Danton Diego Ferreira holds a degree in Industrial Electrical Engineering from UFSJ (2005), a Master’s in Electrical Engineering from UFJF (2007), a Ph.D. in Electrical Engineering from COPPE/UFRJ (2010) in the area of Computational Intelligence, and a Postdoctoral degree from UFJF (2014). He is currently an Associate Professor under a full-time dedication regime at UFLA, where he serves as Head of the Department of Automation. He is a CNPq Level 1D Productivity Researcher. He is the coordinator of the CNPq Research Group entitled Automation and Artificial Intelligence (AIA/UFLA) and a member of the groups Control and Modeling Group (GCOM/UFSJ), Information and Communication in Health (InfComSaude/UNICAMP), Information Processing (UFRJ), and Signal Processing Applied to Power Systems (PSCOPE/UFJF). He is the coordinator and founder of the Center for Innovation in Automation and Artificial Intelligence (AIA) at UFLA (http://www.aia.ufla.br/ ). He is also the coordinator of the Specific Technical Committee of the Brazilian Society of Automation (SBA) called System Identification and Data Science and a member of the Technical Council of the Brazilian Society of Computational Intelligence (SBIC). He has been a member of the Architecture and Engineering Chamber of FAPEMIG since March 2025. His research interests include signal processing, computational intelligence, neural networks, independent component analysis, electric power quality, Industry and Agriculture 4.0, and biomedical engineering.

Research Areas

Signal Processing and Computational Intelligence for Smart Grids

Objective: To study and develop computational intelligence and signal processing techniques for power quality analysis in Smart Grids.
Major Area: Engineering
Activity Sectors: Electricity, gas, and other utilities.

