Search

Optimal power flows play a key role in power system operation planning. While most papers in the literature focus on attaining optima, sequencing paths of optimal control adjustments that lead the system from an initial operating point towards the optimum remain scarcely accounted for. Thus, this work proposes a practical framework based upon power system steady-state analysis for sequencing strictly feasible paths of optimal control adjustments determined by the Optimal Reactive Dispatch (ORD) via Lagrange multiplier sensitivity analysis. The proposed framework is methodologically founded on the reformulation of the ORD in terms of optimal control adjustments rather than optimal control values, successive Newton’s power flow calculations to assure a strictly feasible path from the initial operating point towards the optimum, and successive resolutions of the reformulated ORD’s associated dual problem to determine Lagrange multipliers along such sequence path. Thus, pondering optimal control adjustments by their respective Lagrange multipliers indicates which control action must be realised. Numerical results for IEEE test-systems with up to 300 buses with an increased number of controllable variables are obtained to validate and illustrate the efficiency and robustness of the proposed framework.

The appeal for more reliable, efficient and resilient power grids has become indispensable with the recent technological development and ever-increasing demand for electrical energy. These goals might be achieved by introducing sensors and actuators in the power grid, with the purpose of enabling a smart control of the power network. All power devices must be connected using a reliable communication network, which should be able to operate even when the power grid fails. Currently, several communication technologies have been applied for supporting this new application scenario. This article aims for reviewing both academic and market use cases for technologies applied to the mission-critical, control networks and Smart Grid applications, such as teleprotection, self-healing, communication with control centers and field devices, among others. More specifically, the key aspects, potential technologies, main protocols and use cases of operating data networks in energy transmission, power distribution and smart grid environments, including circuit and packet switching technologies, are discussed into details. Furthermore, resilience and the main telecommunications technologies used in power grids, as well, the correlation among them, are exploited in practical point of view. The article provides a broad discussion on the best telecommunication options to build the emerging intelligent energy distribution systems, covering control networks, teleprotection, and smart grid applications.

Seeking to solve problems in the power electric system (PES) related to exacerbated and uncontrolled energy consumption by final consumers such as residences, condominiums, public buildings and industries, electric power companies (EPC) are increasingly seeking new information and communication technologies (ICTs) to transform traditional electric power distribution networks into smart grids (SG). With this implementation, PES will be able to remotely control electric power consumption as well as monitor data generated by smart meters (SM). However, Internet-of-Things (IoT) technologies will enable all this to happen quickly and at low cost, since they are low-cost devices that can be deployed quickly and at scale in these scenarios. With this in mind, this work aimed to study, propose, and implement a hybrid communication infrastructure with LoRaWAN and LoraMesh for the demand-side management as a service (HyDSMaaS) using IoT devices such as long range (LoRa) to provide an advanced metering infrastructure (AMI) capable of performing all these applications as a service offered by EPC to end consumers. Additionally, services such as demand-side management (DSMaaS) can be used in this infrastructure. From the preliminary results it was found that the LoRaWAN network achieved a range of up to 2.35 km distance and the LoRaMESH one of 600 m, thus, the latter is more suitable for scenarios where there is little interference and the SMs are at long distances, while the other is used for scenarios with greater agglomeration of nearby SMs. Considering the hybridized scenario between LoraWAN and LoRaMESH, it can be seen that the implementation possibilities increase, since its range was approximately 3 km considering only one hop, and it can reach 1023 devices present in a mesh network. Thus, it was possible to propose the actual implementation of LoRaWAN and LoRaMESH protocols as well as the hybridization of the two protocols for HyDSMaaS. Additionally, the results obtained are exclusively from Radioenge’s LoRa technology, which can be further improved in the case of using more powerful equipment.

