EOElder Oroski

Elder Oroski

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Elder Oroski has a total of 51 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Endoscopic Endonasal Stereoscopic Images of the Skull Base: From Creation to Automatic Evaluation

    by Milton Manrique Rastelli Júnior, João Pedro Murback Wiedmer, Pedro Marochi, Rubens Alexandre De Faria, Elder Oroski

    The study of neuroanatomy is an integral part of the education and training of neurosurgeons and aspiring professionals in this field, in which the use of images, videos, models, simulators, and virtual reality can fit into the pedagogy of this medical area. However, individuals who lack a solid understanding of three-dimensionality and proportion encounter challenges when interpreting two-dimensional images. Regardless of all the technology currently available, the stereoscopic images, a technique used by superimposing two images taken from the same region by moving a camera laterally, keeping the same horizontal plane and the same focal length, become valuable by presenting the viewer with the notion of depth and decreasing the learning time. However, when stereoscopic images are generated, many of them do not present the expected results, and could cause discomfort to the observer when analyzed, especially because their depth perception could be altered. In this study, a computational algorithm uses Pearson’s Correlation Coefficient, Angle between two vectors, and Oriented FAST and Rotated BRIEF (ORB) to analyze and compare the two stereo pair images. Thus, six parameters are described that best correlate the two images, classified in order using the Fisher Discriminant Ratio (FDR), and used to evaluate the quality of a set of 180 endoscopic endonasal stereo images of the skull base. This study evaluated stereo pairs to determine their suitability for accurate 3D pictures using digital image processing and classification methods. Six parameters tested were effective for class separation, with the most discriminative being the angle between column vectors in the red channel, determined by the highest FDR, and it can serve as a standalone classifier or be used in a multivariate discriminant function for automated quality assessment. This research also determines that existing quality metrics such as Unified No-Reference Quality Assessment via Monocular and the Binocular Local Visual Primitives and Top-Down JND Estimation were ineffective for this specific classification, highlighting the potential of the proposed methodology.

  • System Identification Study for a Digital Twin of Fuel Injection, Intake Valve, and Exhaust Valve in an HCCI Engine

    by Francisco Javier Prieto Coronel, Elder Oroski, José Antonio Andrés Velásquez Alegre

    This paper examines the application of nonlinear system identification techniques to the actuator system of a Homogeneous Charge Compression Ignition (HCCI) engine. The study focuses on three essential actuators: (i) the fuel injector; (ii) the intake valve; and (iii) the exhaust valve. In this context, a dynamic model is developed to accurately capture the empirical behavior of each actuator, with the goal of integrating them into a modular digital twin that represents the dynamics of the main HCCI engine subsystems. Furthermore, the study evaluates the performance of different nonlinear model structures to determine their effectiveness in capturing the inherent dynamics of these systems. Finally, the systems were described using Hammerstein-Wiener models, resulting in a Normalized Root Mean Square Error (NRMSE) of 86.04% for fuel injection, 88.00% for the intake valve, and 83.02% for the exhaust valve.

  • System Identification Study to Model a Digital Twin for a Heavy-Duty Vehicle's Oil System

    by Mayron Vinicius Lopes de Camargo, Elder Oroski, Marco Wehrmeister

    This paper explores the application of system identification to a lubrication system found in heavy-duty diesel engines. These engines are equipped with a variable oil pump and a cooling piston jet. The objective is to establish a dynamic model that accurately captures the relationship between the duty cycle of the valves and the resulting pressure values under normal road operating conditions to be used as a digital twin of the system. Additionally, the study aims to determine whether a simple recursive model can sufficiently describe the system with enough precision. Different linear and nonlinear models were evaluated and validated to identify the best fit for the system. Ultimately, the system was described using a Hammerstein-Wiener model, resulting in an 83.86% Normalized Root Mean Squared Error (NRMSE) for main gallery pressure and 82.69% for piston cooling jet gallery pressure.