Biomedical Signal Processing

Pattern Recognition

---

Research Projects

Title	Description	Start Date	End Date
Development of Pattern Recognition Systems with Low Computational Complexity	This project proposes the development of pattern recognition systems with low computational complexity. The use of low computational complexity systems can lead to gains in computational efficiency, response time, cost reduction, accessibility and portability, lower energy consumption (reduced carbon footprint), and ease of implementation. To this end, pattern recognition models based on statistical signal processing techniques such as principal curves (PC), principal component analysis (PCA), and independent component analysis (ICA) will be proposed. Few-shot learning and transfer learning techniques will also be explored comparatively and in conjunction with PC, PCA, and ICA. The developed models will be applied in two distinct scenarios: (1) Monitoring of electrical signals in power systems; and (2) Detection of active tuberculosis via X-ray images. Both scenarios require solutions that enable real-time implementations with low computational complexity. Status: In progress; Nature: Research.	2024
Project of Artificial Intelligence Models with Low Computational Complexity and Reduced Carbon Footprint	This project proposes the development of AI-based systems with low computational complexity and a reduced carbon footprint. The use of low computational complexity systems can lead to gains in computational efficiency, response time, cost reduction, accessibility and portability, lower energy consumption, and ease of implementation. AI models will be proposed that will include the following objectives in a multi-objective optimization problem: 1) CO2 emissions measured at all stages of the model's design; 2) Suitability for validation data; 3) Average Degree of Explainability. The proposed method will be developed and tested considering mainly two application areas, without being limited to them: a) Electric Power Systems (EPS): monitoring of electrical signals; b) Health: detection and screening of diseases and complications from data and images. Status: In progress; Nature: Research.	2024
Monitoring Power Quality with Statistical Signal Processing and Computational Intelligence	This proposal is a continuation of the projects covered by FAPEMIG Calls 02/2015 and CNPq 28/2018, which aimed to analyze electrical disturbances in Power Systems using computational intelligence and yielded promising results, published in journals and conferences. This project addresses the problem of power quality (PQ) in the current scenario characterized by the strong penetration of solar and wind energy in the context of Smart Grids. In this new scenario of high penetration of distributed and dispersed generation, efforts in research and the development of new technologies will be necessary to solve the problems that are already starting to appear in the grids. In this sense, the general objective of this project is to develop and improve advanced techniques of statistical signal processing and computational intelligence for the decomposition and analysis of electrical signals in power systems. For decomposition, techniques based on statistical signal processing such as Independent Component Analysis (ICA) and Principal Component Analysis (PCA) will be explored. In this step, new online ICA and PCA algorithms will be introduced with the aim of real-time implementation. For the analysis of signals, Higher-Order Statistics (HOS) combined with classifiers/estimators based on Computational Intelligence and Pattern Recognition will be employed. In this step, the concept of novelty detection will be introduced with the aim of detecting unknown events. As a result, it is expected to develop new methodologies for: (1) extraction of harmonics, interharmonics, subharmonics, and supraharmonics; (2) classification and location of faults in power systems; and (3) detection of islanding. The methods will be tested on simulated and real signals via MatLab. It is expected to implement them in FPGA and LabView. Status: Concluded; Nature: Research.	2021	2024
Computer-Aided Diagnosis for Active Tuberculosis Exclusion in Contacts of Patients with Pulmonary Tuberculosis, Breaking the Chain of Transmission	This project aims to validate and develop applications to read chest X-rays of contacts and distinguish between (a) normal patterns; (b) parenchymal lesions not related to TB; and (c) active TB lesions. The application developed by the group will be made available for incorporation into the SUS (Brazil's public healthcare system), at no cost. This would allow for same-day exclusion of active TB and the possibility of prescribing TPR. For this, a multidisciplinary team will be available, which will include TB specialists, public health experts, and engineers specializing in artificial intelligence. For machine learning, a database of normal chest X-rays, from participants of two clinical trials for different regimens to treat LTBI, and publicly available databases of multiple X-rays with and without TB. Status: Concluded; Nature: Research.	2020	2024
Artificial Intelligence System for Decision-Making Support in Screening and Diagnosing COVID-19 Patients: Score and Risk Groups Based on Chest Imaging Exams	At this moment when the COVID-19 pandemic has overwhelming characteristics worldwide and is progressing in a worrying way in Brazil, telemedicine has become a daily option for many people. COVID-19 infection produces a very aggressive condition in the respiratory system in a part of the population that has a poor disease evolution. Although the systemic nature of the infection is now known, pulmonary clinical analysis remains one of the most important items in the screening and diagnosis of the disease and is a central element of telemedicine care. As an indicator, compared to 2019, hospitalizations for Severe Acute Respiratory Syndrome (SARS) grew by 606 in the first 18 epidemiological weeks of 2020. In a recent interview, researcher Margareth Dalcolmo from Fiocruz highlights that the pattern of pulmonary infection by COVID-19 is quite different from what was known until now. Thus, analysis by radiography, computed tomography, or ultrasound becomes quite challenging, especially considering other possible diseases to which the patient has been exposed throughout life and that concern the respiratory system (pneumonia, tuberculosis), leaving scars that must be considered in image analysis. The interpretation of the imaging exam requires the presence of a specialist, which delays the process in many primary care settings. To overcome this barrier, we are proposing the development of an open-source application that allows for a quick evaluation of the images. The analysis of imaging exams can be combined with clinical information, so that risk groups can be associated with patients, going beyond a predictive support for the disease. The proposed application will be developed using reproducibility techniques, which facilitates its incorporation into the SUS, and will be validated and made available free of charge to the SUS. Computational intelligence techniques will be used, more precisely deep learning. It should be noted that computational intelligence methods rely on the available data, so the quality of this data is of fundamental importance. Additionally, it is on quality data that the information to be processed can be obtained and, with that, the patterns of interest can be recognized, in this case, for the screening and diagnosis of COVID-19 infection. Therefore, techniques that evaluate data quality (DQ, in English) and, consequently, information quality, have been receiving increasing attention in the areas of data science and will be used here. Status: Concluded; Nature: Research.	2020	2023