Currently, people are very concerned with the quality of life in their cities, especially with the level of air quality. Due to this concern, this work aims to analyze the communication of Internet of Things (IoT) devices using Long-Range technology (LoRa) with the Long-Range Wide Area Network (LoRaWAN) protocol for monitoring air quality and smart cities. In this context, factors such as distance between nodes, types of antennas and different Spreading Factor (SF) are considered. In addition, metrics Received Signal Strength Indicator (RSSI), Latency, Signal-to-Noise Ratio (SNR) and Packet Loss Rate (PDR) are collected from a real scenario in the city of Teresina in Brazil. Furthermore, a simulation to analyze the lifetime of the sensor nodes using energy consumption is run. Results show that the LoRaWAN technology proved to be a viable option, with low cost, low energy consumption, covering a distance of up to 2000 meters with packet loss rate of 0.3%, low SNR and latency of 23ms, with shipment of up to 12000 packages with node life of 2081 days, reaching a total consumption of 0.0105 mWh, even with all the barriers of the real scenario.

Given the recent progresses in designing and implementing electricity market optimization-based tools to support power allocations that bridge the gap between supply and demand, the realistic modeling of optimal power flow-based auction systems provide transparent and higher transaction levels among market players, as well as more accurate locational marginal prices. Nevertheless, actual voltage stability/security representation in these auction systems remains an open question. Thus, this paper proposes a two-stage resolution framework for a nonlinear Voltage-Stability/Security-Constrained Optimal Power Flow (VSCOPF)-based auction system. The first stage of the proposed framework consists in solving a single-period two-sided VSCOPF-based auction system model by a Particle Swarm Optimization (PSO) algorithm, which adjusts active power allocations and generator voltage magnitudes while verifying if the grid's technical-operational constraints are satisfied by power flow calculations. The system voltage stability/security is addressed in the second stage by a maximal loading problem solved by the Continuation Power Flow method. Then, if a minimal voltage stability/security margin condition is not satisfied, active power allocations and generator voltage magnitudes are readjusted by the PSO to comply with stable/secure power allocations. Results with a 6-bus test-system are presented to illustrate this proposal and, therefore, show the effectiveness of proposing a two-stage resolution framework for VSCOPF-based auction systems.

Electric power distribution systems are often configured to operate radially and without islands. Radiance is important for the sake of improving system protection. Also, for legal reasons, distribution systems should not operate with islands. In this work, we present a procedure for generating radial configurations. In addition, the methodology proposed in this paper is able, through switches, to detect and remove configurations with islands. Tests were carried out in several systems and, after analysis, we can verify the efficiency and robustness of the proposed methodology even for large systems.

The region of Ibiapaba mountains, in Ceara, stands out as a region with excellent tourist potential in the state. Due to this, the present paper carried out comparative tests between four algorithms to solve the Traveling Salesman Problem applied to tourist places in the region. The general objective is to verify which of the algorithms obtained the best performance, measured through the execution time, to find an optimal route between a set of tourist sites. The studied algorithms were Brute Force, Ant Colony Optimization, Genetic Algorithm, and 2-OPT algorithm. We modeled the Traveling Salesman Problem using graphs and distances and captured real locations using Google Maps. The results showed the faster 2-OPT algorithm during tests, although any of the tested algorithms can perform this function, depending on the number of points of interest. This work is a basis for future optimized tourist itinerary planner applications.

As a consequence of a short circuit in the power electrical system, there is a power quality disturbance occurring in the voltage/current signal that is propagated along the bars that compose the system with different characteristics and it is influenced by several factors, which makes it difficult to develop methods that can be used to locate the fault. Considering the advanced measurement infrastructure originating from the smart grid, these signals can be measured in real time and analyzed in a control and operation center improving the reliability of the system. This paper performs an exploratory data analysis on a database of voltage signals during the occurrence of a short circuit applied to the IEEE 34Bus system, aimed at the discovery of heuristics that can assist in decision making. The database includes information from voltage signals measured at each bar of the system and was assembled from simulations of three-phase, two-phase, and single-phase short circuits. Statistical metrics and data visualization are applied on the data subset to infer how the disturbance propagation occurs according to characteristics that are extracted from the measured signal during the presence of a fault in the power distribution system. This work shows that the combination of graphical, statistical modeling techniques is a powerful approach to explore these events. Multivariate data analysis has benefits over classical approaches that are based on raw data.

