection, a microcontroller for receiving measurements, a memory card for storage of each sensor node¿s information, and Ethernet interface to provide sensor node¿s information in an embedded webpage, accessed through a static IP in a common internet browser. Tests performed in laboratories and in the field attested the node¿s autonomy (around 90 days for a 500mAh battery) and reach (approximately 150m line-of-sight), and proved the compliance of measurements performed by each sensor node, through comparison with a reference system Mestrado Eletrônica, Microeletrônica e Optoeletrônica Mestre em Engenharia Elétrica

Orientador: Talía Simões dos Santos Ximenes Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Tecnologia Resumo: Hoje, o interesse em explorar a utilização de radiação solar vem adquirindo grandes proporções, abrangendo o aproveitamento da radiação como fonte de energia limpa e renovável. O Brasil apresenta grande potencial de energia solar em quase todo seu território durante o ano, tal característica contribui para o uso de sistemas de rastreamento solar com o intuito de aumentar a eficiência na captação de energia solar. Porém, a implementação de tal sistema apresenta custo elevado em território brasileiro. Considerando isso o presente trabalho teve como objetivo desenvolver um sistema elétrico mecânico com correção automática de ângulo dos painéis solares para otimização da captação de energia, acoplado ao um sistema de monitoramento com acesso remoto. Foi construído um rastreador solar ativo, que se orienta em direção ao Sol com o intuito de obter maior eficiência energética. Para isso, foram utilizados três painéis solares simples. Sendo um painel fixo (P1), o outro painel com rotação de apenas um eixo (P2), e o último com rotação em dois eixos (P3). Os painéis P2 e P3 foram conectados a um sistema de controle embarcado que movimentava os painéis por temporizador. Os dados coletados foram enviados, em tempo real, para um servidor que os armazenou para posteriores análises. O sistema de rastreamento solar funcionou em um período de 9 meses, contemplando as 4 estações do ano (final do inverno, primavera, verão e início de outono). Concluiu-se com o presente projeto que é possível otimizar a captação energética por placas fotovoltaicas com sistema de rastreamento solar, obtendo eficiência energética de aproximadamente 60% superior aos painéis fixos convencionais. Esse incremento na eficiência energética do painel móvel de dois eixos (P3) é resultado do aumento do tempo de captação energética ao longo do dia, uma vez que tal painel apresentou alta captação solar desde o nascer até o pôr-do-sol Abstract: Today, the interest in exploring the use of solar radiation has acquired great proportions, including the use of radiation as a source of clean and renewable energy. Brazil has great potential for solar energy in almost all its territory during the year, this characteristic contributes to the use of solar tracking systems in order to increase the efficiency in the capture of solar energy. However, the implementation of such a system has a high cost in Brazil. Considering this the present work had as objective to develop a mechanical electrical system with automatic correction of angle of the solar panels for optimization of the power capture, coupled to a monitoring system with remote access. An active solar tracker has been built, which is oriented towards the Sun in order to obtain greater energy efficiency. For this, three simple solar panels were used. Being a fixed panel (P1), the other panel with one axis rotation (P2), and the other with two axis rotation (P3). Panels P2 and P3 were connected to an embedded control system that moved the panels by timer. The collected data was sent in real time to a server that stored it for further analysis. The solar tracking system operated over a period of 9 months, covering the 4 seasons of the year (late winter, spring, summer and early fall). It was concluded with the present project that it is possible to optimize the energy uptake by photovoltaic panels with solar tracking system, obtaining energy efficiency of approximately 60% higher than conventional fixed panels. This increase in the energy efficiency of the two-axis mobile panel (P3) is a result of the increase in the energy capture time throughout the day, since this panel showed high solar pickup from sunrise to sunset Mestrado Sistemas de Informação e Comunicação Mestre em Tecnologia CAPES

