ESTiG - Publicações em Proceedings Indexadas à WoS/Scopus
URI permanente para esta coleção:
Navegar
Entradas recentes
- A virtual reality-based learning environment for human - robot collaboration training in construction 4.0Publication . Sabiri, Khadija; Afonso, Luís; Camargo, Caio; Gonçalves, Estefânia; Fernandes, RuiThe rapid digitalization of the construction sector under the Construction 4.0 paradigm demands new training approaches for human-robot collaboration (HRC). Integrating HRC with Virtual Reality (VR) offers a powerful means to enhance workforce competence, safety, and efficiency through immersive and realistic simulations of construction scenarios. This paper presents the design and evaluation of a VR-based Virtual Learning Environment (VLE) developed in Unity 3D to train construction professionals for safe and effective human-robot teamwork. The platform comprises four integrated modules: (i) safety fundamentals, (ii) robot familiarization, (iii) immersive collaboration scenarios, and (iv) performance assessment. A virtual prototype of a construction robot and typical interaction workflows were modeled to simulate real-world operations. An experimental study involving 10 construction professionals evaluated the system using a mixed-methods approach grounded in the Technology Acceptance Model (TAM) and objective task performance metrics. Results demonstrate strong perceived usefulness and ease of use, with participants reporting improved communication, safety awareness, and trust in robotic co-workers. The findings highlight the potential of VR-based training to enhance workforce readiness and advance safe, collaborative practices in the era of Construction 4.0.
- A proposal for a mobile application pipeline for posture-based muscle rehabilitation at homePublication . Lopes, Júlio Castro; Van-Deste, Isaac; Lopes, Rui PedroThis work outlines the development process of a rehabilitation system from smartphones to provide remote monitoring and real-time feedback assistance in muscle rehabilitation exercises. The system leverages advanced machine learning techniques, such as pose estimation techniques and anomaly detection models, in tracking users’ movements and assessing the performance of exercises. The system is designed to identify deviations from correct movement patterns and provide corrective feedback, allowing patients to engage in home-based rehabilitation withoutspecialized hardware. This paper details the system design, data acquisition methods, and the machine learning models used to detect incorrect movements. The outlined pipeline is intended to facilitate the creation of an affordable, cost-effective, and scalable rehabilitation platform for strengthening patient adherence to exercises and optimizing recovery outcomes.
- End-of-life waste from photovoltaic systems: current practices, challenges, and opportunitiesPublication . Silva, Margarete Heleno da; Boloy, Ronney; Ferreira, Ângela P.The rapid growth of photovoltaic (PV) solar energy has introduced significant challenges related to the management of End-of-Life (EoL) panel waste. Projections estimate that by 2050, global PV waste could exceed 60 million tons. The absence of specific regulatory frameworks and the limited development of recycling infrastructure further aggravate the situation, underscoring the urgent need for sustainable waste management practices and the integration of circular economy principles within the solar industry. This paper discusses existing regulatory policies across various countries, along with recent technological advances in recycling processes. Furthermore, it emphasizes the importance of implementing a circular economy-based waste management model to enable the recovery of valuable materials, reduce reliance on virgin raw materials, and mitigate the environmental impact of the PV sector.
- Markov transition field for fall detection using time-series dataPublication . Kalbermatter, Rebeca B.; Silva, Felipe G.; Pereira, Ana I.; Valente, António; Lima, José; Yahiaoui, Réda; Fayad, MoustafaFall detection systems have traditionally relied on sequential pattern recognition methods, using, for example, time series data obtained from inertial sensors, such as accelerometers. This paper proposes a methodology for fall detection based on converting time series from accelerometer sensors into visual representations using the Markov Transition Field (MTF) method. The UP-Fall dataset was used to test the performance of a Convolutional Neural Network (CNN) model trained on the MTF images generated. A systematic analysis of the image generation parameters was carried out, including the window size, the percentage of overlap, and the number of bins used in the discretizations. The experiments showed that the configuration with 55 bins, a window of 200 samples, and 40% overlap resulted in the best accuracy (97.13%), demonstrating that the conversion of sensory signals into MTF images is a promising alternative for fall detection, allowing computer vision models to capture relevant temporal patterns with high efficiency.
- A low power photovoltaic water pumping system based on a DC-DC step-up converter and standard frequency convertersPublication . Fey, Alice Nogueira; Leite, Vicente; Romaneli, Eduardo Felix RibeiroPhotovoltaic Water Pumping Systems (PVWPS) are a solution to supply water to populations living in arid and remote regions. One problem is that these systems are usually sold as closed kits and with solar energy dedicated equipment. This situation makes it difficult to replace damaged equipment for others that would be available on the market. Also, solar dedicated equipment are more expensive than the general-purpose ones. Another issue is the oversizing of the photovoltaic string in terms of power, when installing low power PVWPS that employ conventional AC water pumps. The oversizing occurs in order to achieve the voltage requirements of these pumps (rated with less than 1kW). In the interest of solving these problems, this work proposes a solution for a low power PVWPS. The proposed system employs a maximum of 4 photovoltaic modules; a DC-DC step-up converter; a standard frequency converter and an 1HP AC water pump. It was designed and tested a 750W DC-DC step-up converter with a static gain of 3.44 for composing the proposed solution. The whole set of the low power PVWPS was tested in a laboratory environment and showed a full day of operation. The present work was held in cooperation with the company VALLED.