Among the numerous alternatives used in the world of risk balance, it highlights the provision of guarantees in the formalization of credit agreements. The objective of this paper is to compare the achievement of fuzzy sets with that of artificial neural network-based decision trees on credit scoring to predict the recovered value using a sample of 1890 borrowers. Comparing with fuzzy logic, the decision analytic approach can more easily present the outcomes of the analysis. On the other hand, fuzzy logic makes some implicit assumptions that may make it even harder for credit-grantors to follow the logical decision-making process. This paper leads an initial study of collateral as a variable in the calculation of the credit scoring. The study concludes that the two models make modelling of uncertainty in the credit scoring process possible. Although more difficult to implement, fuzzy logic is more accurate for modelling the uncertainty. However, the decision tree model is more favourable to the presentation of the problem.

Intelligent Energy Networks are comprised of devices capable of fulfilling their functions in an energy-efficient fashion and with communication and remote control capabilities. Therefore, some of these devices, such as smart energy meters, become attractive for use in the power generation and distribution industry, achieving the vision of Smart Grids. However, many are the challenges that need to be overcome in order to reach a fully-functional and security-aware smart grid. Providing measurement, control, communication, power, display, and synchronization capabilities shall be no easy task for smart meters. In this context, this paper elaborates on a detailed description of the main functionalities that smart meters must provide, along with the analysis of existing solutions that make use of smart meters for smart grids. Moreover, open challenges in the topic are identified and discussed. By the end of this research piece, the reader should be able to have a detailed view of the capabilities already offered by smart meters and the ones they will have available in order to tackle the challenges smart grids present.

Segmentation is a crucial step to obtain success for classifying medical images. However, it is a highly complex task due to the abnormal shapes and the presence of other artifacts. In this study, a melanoma segmentation approach based on deep learning is proposed. In conjunction with post-processing techniques, the proposed modified U-net network has proven to be highly effective in lesions segmentation. The experiments were performed in two public datasets (PH2 and DermIS) and reached an average Dice coefficient of 0.933 in the PH2 dataset and Dice = 0.872 in the DermIS dataset. Considering the high-performance methodologies available in the literature, the proposed solution is very promising, surpassing other methods with very promising results.

Lung cancer is one of the leading death causes by cancer worldwide. Early diagnosis increases the patient's cure chances. This diagnosis is made by computed tomography, an imaging exam that provides accurate information about the nodule. However, it depends on many external factors, from equipment quality to the fatigue of expert who analyzes. Image processing techniques might be great allies in early nodule detection, once it has no human limitations. This study presents an evaluation of two deep learning approaches, 3D U-Net and 3D V-Net, with different configurations of architectures, parameters, and data augmentation distribution applied to pulmonary nodules segmentation. The best results obtained mean an IoU of 0.74 and 0.99 for 3D U-Net and 3D V-Net, respectively. The second network obtained the best results because it is a much more robust network than the 3D U-Net, since it is a network developed for volumetric data processing.