Modeling the Occurrence of Coffee Pests and Diseases Using Knowledge Discovery Techniques	This project aims to develop mathematical models from meteorological data associated with the occurrence of coffee pests and diseases in the southern regions of Minas Gerais. The specific objectives are: (i) To develop software prototypes for monitoring the occurrence and incidence of pests and diseases in coffee plants; (ii) To monitor the occurrence of the main pests and diseases in coffee plants; (iii) To monitor the most relevant meteorological variables for the occurrence and incidence of pests and diseases in coffee plants; (iv) To establish progress curves for the main pests and diseases of coffee plants; (v) To use knowledge discovery techniques to identify meteorological patterns aimed at modeling the occurrence of coffee pests and diseases. Status: Concluded; Nature: Research.	2019	2024
Modeling Spectral Variables from Orbital and Suborbital Sensors to Determine the Hydric Conditions of Coffee Plants	This project aims to develop technologies to determine leaf reflectance patterns as indicators of the hydric conditions of coffee plants; To propose technology for monitoring the hydric conditions of coffee plants using precision agriculture. Status: Concluded; Nature: Research.	2019	2024
Statistical Signal Processing Applied to Smart Grids	This proposal's main objective is to develop methodologies for the application of statistical signal processing (SSP) to Smart Grids. Among the statistical signal processing tools, Principal Component Analysis (PCA), Independent Component Analysis (ICA), and Higher-order Statistics (HOS) will be explored in this proposal. These tools have in common the good ability to: (i) deal with high-dimensional data; (ii) extract relevant (often hidden) information from electrical signals; (iii) decompose signals into isolated components; (iv) estimate parameters; and (v) design adaptive filters. The methodologies to be developed will focus on application in the much-discussed smart grids, in a context of distributed generation (DG). Smart grids must have a set of basic functions that allow for the modernization of the electrical infrastructure, among which stand out: (1) self-reconfiguration capability; (2) being fault-tolerant, resisting hacker attacks; (3) allowing the integration of all options of energy and storage sources; (4) allowing the dynamic optimization of grid operation; (5) allowing the active participation of consumers; and (6) improving the reliability, power quality, security, and efficiency of the energy system. Some of these functions are not evidently new, as the energy infrastructure has always relied on intelligent technologies for its operation, control, and protection, etc., but in this new scenario of high penetration of distributed and dispersed generation, it will be necessary to make efforts in research and the development of new technologies to solve the problems that are already beginning to appear in the grids. In this context, the main contributions of the present project are: (i) the development of a classification system for the causes of voltage sags; (ii) the development of an islanding detection system; (iii) the development of a method for the extraction of harmonic, sub-harmonic, and interharmonic components; and (iv) the proposal of a methodology for signal compression. Status: Concluded; Nature: Research.	2018	2022
Development of Computational Intelligence and Signal Processing Techniques for Power Quality Analysis and Identification of Electrical Loads in Smart Grids	This proposal is a continuation of the project covered by FAPEMIG Call 01/2011, which aimed to analyze electrical disturbances in Power Systems using computational intelligence and presented promising results (one scientific article in an international Qualis A1 journal and three full articles in national conferences). Still in the line of electrical signal analysis with computational intelligence and signal processing, this proposal now has a focus on the much-discussed smart grids. Smart grids must have a set of basic functions that allow for the modernization of the electrical infrastructure, among which stand out: (1) self-reconfiguration capability; (2) being fault-tolerant, resisting hacker attacks; (3) allowing the integration of all options of energy and storage sources; (4) allowing the dynamic optimization of grid operation; (5) allowing the active participation of consumers; and (6) improving the reliability, power quality, security, and efficiency of the energy system. Some of these functions are not evidently new, as the energy infrastructure has always relied on intelligent technologies for its operation, control, and protection, etc., but in this new scenario of high penetration of distributed and dispersed generation, it will be necessary to make efforts in research and the development of new technologies to solve the problems that are already beginning to appear in the grids. In this sense, the general objective of this project is to develop an electrical energy monitoring system capable of identifying the electrical loads activated in a given electrical system and providing information on the power quality of the grid that powers such loads, as well as the changes in it due to the activation of each electrical load, considering the Smart Grids scenario. Status: Concluded; Nature: Research.	2015	2017
Instrumentation, Signal Processing, and Computational Intelligence for Pattern Recognition	This project proposes the development of instrumentation tools with signal processing and computational intelligence techniques for applications in pattern recognition and systems that require careful data (or signal) analysis. For this purpose, the project addresses three subprojects that are being developed by the coordinator and are directly related to the main theme of this proposal. These subprojects have in common the development and improvement of instrumentation, signal processing, and computational intelligence techniques for the following purposes: 1. Power Quality Analysis and Identification of Electrical Loads in Smart Grids; 2. Analysis of biomedical signals: electroencephalogram (EEG), electrocardiogram (ECG), and electromyogram (EMG); 3. Diabetic foot prevention: a multidisciplinary approach. Status: In progress; Nature: Research.	2015