- A proposal for a mobile application pipeline for posture-based muscle rehabilitation at homePublication . Lopes, Júlio Castro; Van-Deste, Isaac; Lopes, Rui PedroThis work outlines the development process of a rehabilitation system from smartphones to provide remote monitoring and real-time feedback assistance in muscle rehabilitation exercises. The system leverages advanced machine learning techniques, such as pose estimation techniques and anomaly detection models, in tracking users’ movements and assessing the performance of exercises. The system is designed to identify deviations from correct movement patterns and provide corrective feedback, allowing patients to engage in home-based rehabilitation without specialized hardware. This paper details the system design, data acquisition methods, and the machine learning models used to detect incorrect movements. The outlined pipeline is intended to facilitate the creation of an affordable, cost-effective, and scalable rehabilitation platform for strengthening patient adherence to exercises and optimizing recovery outcomes
- Study of the effects of partial cement replacement with spent diatomaceous earth from the brewing industry in mortarPublication . Zolin, Renan; Angulski, Caroline; Ferreira, Débora; Meira, Ana;The present study evaluates the effect of partial replacement of cement in mortars with spent diatomaceous earth (SDE) from beer filtration, aiming to explore sustainable alternatives for civil construction. Three compositions were tested: one with 15% cement replacement by SDE, another with 5% sand replacement by SDE, and a reference composition without SDE addition. The diatomaceous earth (DE), after being used in beer filtration until saturation, was calcined in a furnace to eliminate organic matter and subsequently incorporated into the mortar composition. This research, as an extension of a previous study using residues from wine filtration, evaluates the impact of this replacement on the mechanical properties of the mortar, including compressive strength, tensile strength, and consistency. A detailed characterization of the spent diatomaceous earth was performed through X-ray diffraction and laser granulometry, enabling a better understanding of its structural properties.
- Diagnosis of broken bar fault in three-phase induction motors using fibre bragg grating strain sensors assisted by an algorithmPublication . Cavalcanti, Rafael; Dreyer, Uilian José; Aguiar, Everton Luiz de; Silva, Jean Carlos Cardozo da; Sousa, Kleiton Morais; Mendes, Andre C.This study developed an algorithm running on the cloud that makes data process and diagnoses broken rotor bar faults in three-phase induction motors (TIMs), by analyzing stator dynamic deformation using fibre Bragg gratings (FBGs) as sensors. This method can diagnose mechanical faults (misalignment, imbalance) and electrical faults (fractures or cracks in rotor rings or bars). FBG-based sensors were used due to their high multiplexing capability, electromagnetic radiation immunity, and long-distance operation. Tests were conducted on a small-scale induction motor (3 HP) coupled to a generator to simulate load and the generator supplied by the grid. Were used 1 healthy rotor and one rotor with a broken bar fault running at two load conditions, 75% and 100%. The algorithm successfully identified broken bar faults in two frequency regions: around the mechanical rotational frequency of the rotor 28.06 Hz and 31.014 Hz operating at 75%; 25.325 Hz and 32.995 Hz operating at 100% and approximately twice the electrical frequency supply, 118.455 Hz and 123.942 Hz operating at 75%; 117.237 Hz and 124.683 Hz operating at 100%. The system showed high sensitivity, a good signal-to-noise ratio, and advantages over conventional methods for detecting broken bar faults in induction motors.
- Analysis of the acoustical environment of classrooms in three Brazilian public schools through measurements and 3D simulationPublication . Andrade, Fernanda Horst; Ribeiro, Rodrigo Scoczynski; Braz-César, ManuelThe present study analyses the outdoor and indoor sound pressure levels (SPL) and the reverberation time (RT) measured in three Brazilian public classrooms. For the SPL, a sound level analyzer (class II) was used, and for the RT it was used a smartphone for the measurements. The sound sources were the impulses of bursting balloons and the data was processed in a MatLab toolbox (ITA-Toolbox). The classrooms were also simulated in an open source modeling software (I-SIMPA), using ray-tracing principles. Based on the results of the simulations, supported by the low-cost measurements, it was observed that the classroom didn't reach the national standards for classroom acoustics. Some improvements were designed with sustainable materials in order to reach the lower limits of the standards using the same room acoustics software. It was observed that the low-cost measurements helped on the diagnosis of classroom's acoustic issues which was also verified in the 3D simulation. This procedure showed itself as a cheap solution for classroom acoustic designs.
- Nonperiodic pathologic voice signals classification using mel-spectrogram and VGGishPublication . Fernandes, Joana; Pinto, João; Moura, Carla; Vilarinho, Helena; Teixeira, Felipe; Freitas, D.; Teixeira, João PauloIn this work and the literature, voice signals can be classified as peri-odic (type 1) or either some periodicity (type 2) and chaos (type 3). This work aims to classify signs into types 1, 2 or 3 to be subsequently applied in a classifi-cation system for pathological/control signs. The original dataset is composed of 466 type 1 individuals, 900 type 2 individuals, and 84 type 3 individuals classi-fied by an otolaryngologist. 15% of the data was used for testing and the remain-ing 85% was used for training and validation. A data augmentation technique was applied to balance the data in training set. Therefore, for the test set, 3380 sounds were used, 1020 type 1, 1280 type 2 and 1080 type 3. Of these, 80% were used for training and 20% for validation. The Mel spectrograms of the signals were used in the input of a VGGish to retrain the model in classifying the 3 types of signals. Regarding test accuracy, this network obtained 71.2%.