Lung cancer is the leading cause of cancer-related death worldwide. Early diagnosis of pulmonary nodules on chest CT scans provides a chance to design an effective treatment. The focus of this study is the classification problem of benign and malignant pulmonary nodules in CT images. Thus, it is proposed to apply texture maps directly to the 3D nodules as a previous of the feature extraction process. For this, the local binary patterns (LBP), with branches, such as using neighbors with borders (LBP-6), average dimensions (LBP-M), and a 3×3×3 neighborhood (LBP-3×), to highlight the nodule texture. Convolutional Neural Networks, such as DenseNet, ResNet, and LeNet, were used as attribute extractors using the 3D texture maps computed. Then, those deep features are used as input to train a Random Forest classifier. In the experiments, it is used LIDC-IDRI image database. The LIDC-IDRI database was used with two segmentation process, one made by radiologists, present in the base itself (B1), and one performed automatically by a third party (B2). In B1, the best result was the original nodules' attributes extracted with the DenseNet architecture reaching an accuracy of 0.8371, a specificity of 0.9130, sensitivity of 0.7328, and Kappa of 0.6591. In B2, the best result was a combination of attributes of the original nodule combined with the extracted LBP-6 with LeNet architecture that reached an accuracy of 0.9037, a specificity of 0.8453, sensitivity 0.9266, and Kappa of 0.7641. In conclusion, it is possible to improve the classification accuracy by including a texture map computation as part of the process.

COVID-19 leads to radiological evidence of lower respiratory tract lesions, which support analysis to screen this disease using chest X-ray. In this scenario, deep learning techniques are applied to detect COVID-19 pneumonia in X-ray images, aiding a fast and precise diagnosis. Here, we investigate seven deep learning architectures associated with data augmentation and transfer learning techniques to detect different pneumonia types. We also propose an image resizing method with the maximum window function that preserves anatomical structures of the chest. The results are promising, reaching an accuracy of 99.8% considering COVID-19, normal, and viral and bacterial pneumonia classes. The differentiation between viral pneumonia and COVID-19 achieved an accuracy of 99.8%, and 99.9% of accuracy between COVID-19 and bacterial pneumonia. We also evaluated the impact of the proposed image resizing method on classification performance comparing with the bilinear interpolation; this pre-processing increased the classification rate regardless of the deep learning architectures used. We c ompared our results with ten related works in the state-of-the-art using eight sets of experiments, which showed that the proposed method outperformed them in most cases. Therefore, we demonstrate that deep learning models trained with pre-processed X-ray images could precisely assist the specialist in COVID-19 detection.

The modernization of Power Systems (PSs) to smart grids, the expansion of microgrids, the ever-increasing presence of distributed power generation, the more frequent use of non-linear and voltage-sensitive loads by the consumers have caused problems to the Power Quality (PQ). The studies in PQ are commonly related to disturbances that alter the sinusoidal characteristics of the voltage waveforms and/or current. The first step to analyzing the PQ is to detect and then classify the disturbances, since by identifying the disturbance, it is possible to know its causes and deliberate over strategies to mitigate it. Thus, this paper proposes a deep-learning approach using voltage signals, without pre-processing, extraction, nor manual selection of features in order to detect and classify PQ disturbances automatically. The proposed approach is composed of convolution layers, a pooling layer, a long short-term memory layer, and batch normalization. A 1D convolution was used to adapt the data from the voltage signals. Overlapping windowed signals with different Signal–Noise Ratio (SNR) (40 dB, 30 dB, 20 dB and 0 dB) and with different sampling rates (16, 32, and 64 samples/cycle) were used. For a more in-depth view of the results, the proposed approach was evaluated for its accuracy, precision, recall, and F1-Score in different scenarios. An analysis of the obtained results shows that even for the worst case scenario (SNR of 20 dB and sampling rates of 16 samples/cycle), the approach performs satisfactorily with values above 0.97 for the analyzed metrics, allowing, thus, consumer action in a demand-side management scenario.