Development of Computational Intelligence and Signal Processing Techniques for Smart Grids	This proposal is a continuation of the project covered by FAPEMIG Call 01/2011, which aimed to analyze electrical disturbances in Power Systems using computational intelligence and presented promising results (one scientific article in an international Qualis A1 journal and three full articles in national conferences). Still in the line of electrical signal analysis with computational intelligence and signal processing, this proposal now has a focus on the much-discussed smart grids. Smart grids must have a set of basic functions that allow for the modernization of the electrical infrastructure, among which stand out: (1) self-reconfiguration capability; (2) being fault-tolerant, resisting hacker attacks; (3) allowing the integration of all options of energy and storage sources; (4) allowing the dynamic optimization of grid operation; (5) allowing the active participation of consumers; and (6) improving the reliability, power quality, security, and efficiency of the energy system. Some of these functions are not evidently new, as the energy infrastructure has always relied on intelligent technologies for its operation, control, and protection, etc., but in this new scenario of high penetration of distributed and dispersed generation, it will be necessary to make efforts in research and the development of new technologies to solve the problems that are already beginning to appear in the grids. In this sense, the general objective of this project is to develop an electrical energy monitoring system capable of identifying the electrical loads activated in a given electrical system and providing information on the power quality of the grid that powers such loads, as well as the changes in it due to the activation of each electrical load, considering the Smart Grids scenario. Status: Concluded; Nature: Research.	2013	2016
Development of Computational Intelligence and Signal Processing Techniques for Power Quality Analysis and Identification of Electrical Loads in Smart Grids	This proposal is a continuation of the project covered by FAPEMIG Call 01/2011, which aimed to analyze electrical disturbances in Power Systems using computational intelligence and presented promising results (one scientific article in an international Qualis A1 journal and three full articles in national conferences). Still in the line of electrical signal analysis with computational intelligence and signal processing, this proposal now has a focus on the much-discussed smart grids. Smart grids must have a set of basic functions that allow for the modernization of the electrical infrastructure, among which stand out: (1) self-reconfiguration capability; (2) being fault-tolerant, resisting hacker attacks; (3) allowing the integration of all options of energy and storage sources; (4) allowing the dynamic optimization of grid operation; (5) allowing the active participation of consumers; and (6) improving the reliability, power quality, security, and efficiency of the energy system. Some of these functions are not evidently new, as the energy infrastructure has always relied on intelligent technologies for its operation, control, and protection, etc., but in this new scenario of high penetration of distributed and dispersed generation, it will be necessary to make efforts in research and the development of new technologies to solve the problems that are already beginning to appear in the grids. In this sense, the general objective of this project is to develop an electrical energy monitoring system capable of identifying the electrical loads activated in a given electrical system and providing information on the power quality of the grid that powers such loads, as well as the changes in it due to the activation of each electrical load, considering the Smart Grids scenario. Status: Concluded; Nature: Research.	2013	2015
Analysis of Electrical Disturbances in Power Systems Using Computational Intelligence	The general objectives of this project are the development and study of methodologies for the analysis of electrical disturbances in power systems using advanced computational intelligence techniques considering the occurrence of both isolated and multiple disturbances. The development of these methodologies will aim to propose systems with reduced computational complexity to enable their real-time application. Status: Concluded; Nature: Research	2012	2014
Computational Intelligence Applied to Biomedical Signals	This work aims to process biomedical signals, such as electroencephalogram (EEG), electrocardiogram (ECG), and electromyographic (EMG) signals, using advanced computational intelligence techniques to make clinical analysis from these signals more objective. Status: Concluded; Nature: Research.	2011	2024
Identification of Nonlinear Dynamic Systems: Applications Using Natural Computing and Hybrid Systems	This project aims to study and develop advanced gray-box or black-box identification techniques for nonlinear dynamic systems, comparing the different entities of prediction error and simulation error, as well as proposing other cost functions, especially in the training of artificial neural networks, whether or not using the multi-model approach and using natural computing techniques. Status: Concluded; Nature: Research.	2011	2013
Acquisition and Analysis of Electrical Signals for Computational Intelligence Applications	This project aims to acquire and analyze electrical signals using advanced computational intelligence and signal processing techniques for teaching and research purposes. Status: In progress; Nature: Research.	2011	2012
Computational Intelligence Applied in the Identification of Nonlinear Dynamic Systems	The objective of this project is to analyze the use of different cost functions (such as prediction error and simulation error) and the use of different model structures (such as artificial neural networks and rational and polynomial models) for the identification of nonlinear dynamic systems. Furthermore, this project aims to compare and propose methods for the identification of hybrid systems and the implementation of new approaches for the insertion of a priori information in the training of artificial neural networks. Evolutionary algorithms will be implemented for the estimation of model parameters in both mono-objective and multi-objective problems. Status: In progress; Nature: Research.	2010	2012
Signal Processing and Computational Intelligence Applied to Power Quality Monitoring	The present project aims to develop power quality (PQ) monitoring systems using sophisticated signal processing and computational intelligence techniques. In the context of PQ monitoring, emphasis will be given to the detection and classification of disturbances and the identification and location of disturbance-generating sources. Status: Concluded; Nature: Research.	2010	2012