  • Open Access

    Investigation of Features to Evaluate the Quality of Endoscopic Endonasal Stereoscopic Images of the Skull Base

    by Milton Manrique Rastelli Júnior, João Pedro Wiedmer, Vinicios Pereira Possagno, Lucas Cheslak Rogge, Célia Cristina Bojarczuk, Elder Oroski

    The study of neuroanatomy forms an integral part of the education and training for neurosurgeons and aspiring professionals in this field, in which the utilization of images and videos can fit in the pedagogy of this medical discipline. Furthermore, students who lack a solid understanding of three-dimensionality and proportion encounter challenges when interpreting two-dimensional images. Thus, stereoscopic images, a technique used by superimposing two images taken from the same region by moving the camera laterally, keeping the same horizontal plane and the same focal length, becomes valuable by presenting the viewer with the notion of depth and decreasing the learning time. However, when generating stereoscopic images, many of them do not present the expected result, could have their depth perception altered, and could cause discomfort to the observer when analyzed. In this study, a computational algorithm uses Pearson’s Correlation Coefficient, Angle between two vectors, and Oriented FAST and Rotated BRIEF (ORB) to analyze and compare the two stereo pair images. Thus, 6 parameters that best correlate the two images are described, classified in order using the Fisher Discrimant Ratio, and used to evaluate the quality of a set of 180 endoscopic endonasal stereo images of the skull base. The parameters are showcased in graphical results and discussed about its use to be employed as direct indicators for stereoscopic skull base images quality and expedite the analysis of stereoscopic brain images.


  • Open Access

    Nonlinear System Identification of a Homogeneous Charge Compression Ignition Engine

    by T. Mottin, E. Oroski, T. M. Carlos, L. B. Silveira, J. A. A Velásquez Alegre

    The primary aim of this study was to develop a comprehensive multi-variable model for Homogeneous Charge Compression Ignition (HCCI) engines and evaluate its efficacy by comparing it to simulated HCCI engine data. The complexity of the engine’s nonlinear dynamics necessitated the development of appropriate models to address the associated control challenges. Multiple Inputs Single Output (MISO) models were employed to capture the pressure and temperature variables. The estimated outputs were generated using Nonlinear Autoregressive with Exogenous Inputs (NARX) and Hammerstein-Wiener (HW) models, both were used as black box models. The performance of each model was assessed using metrics such as Normalized Root Mean Square Error (NRMSE), Mean Square Error (MSE), and Akaike’s Final Prediction Error (FPE). The results highlight the effectiveness of the Hammerstein-Wiener models in accurately representing the intricate dynamics of complex combustion, as observed in HCCI engines. The consistency of these models in delivering reliable outcomes further underscores their suitability for modeling such intricate systems.

  • Open Access

    Botnet Detection and Mitigation Model for IoT Networks Using Federated Learning

    by Francisco Lopes de Caldas Filho, Samuel Carlos Meneses Soares, Elder Oroski, Robson de Oliveira Albuquerque, Rafael Zerbini Alves da Mata, Fábio Lúcio Lopes de Mendonça, Rafael Timóteo de Sousa Júnior

    The Internet of Things (IoT) introduces significant security vulnerabilities, raising concerns about cyber-attacks. Attackers exploit these vulnerabilities to launch distributed denial-of-service (DDoS) attacks, compromising availability and causing financial damage to digital infrastructure. This study focuses on mitigating DDoS attacks in corporate local networks by developing a model that operates closer to the attack source. The model utilizes Host Intrusion Detection Systems (HIDS) to identify anomalous behaviors in IoT devices and employs network-based intrusion detection approaches through a Network Intrusion Detection System (NIDS) for comprehensive attack identification. Additionally, a Host Intrusion Detection and Prevention System (HIDPS) is implemented in a fog computing infrastructure for real-time and precise attack detection. The proposed model integrates NIDS with federated learning, allowing devices to locally analyze their data and contribute to the detection of anomalous traffic. The distributed architecture enhances security by preventing volumetric attack traffic from reaching internet service providers and destination servers. This research contributes to the advancement of cybersecurity in local network environments and strengthens the protection of IoT networks against malicious traffic. This work highlights the efficiency of using a federated training and detection procedure through deep learning to minimize the impact of a single point of failure (SPOF) and reduce the workload of each device, thus achieving accuracy of 89.753% during detection and increasing privacy issues in a decentralized IoT infrastructure with a near-real-time detection and mitigation system.


  • Estimation of Kautz Poles in Wiener-Volterra Models Using Levenberg-Marquardt Algorithm

    by Higor de Souza Serafin, Elder Oroski, André Eugênio Lazzaretti

    This work approaches the problem of estimating the Kautz optimal poles in kernel expansion in Wiener-Volterra models. The analytical solution for the suboptimal case is already established in the literature. However, the solution for the two parameters that compose the poles is still open. In this paper, an optimization strategy using the Levenberg-Marquardt is presented. This algorithm is used to find kernel expansion parameters, with the same base for all dimensions. The construction of bases using digital filter is considered. To validate the implemented algorithm, data collected from the excitation of an electrically coupled drive system was used to analyze the impact of the search space thresholds and the behavior of Levenberg-Marquardt’s parameters. It was also analyzed the impact on the model accuracy, as the number of functions in the base is increased. As a result, the models determined have achieved better results than the works found in the literature.