Demand-side management solutions reward flexible customers for achieving desired goals. However, under price-based demand-response programs, uncoordinated load shifting among many customers may lead to rebound peaks, thus incurring financial costs on the energy service provider (ESP). This study addresses this issue from both the ESP (utility company) and customer’s perspectives in a two-step approach. First, each customer solves a local multi-objective load scheduling problem. Thereafter, the ESP solves a system-wide demand profile optimization problem. More precisely, we formulate in the first step a multi-objective optimization model to minimize consumption costs and load schedule discomfort while considering multiple energy sources and customer preferences. In the second step, we design an approach that combines Pareto-optimal solutions from all customers and minimizes their aggregate demand profile’s peak-to-average ratio (PAR). Experimental results show that the proposed approach outperforms the equivalent single-objective optimization models, and it can reduce the PAR metric by up to 11%. In other words, the proposed approach successfully improves the ESP’s system-wide demand profile aggregation while reducing customers’ expenses and discomfort.

The evolution of e-commerce has contributed to the increase of the information available, making the task of analyzing the reviews manually almost impossible. Due to the amount of information, the creation of automatic methods of knowledge extraction and data mining has become necessary. Currently, to facilitate the analysis of reviews, some websites use filters such as votes by the utility or by stars. However, the use of these filters is not a good practice because they may exclude reviews that have recently been submitted to the voting process. One possible solution is to filter the reviews based on their textual descriptions, author information, and other measures. This chapter has a propose of approaches to estimate the importance of reviews about products and services using fuzzy systems and artificial neural networks. The results were encouraging, obtaining better results when detecting the most important reviews, achieving approximately 82% when f-measure is analyzed.

The Radial Basis Function Network centres determination is a open problem. In this work, the cluster centres are determined by a proposed gradient algorithm using the information forces acting on each data point. These centres are applied to a Radial Basis Function Network for data classification. A threshold is established based on Information Potential to classify the outliers. Combined, the threshold and the centres determined by information forces show good results in comparison to a similar Network with a k-means clustering algorithm.

The supply of electricity is an increasingly essential service in modern society and, like this one, has passed through several changes during its evolution. The increased complexity of the electrical energy systems from the generation to the distribution has brought with it several adapables, practical and legal impositions, restrictions, as well as more attention to certain areas of research. Since the continuity of the electricity supply one of its main characteristics, many researches turn their attention to problems involving the breaking this continuity, being the faults one of the well-known reasons. When a fault is detected, the protection systems must act quickly so that the system stability is preserved. Given the importance of the fault event, several studies are focused on its detection, classification and location. This work compared the faults classification methods most used in modern Electricity Distribution Systems like Neural Networks, Trees, Logistic Regression and others. The accuracy and robustness of the algorithms were performed using the data collected in simulations of different types of fault such as line to ground (LG), double Line (LL), double line to ground (LLG) and three phase fault involving ground (LLLG). An analysis of possible interference in the acuracies of algorithms was also carried out in the face of variations in the intensity of the fault impedance, presence of distributed generation, noisy data and reduced dimensionality. All fault scenarios have been modeled using the OpenDSS software and IEEE-13 bus scenario was used. After the simulations, it was noticed that the Neural Networks showed a better ability to generalize the problem even in scenarios with noisy data. Dimensionality reduction was more effective only in the noiseless scenarios, decreasing the accuracy of the methods when the data were noisy.

New and more transmissible SARS-COV-2 variants aggravated the SARS-COV-2 emergence. Lung X-ray images stand out as an alternative to support case screening. The latest computer-aided diagnosis systems have been using Deep Learning (DL) to detect pulmonary diseases. In this context, our work investigates different types of pneumonia detection, including COVID-19, based on X-ray image processing and DL techniques. Our methodology comprehends a pre-processing step including data-augmentation, contrast enhancement, and resizing method to overcome the challenge of heterogeneous and few samples of public datasets. Additionally, we propose a new Genetic Fine-Tuning method to automatically define an optimal set of hyper-parameters of ResNet50 and VGG16 architectures. Our results are encouraging; we achieve an accuracy of 97% considering three classes: COVID-19, other pneumonia, and healthy. Thus, our methodology could assist in classifying COVID-19 pneumonia, which could reduce costs by making the process faster and more efficient.