Orientador: José de Alencar Simoni Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Química Resumo: Neste trabalho, utilizando uma coleta automatizada de dados, realizaram-se experimentos tradicionais de Química, como destilação fracionada de mistura água/etanol e água/acetona, reação oscilante Briggs-Rauscher, cinética da decomposição de tiossulfato catalisada por ácido e entalpia de decomposição catalítica de peróxido de hidrogênio. Para esse conjunto de experimentos, utilizou-se uma coleta local de dados, em que o objetivo era adequar as condições de medida à instrumentação alternativa, que foram utilizadas para a realização de experimentos remotos. A instrumentação alternativa significa o uso de pequenos dispositivos baratos e facilmente encontrados no mercado, como: um sensor de temperatura LM35, laser vermelho, sensor LDR, sensor de temperatura DS18B20, Plataforma Arduino, ESP8266-NodeMCU e Módulo WiFi ESP8266. Na primeira etapa, um sensor de temperatura e um sistema laser-LDR foram acoplados à placa de aquisição, e os dados foram coletados usando o software da placa e, posteriormente, analisados usando um software de processamento de dados. Com os sistemas desenvolvidos nessa etapa, quatro experimentos foram realizados remotamente, via Internet, usando o site do YouTube e a plataforma Moodle. Nessa etapa, todos os experimentos foram transmitidos a estudantes de diferentes disciplinas, e posteriormente avaliados por questionários qualitativos e quantitativos aplicados aos alunos. Em todos os experimentos os resultados experimentais concordaram adequadamente com os da literatura, com precisão e exatidão compatíveis com o que se espera no ensino experimental convencional. A aplicação dos questionários aos alunos que vivenciaram os experimentos remotamente mostrou uma excelente aceitação por parte dos mesmos, evidenciando a positividade na utilização do recurso da experimentação remota Abstract: In this work, using automated data collection, we carried out traditional chemistry experiments, such as fractional distillation of water / ethanol and water / acetone mixture, Briggs-Rauscher oscillating reaction, kinetics of acid catalysed thiosulfate decomposition and enthalpy of catalytic decomposition hydrogen peroxide. For this set of experiments, a local data collection was used, in which the objective was to adapt the measurement conditions to the alternative instrumentation, that were used to carry out remote experiments. Alternative instrumentation means the use of small, inexpensive devices easily found in the market, such as: a LM35 temperature sensor, red laser, LDR sensor, DS18B20 temperature sensor, Arduino platform, ESP8266-NodeMCU and ESP8266 Wi-Fi module. In the first step, a temperature sensor and a laser-LDR system were coupled to the acquisition board, and data were collected using the board software and then analyzed using data processing software. With the systems developed in this stage, four experiments were performed remotely, via the Internet, using the YouTube site and the Moodle platform. In this stage, all the experiments were transmitted to students from different disciplines, and later evaluated by qualitative and quantitative questionnaires applied to them. In all experiments, the experimental results were in agreement with those of the literature, with precision and accuracy compatible with what is expected in conventional experimental teaching. The application of the questionnaires to the students who experienced the experiments remotely showed an excellent acceptance on the part of them, evidencing the positivity in the use of the remote experimentation resource Doutorado Físico-Química Doutor em Ciências

Ultra-low-power microcontrollers (ULP MCUs) face a performance trade-off between energy-efficient computing during activity periods and low sleep power, associated with limited wake-up time and energy. Adaptive back-biasing in FD-SOI, along with near-threshold operation at ultra-low voltage, has brought significant improvements by dynamically shifting the minimum energy point (MEP) along the frequency axis. This work introduces a highly-integrated 64-MHz ULP Cortex-M4 MCU with 96-kB SRAM in 28nm FD-SOI. A clock and power management unit (CPMU) generates all internal supplies and clocks from a 1.8-V supply, while unified frequency and back-bias regulation (UFBBR) performs PVT compensation. Custom 16-kB ULP SRAMs achieve low read/write access energy, 1.2/0.84pJ/32-bit access respectively, and provide 0.98nW/kB ultra-low-leakage data retention. A low-power biomedical analog front-end enables biopotential monitoring. The MEP is 5.5µW/MHz (8.2µW/MHz including conversion losses). Sleep power is 7.7µW with retention of logic state and 32-kB memory.