---

Year	Publication
2025	KACZMARCZYK, GRZEGORZ ; STANISLAWSKI, RADOSLAW ; KAMINSKI, MARCIN ; KASPRZYK, KACPER ; FERREIRA, DANTON DIEGO . A neural-enhanced active disturbance load-side speed control of an electric drive with a flexible link. Archives Of Electrical Engineering, v. 74, p. 269-269, 2025.

Extension Projects

                                                                                                                                                     

Title	Description	Start Date	End Date
Artificial Intelligence for Process Automation	The project aims to disseminate technical information about artificial intelligence (AI) for general process automation to the community. The project will be developed at the Automation and Artificial Intelligence Laboratory (AIA) of the Automation Department (DAT) by the CNPq research group called Artificial Intelligence and Automation. With the advent of the AI field in the last 5 years, mainly driven by the limitations imposed by the pandemic, the term "artificial intelligence" has become quite popular. Consequently, the demand for AI-based products has increased. However, technical knowledge about the tool is still very restricted to scientists and engineers. The project's objective is, therefore, to show the community the possibilities and limitations of the tool, in addition to publicizing the work of the UFLA research group. Status: In progress; Nature: Extension.	2022
Study on Energy Efficiency and Automation in Vehicles	The project aims to promote the importance of "clean" energy and, in parallel, attract students' attention to the need to train engineers for the development of Brazil, in addition to imparting basic knowledge of mechanics and electricity using a prototype produced by the group. Initially, a prototype of an electric car will be built to participate in the University Energy Efficiency Marathon Competition. This will be used by the group to transmit knowledge of mechanics and electricity to high school students, contextualizing the topics of mathematics and physics learned in school. Status: Concluded; Nature: Extension.	2013	2018
Development of an Automatic System for the Identification of Diabetic Patients with the Potential to Develop Diabetic Foot	Diabetes is a serious health problem, referring both to the number of people affected, generating disability and mortality, and to the high government investment for the control and treatment of its complications, with emphasis on infections affecting the feet, known as diabetic foot. These infections are the main cause of amputation. Knowing that the main measure in the treatment of diabetic foot is early detection, this project aims to develop an automatic system using neural networks for the identification of patients who have the potential to develop diabetic foot. The system will process patient information, collected from questionnaires, and present as output their classification into three risk groups for developing diabetic foot: high risk, medium risk, and low risk. The system will allow for more dynamic control of the problem and will direct more specific campaigns to each group, thus preventing the development of diabetic foot. Status: In progress; Nature: Extension.	2012
Computerization of Procedures for High-Impact Infectious Diseases in Rio de Janeiro	The project aims to develop an online platform to support the medical diagnosis of Tuberculosis. Status: Concluded; Nature: Extension.	2008	2009