Routing, Modulation Level and Spectrum Assignment (RMLSA) are some of the main problems studied in elastic optical networks. This work focuses on the study of guard band selection, with one or more free slots between the circuits, which are used in the solutions of the RMLSA problem in order to reduce the interference between adjacent circuits in the spectrum optical. In this context, a new approach, called ADVANCE, which uses a convolutional neural network to adaptive guard band selection is proposed. A proposal performance is compared to other adaptive proposals: AGBA, GBUN and UTOPIAN. The proposal achieves a reduction in the bandwidth blocking probability of at least 86.56% relative to AGBA, 84.60% relative to GBUN and 73.26% relative to UTOPIAN.

This paper presents an efficient methodology based on Graph Theory for imposing the radial topology constraint in unfeasible solutions that appear in Evolutionary Algorithms during the evolution of the optimal solution in Distribution Network Reconfiguration. To reach a better computational time, this paper also presents an effective Backward-Forward method to solve power flow in radial DS (RDS). The algorithm developed was tested in distribution systems With the well-known switch reduction problem to achieve a lower number of active power losses.

The increased demand for electricity has caused several problems for traditional electrical power systems, such as voltage fluctuations and interruptions in supply. These events, power quality disturbances, cause several losses for both the concessionaire and its consumers, either by damaging appliances or interrupting their operation. Among these power quality disturbances, the voltage sag stands out for being the most frequent event, causing several losses. Therefore, it is extremely important to locate the source of these disturbances in the electrical distribution system, in order to mitigate the problem. In general, methods for locating disturbances use few electrical meters and an analysis of the characteristics of voltage and current signals, which results in the estimation of a large region as a result. This paper proposes a approach to find not a region, but the bus in the power distribution system in which the voltage sag disorder originated by using a model of deep learning.

A transmission power system carries the energy produced by generators to load points over long distances and to ensure the power demand is satisfactorily reached, the electric utilities use devices running applications for monitoring, control and protection, in order to maintain system operation in a normal state. These applications mainly require continuous data communication and low latency, which makes crucial to define the appropriate communication technology. In this context, this paper proposes a study which simulates a scenario based on the IEEE 14-bus transmission test system, on NS-3 (Network Simulator 3) tool, using the wireless Long Term Evolution (LTE) and WiMAX technologies to move data across data links on the system. The simulations generate metrics like delay, throughput and packet loss, to compare the technologies and express the degree of compatibility of each one with the applications tested, which can guide the utilities in choosing the communication technology to be deployed in its system.

This paper presents a structural recognition system using machine learning algorithms (Multilayer Perceptron, Support Vector Machine and Random Forest) and the environment information to analyzes the feasibility of the use of machine learning methods for the construction of topological maps. The proposed method combines the recognized information from a given scene with a topological graph to create a map. This map can be used to plan high-level tasks of robotic navigation. The topological nodes are used to store semantic information, such as the robot’s poses, sensor data and scene characteristics. The machine learning algorithms classification of the structural information as either rooms, corridors or doors obtained a satisfactory performance. The structural recognition provided by classification presents accuracy greater than 97% and topological maps built efficiently of classification.

In this work, a preference-based, demand response (DR) multi-objective optimization model based on real-time electricity price is presented to solve the problem of optimal residential load management. The purpose of such a model is threefold: 1) to minimize the costs associated with consumption; 2) to minimize the inconvenience caused to consumers; and 3) to minimize the rebound peak occurrence. Potential solutions to the underlying multi-objective optimization problem are subject to a set of electrical and operational constraints related to home appliances categories and the utilization of distributed energy resources (DER) and energy storage systems (ESS). The use of the proposed model is illustrated in a realistic microgrid scenario where suitable solutions were found by the Non-Dominated Sorting Genetic Algorithm III (NSGA-III). These solutions determine new operational periods for home appliances as well as the utilization of DER and ESS for the planning horizon while considering consumer preferences. Besides helping consumers to take advantage of the benefits offered by DR, the experimental results show that the multi-objective DR model together with the NSGA-III algorithm can effectively minimize energy-consumption costs as well as reduce inconvenience costs and rebound peaks occurrences.