Smart homes/offices based on wireless sensor networks (WSNs) can provide an assisted living and working environment to the users. In these applications, the distributed network nodes are made up of low-power low-cost high-energy-efficient electronic platforms equipped with sensors, microcontroller, radio, and antenna, able to periodically sense, receive, store, pre-process, and transmit ambient data to a remote host station. Conventional nodes are usually supplied by batteries, resulting in a significant limitation to the lifetime and to the maximum number of deployable devices. To meet the demand of the next Internet-of-Things (IoT) applications, requiring a vast plurality of interconnected wireless network nodes, this paper presents the design and implementation of a WSN platform whose nodes are energetically autonomous thanks to an embedded photovoltaic panel associated to a rechargeable battery and a power-efficient design with optimized power-management strategy. The implemented node is able to harvest indoor ambient light starting from 100 lux and, according to the available energy, adaptively sets the sensors acquisition and RF transmission rate. Moreover, it provides long-distance data transmission with air data rate from 1 to 500 kbps. The WSN node device is implemented on an $8.6\times 5.4$ cm2 flexible PCB, being therefore amenable to conform even to curved surfaces. Comparison with the commercial IoT nodes reveals a significant improvement in the state of the art.

9th International Conference on Hydrogen Production (ICH2P), WOS: 000466455000007, Software and hardware advancements on a hydroxy gas (HHO) production system is presented in this experimental study. A unique control system that enables to generate hydroxy gas at the specified volumetric flow rates with pulse width modulation (PWM) technique is developed and manufactured. The novelty of the method is automatic control of the hydroxy production according to the instantaneous demand, in other words, the adaptation of the HHO system to the operating principles of internal combustion engines (ICEs). A stepping motor represented the vehicle engine in order to get variable rotational speed values. The instantaneous rotation speed range of the motor divided into four sections as no rotation, low (1000-2001 rpm), middle (2001-4001 rpm) and high (>4001 rpm). A computer-driven microcontroller (MCU) card managed the entire system. The MCU card received motor speed input parameters and identified the PWM signal. Results of the study which are the electrical current, hydroxy gas flow and dry cell temperature ranges are presented in tables and graphs. The measured experimental data were 0-15 A, 0-1.56 L per minute and 24-48 degrees C for electrical current, hydroxy gas production and dry cell temperature ranges, respectively. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

The recent phenomenon known as the Internet of Things (IoT) shows us a glimpse into a future where people rely on smart connected devices to solve various problems. However, people have used IoT devices primarily in dense, urban areas, and homes because they rely on existing electronic network technologies such as Ethernet, WiFi, and Bluetooth. These technologies depend on devices’ proximities to each other to transmit data reliably. With the emergence and ubiquity of novel wireless networking technologies such as Zigbee and LoRa (with the proper antennas) and ultra-low-power hardware, it is now possible to use IoT devices remotely in rural areas to gather data from a broader geographical surface area. Doing so enables more excellent coverage and data acquisition from locations that were previously difficult to monitor. The emergence of these new technologies enables engineers to solve an even greater host of problems based on wireless technologies’ increased capabilities and reliability. The sensor system that we aim to design is a distributed IoT network that can detect and monitor hazardous conditions, such as wildfires, remotely to help humans respond to these threats and prevent large scale fires and other infrastructure damages.

Teaching electronics can be fun and motivating when you give the students the freedom to develop their own projects but of course, with a proper supervision. This work presents a case study where the teaching of microcontrollers in the third course of the MSc in Industrial Engineering, specialty in electronics, has been moved away from a strictly academic point of view towards a project-based learning (PBL) course. Here, the students are supposed to acquire a basic microcontroller knowledge in a few lecture sessions and autonomous work in order to pass rapidly to the hands-on training and activities and the following autonomously developed project. Proposed approach has enabled to increase the student engagement and motivation during the whole semester as well as improve the student average results compared to previous cases. Moreover, the satisfaction of the students and the word of mouth from student to student has permitted to increase the number of students of the specialty of electronics in the MSc in Industrial year after year. Students' outcomes and satisfaction surveys have been analysed for seven consecutive courses.