  • Feature Extraction of V–I Trajectory Using 2-D Fourier Series for Electrical Load Classification

    by Bruna Machado Mulinari, Lucas da Silva Nolasco, Elder Oroski, Andre Eugenio Lazzaretti, Robson Ribeiro Linhares, Douglas Paulo Bertrand Renaux

    Nonintrusive load monitoring (NILM) techniques allow the individual consumption of devices in an installation to be reported to the user, collaborating in the awareness and managing consumers’ energy use. One of the most critical steps in NILM procedures is feature extraction, which involves the quantitative description of the load signature. Most state-of-the-art methods, particularly those involving voltage–current ( V–I ) trajectories, are based on features that specialists design when manually analyzing signals—the so-called hand-crafted features. On the other hand, this work presents a mathematical method that describes the transient and steady-state (SS) load signature using a 2-D Fourier series, avoiding hand-crafted descriptors. The proposed approach evaluates load identification results with different classifiers and publicly available datasets. In addition to the evaluation of the proposed method’s computational resources, such as execution time and memory footprint in an embedded system, their robustness to noise insertion is also assessed. The results of the proposed approach indicate an average per-class accuracy higher than 90%, equivalent or superior to hand-crafted V–I trajectory-based methods in most cases and even deep learning approaches recently presented in the literature. Such results reinforce the main contributions of this work: 1) a mathematical method to extract interpretable features; 2) results comparable or superior to other state-of-the-art methods; and 3) promising results in computational resources in an embedded system.


  • Open Access

    A Dataset for Non-Intrusive Load Monitoring: Design and Implementation

    by Douglas Paulo Bertrand Renaux, Fabiana Pottker, Hellen Cristina Ancelmo, André Eugenio Lazzaretti, Carlos Raiumundo Erig Lima, Robson Ribeiro Linhares, Elder Oroski, Lucas da Silva Nolasco, Lucas Tokarski Lima, Bruna Machado Mulinari, et al.

    A NILM dataset is a valuable tool in the development of Non-Intrusive Load Monitoring techniques, as it provides a means of evaluation of novel techniques and algorithms, as well as for benchmarking. The figure of merit of a NILM dataset includes characteristics such as the sampling frequency of the voltage, current, or power, the availability of indications (ground-truth) of load events during recording, the variety and representativeness of the loads, and the variety of situations these loads are subject to. Considering such aspects, the proposed LIT-Dataset was designed, populated, evaluated, and made publicly available to support NILM development. Among the distinct features of the LIT-Dataset is the labeling of the load events at sample level resolution and with an accuracy and precision better than 5 ms. The availability of such precise timing information, which also includes the identification of the load and the sort of power event, is an essential requirement both for the evaluation of NILM algorithms and techniques, as well as for the training of NILM systems, particularly those based on Machine Learning.


  • Open Access

    A Multi-Agent NILM Architecture for Event Detection and Load Classification

    by André Eugenio Lazzaretti, Douglas Paulo Bertrand Renaux, Carlos Raimundo Erig Lima, Bruna Machado Mulinari, Hellen Cristina Ancelmo, Elder Oroski, Fabiana Pöttker, Robson Ribeiro Linhares, Lucas da Silva Nolasco, Lucas Tokarski Lima, et al.

    A multi-agent architecture for a Non-Intrusive Load Monitoring (NILM) solution is presented and evaluated. The underlying rationale for such an architecture is that each agent (load event detection, feature extraction, and classification) outperforms others of the same type in particular scenarios; hence, by combining the expertise of these agents, the system presents an improved performance. Known NILM algorithms, as well as new algorithms, proposed by the authors, were individually evaluated and compared. The proposed architecture considers a NILM system composed of Load Monitoring Modules (LMM) that report to a Center of Operations, required in larger facilities. For the purposed of evaluating and comparing performance, five load event detect agents, five feature extraction agents, and five classification agents were studied so that the best combinations of agents could be implemented in LMMs. To evaluate the proposed system, the COOLL and the LIT-Dataset were used. Performance improvements were detected in all scenarios, with power-ON and power-OFF detection improving up to 13%, while classification accuracy improved up to 9.4%.