Orientadores: Francisco José Arnold, Rangel Arthur Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Tecnologia Resumo: A medição da impedância elétrica em transdutores piezelétricos é um procedimento fundamental para a caracterização desses dispositivos. A larga variação do módulo da impedância elétrica ao longo da frequência representa uma dificuldade para a realização dessas medições. Neste trabalho é apresentado o protótipo de um medidor de impedâncias elétricas de transdutores de ultrassom utilizados em aplicações de potência elevada. O sistema opera baseado na captação dos valores de pico de tensão e corrente sob uma faixa de frequência de utilização dos transdutores. O protótipo é controlado por um microcontrolador que se comunica com um computador no qual um software, desenvolvido em LabView, armazena e mostra resultados numa interface. A validação do medidor de impedâncias foi feita pela comparação de resultados experimentais com os medidos por um analisador de impedâncias comercial. Os resultados mostram que o sistema proposto pode medir impedâncias entre 100 W e 10 kW, aproximadamente, numa faixa de frequências de até 43 kHz Abstract: The measurement of electrical impedance of piezoelectric transducers is a fundamental procedure for the characterization of these devices. The large variation in the electrical impedance module over frequency is a difficulty to carry out these measurements. This dissertation presents the prototype of a meter electrical impedance of ultrasound transducers used in high power applications. The system operates based on the capture of peak voltage and current values of the transducers in a usage frequency band. The prototype is controlled by a microcontroller that communicates with a computer on which a software, developed in LabView, stores and displays the results in an user interface. The validation of the impedance meter was made by comparison with the experimental results measured by a commercial impedance analyzer. The results show that the proposed system can measure impedances between 10 W and 10 kW at frequencies up to 43 kHz Mestrado Tecnologia e Inovação Mestre em Tecnologia

Orientadores: Eduardo Tavares Costa, Francisco Ubaldo Vieira Júnior Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação Resumo: Circulação extracorpórea é um processo no qual o sangue circula externamente ao corpo com o intuito de manter a oxigenação e o fluxo do sangue adequados durante a realização de determinados procedimentos cirúrgicos; para realizá-la é utilizado um conjunto de técnicas e equipamentos cuja função é substituir temporariamente o coração e os pulmões. Contudo, a circulação extracorpórea é percebida pelo organismo como um agente agressor e um dos principais parâmetros relacionados aos danos que ela provoca é a hemólise. O uso de bombas propulsoras é um dos causadores de hemólise em procedimentos envolvendo circulação extracorpórea e um dos principais tipos de bombas propulsoras é a bomba de roletes. A hemólise provocada pelo uso de bomba de roletes é devida principalmente ao grau de oclusão utilizado. Há dois métodos geralmente utilizados para calibração da bomba de roletes: o método de velocidade de queda e o método de calibração dinâmica. Nesse trabalho é proposto um dispositivo para auxiliar o perfusionista a ajustar bombas de roletes por um método menos oclusivo, a calibração dinâmica. O dispositivo é baseado em um microcontrolador PIC 18F4523 e é utilizado em conjunto com transdutores de pressão descartáveis normalmente utilizados em procedimentos cirúrgicos. Foram determinadas as curvas características de três transdutores de pressão na faixa de 0 a 1000 mmHg. Os resultados das calibrações realizadas com o protótipo desenvolvido foram comparados com os resultados obtidos com uma placa de aquisição de dados comercial. Dois, dos três transdutores, foram submetidos a testes de fadiga e um foi mantido como referência. Para validação do dispositivo, sendo utilizado para realizar calibração dinâmica, foram testados ajustes em 150, 250, 350, 450 e 500 mmHg. Em cada ajuste foram realizadas 8 medições simultâneas de pressão média de calibração dinâmica com o dispositivo e com uma placa de aquisição de dados comercial. Após a realização de todos os testes, os transdutores foram novamente caracterizados. As curvas características dos três transdutores obtidas na caracterização inicial mostraram igualdade estatística (p > 0,05). Após os testes de fadiga não foram observadas alterações nas respostas dos transdutores de pressão no início e após a realização dos testes, nos mesmos pontos de pressão (p > 0,05). As medidas de pressão de calibração dinâmica realizadas com o protótipo apresentaram igualdade estatística (p > 0,05) para toda a faixa de pressão testada, quando comparados com os respectivos resultados obtidos com a placa de aquisição de dados. Conclui-se que os transdutores utilizados atualmente em procedimentos cirúrgicos podem ser utilizados na calibração dinâmica sem perdas de características e que o dispositivo construído pode ser utilizado em ambiente operatório para ajustes de bombas de roletes pelo método de calibração dinâmica Abstract: Cardiopulmonary bypass is a procedure in which blood circulates outside the body in order to maintain oxygenation and blood flow conditions during the performance of certain surgical procedures; for this procedure, it is used a set of equipments whose function is to temporarily replace the heart and lungs. However, the cardiopulmonary bypass is perceived by the body as an aggressor agent and one of the main parameters related to the damage it causes is hemolysis. The use of driving pumps is a cause of hemolysis in procedures involving cardiopulmonary bypass and one of the main types of driving pumps is the roller pump. Hemolysis caused by the use of roller pump is mainly due to the degree of roller occlusion used. There are two methods commonly used for adjusting the roller pump: the method of drop rate and the dynamic calibration method. In this work, we developed a device to assist the perfusionist adjust roller pumps by a less occlusive method, the dynamic calibration method. The device is based on a PIC 18F4523 microcontroller and is used in conjunction with disposable pressure transducers commonly used in surgical procedures. We have characterized three pressure transducers in the range of 0 to 1000 mmHg. Calibration results obtained with the prototype were compared with those obtained with a data acquisition board. Two of the three transducers were submitted to fatigue tests and one was kept as a reference. To validate the constructed device, being used to perform dynamic calibration, adjustments were at 150, 250, 350, 450 and 500 mmHg, in each set were performed eight simultaneous measurements of dynamic calibration mean pressure with the device and with a data acquisition board. After all tests were conducted, the transducers were re-characterized. The three characteristic curves obtained in the initial characterization showed statistically equivalence (p > 0.05) among themselves. After the fatigue tests no changes were observed in the responses of pressure transducers in the beginning and in the end of the tests, at the same pressure points (p > 0.05). Measurements of dynamic calibration pressure obtained with the prototype showed statistically equivalence (p > 0.05) throughout the tested pressure range, when compared with the results obtained with the data acquisition board. We conclude that the transducers currently used in surgical procedures may be used for dynamic calibration without losing their characteristics and that the constructed device can be used in surgery to adjust the roller pumps by the dynamic calibration method Mestrado Engenharia Biomédica Mestre em Engenharia Elétrica

This article aims to describe the design of a circuit of a level transducer implemented by means of an ultrasonic sensor and controlled by Arduino Nano, applied to a water tank of a firefighting system. Initially, the integration of the Siemens 1212C programmable logic controller is described, in the connection between the sensor, the controller and the interfaces that allow to generate the monitoring, control and data recording, conditioned by a PWM pulse width modulated signal controlled by Arduino Nano. When developing the research and performing an analysis of the linear regression model, it is established that the behavior of the controlled variable with respect to time, generates a linear voltage response in the range of 0 to 10 volts; expressing in terms of correlational relationship a factor of R2 equal to 0.997, thus establishing that the designed transducer does not show susceptibility to noise or disturbances in the start-up of the firefighting system. Campus Lima Centro

In the last few years, research and development on Deep Learning models and techniques for ultra-low-power devices in a word, TinyML has mainly focused on a train-then-deploy assumption, with static models that cannot be adapted to newly collected data without cloud-based data collection and fine-tuning. Latent Replay-based Continual Learning (CL) techniques[1] enable online, serverless adaptation in principle, but so farthey have still been too computation and memory-hungry for ultra-low-power TinyML devices, which are typically based on microcontrollers. In this work, we introduce a HW/SW platform for end-to-end CL based on a 10-core FP32-enabled parallel ultra-low-power (PULP) processor. We rethink the baseline Latent Replay CL algorithm, leveraging quantization of the frozen stage of the model and Latent Replays (LRs) to reduce their memory cost with minimal impact on accuracy. In particular, 8-bit compression of the LR memory proves to be almost lossless (-0.26% with 3000LR) compared to the full-precision baseline implementation, but requires 4x less memory, while 7-bit can also be used with an additional minimal accuracy degradation (up to 5%). We also introduce optimized primitives for forward and backward propagation on the PULP processor. Our results show that by combining these techniques, continual learning can be achieved in practice using less than 64MB of memory an amount compatible with embedding in TinyML devices. On an advanced 22nm prototype of our platform, called VEGA, the proposed solution performs onaverage 65x faster than a low-power STM32 L4 microcontroller, being 37x more energy efficient enough for a lifetime of 535h when learning a new mini-batch of data once every minute., Comment: 14 pages

En este proyecto se expone como se ha llevado a cabo la construcción de un dispensador automático de comida para animales domésticos a partir de una placa de ordenador "single-board'' y otros componentes electrónicos como sensores de peso y electroimanes. El propósito de este proyecto no es solo llevar a cabo el sistema mencionado sino hacerlo con el menor coste posible. Es por eso que, tanto la placa en sí (una Raspberry Pi Zero WH) como el resto de dispositivos han sido especialmente elegidos por su bajo precio y por su capacidad de integración al sistema. Las utilidades que se le pueden dar al dispensador no solo se basan en la propia alimentación de los animales sino también a la monitorización de esa alimentación y el control de dietas de estos: con el dispensador automatizado se puede controlar, desde una aplicación programada para Android, cuánto han comido los animales en los últimos días, u obtener una media mensual o semanal, que servirá para posteriormente extraer conclusiones como la de si un cambio de comida (p. ej. comprar un pienso A en vez de un pienso B) ha afectado realmente a la frecuencia con la que los animales comen o a la velocidad que lo hacen. This project shows how the construction of an automatic pet food dispenser has been carried out from a single-board computer and other electronic components such as weight sensors and electromagnets. The purpose of this project is not only to carry out the aforementioned system but to do it at the lowest possible cost. That is why both the board itself (a Raspberry Pi Zero WH) and the rest of the devices have been specially chosen for their low price and for their ability to integrate into the system. The utilities that can be given to the dispenser are not only based on the feeding of the animals but also on the monitoring of this feeding and the control of their diets: with the automated dispenser it can be controlled, from an application programmed in Android, how much the animals have eaten in the last days, or obtain a monthly or weekly average, which will serve to later draw conclusions such as whether a change of food (e.g. buying feed A instead of feed B) has really affected how often animals eat or how fast they do it. En aquest projecte s'exposa com s'ha dut a terme la construcció d'un dispensador automàtic de menjar per animals domèstics a partir d'una placa d'ordinador "single-board'' i altres components electrònics com sensors de pes i electroimants. El propòsit d'aquest projecte no és només dur a terme el sistema esmentat sinó fer-ho amb el menor cost possible. És per això que, tant la placa en si (una Raspberry Pi Zero WH) com la resta de dispositius han estat especialment escollits pel seu baix preu i per la seva capacitat d'integració en el sistema. Les utilitats que se li poden donar al dispensador no només es basen en la pròpia alimentació dels animals sinó també a la monitorització d'aquesta alimentació i el control de dietes d'aquests: amb el dispensador automatitzat es pot controlar, des d'una aplicació programada per Android, quant han menjat els animals en els últims dies, o obtenir una mitjana mensual o setmanal, que servirà per a posteriorment extreure conclusions com la de si un canvi de menjar (p. ex. comprar un pinso A en comptes d'un pinso B) ha afectat realment a la freqüència amb què els animals mengen o la velocitat que ho fan.

Aquest document descriu el desenvolupament hardware i software d'un lector de llibres electrònics simple. El dispositiu es basa en un microcontrolador ARM i una pantalla electroforètica, és portable i es monta a una carcassa impresa en 3D. Pot llegir documents markdown des d'una SD a través d'un explorador d'arxius i mostrar el seu contingut. Tot no ser actualment compatible amb altres formats d'arxiu, pot servir de base per a futures extensions. This document describes the development of a simple ebook reader, both software and hardware-wise. The device is based on an ARM microcontroller and an electrophoretic display, is portable and fits in a 3D printed case. It can read markdown documents from an SD card through a file explorer and show its content. Although its currently limited file compaibility, it could be extended to support other formats in the future. Este documento describe el desarrollo hardware i software de un lector de libros electrónicos simple. El dispositivo se basa en un microcontrolador ARM y una pantalla electroforética, es portátil i se monta en una carcasa impresa en 3D. Puede leer documentos markdown desde una SD a traves de un explorador de archivos, y mostrar su contenido. Aunque no es compatible con otros formatos de archivo, puede servir de base para futuras ampliaciones.

In the last few years, microcontrollers became more and more powerful, and many authors have started to use them for different machine learning projects. One of the most popular frameworks for machine learning is TensorFlow, and their authors began to develop this framework for microcontrollers. The goal of this paper is to analyses the full connected neural networks inference speed depending on the number of neurons of one popular microcontroller (Arduino Nano 33 BLE Sense) with simple neural networks implementation, as well as the impact of neural network weights quantisation. We expected a reduction in the size of the model with the selected quantization by four times, which was achieved, but with a large number of neurons in the neural network. TensorFlow Lite for Microcontrollers is used with the Arduino Integrated Development Environment. Neural networks with two hidden layers are used with a different number of neurons.