The voltage sag disturbance stands out as the most evident waveform change that is detected in electric networks, since the presence of these events in the network causes damages to the consumers. The first step in diagnosing the problem is to identify the location in the distribution system that is connected to the source causing the sinking disorder. This work presents a methodology based on clustering algorithm combined with decision rule to point out the region (cluster) that aggregates the place of origin. Clustering algorithm is responsible for analyzing the voltage signal data from different measurement nodes and separating these data into clusters. Then the Partial Decision Trees (PART) algorithm is responsible for defining the decision rule set that will confront the characteristics of each cluster and define which group aggregates the disturbance source location. For the clustering task, the k-means and DBSCAN clustering algorithms are evaluated and compared. The methodology was evaluated using the IEEE 34-bus test feeder system and the results show a hit rate higher than 98%.

The smart grids (SG) present a new architecture for the generation, transmission, and distribution of electrical energy, combined with new information and communication technologies aimed at guaranteeing the required demand, quality, availability and reliability of energy supply to its consumers. The Smart Meter (SM) is installed in each residence whose functionality is to record information related to consumption and send it to the Electric Power Company (EPC) through a two-way communication infrastructure. However, a large amount of data transmitted by the SG network can cause significant congestion in the communication infrastructure, leading to packet loss, increased latency and other issues affecting decision-making, which must be done in real-time. Thus, this work presents a performance analysis of the two-way communication infrastructure of SG in two experimental scenarios: with and without the application of an adaptive data compression mechanism. The communication infrastructure is presented logically and in practice, where a testbed is deployed using IEEE 802.11 communication technologies on the Message Queuing Telemetry Transport (MQTT) protocol to represent the communication infrastructure and analyze network performance using the Wireshark tool in a real scenario.

Lung cancer is the most prevalent cancer in the world and early detection and diagnosis enable more treatment options and a far greater chance of survival. In this work, we propose an algorithm based on 3D Convolutional Neural Network (CNN) to classify pulmonary nodules as benign or malignant in computed tomography images. Three architecture of 3D CNNs are proposed, containing different input sizes and numbers of convolutional layers. In addition, we investigated data augmentation techniques and modifications in the network training cost function to address the problem of imbalanced data. The best result was achieved for input size of 32×32×32 pixels, 2 blocks of convolutional layers and 2 pooling layers. Also, the modification of cost function achieved promising results, with accuracy of 0.9188, kappa of 0.8019, sensitivity of 0.8481, specificity of 0.9479 and AUC of 0.8980 in the test set during malignant nodule detection.

Coronavirus disease (COVID-19) has infected more than 3.6 million people worldwide and it is responsible for more than 250,000 deaths. A major problem faced in the diagnosis of COVID-19 is the inefficiency and scarcity of medical tests. The use of computed tomography (CT) has shown promise for the evaluation of patients with suspected COVID-19 infection. CT exam analysis is complex and requires specialist effort, which can lead to diagnostic errors. The use of CAD systems can minimize the problems generated by the analysis of CTs by specialists. This paper presents a methodology for diagnosing COVID-19 using convolutional neural network (CNN) for feature extraction in CT exams and its classification using XGBoost. The methodology consists of using a CNN to extract features from 708 CTs, 312 with COVID-19, and 396 Non-COVID-19. After the extracted data, we used XGBoost for classification. The results show an accuracy of 95.07, recall of 95.09, precision of 94.99, F-score of 95, AUC of 95, and a kappa index of 90. The results obtained show that the proposed methodology can be used as a diagnostic aid system by specialists.

Glaucoma is an eye disease responsible for the second most common cause of blindness in the world. The need to detect this disease in its early stages is notorious, considering that late treatment can cause loss of vision. In this context, computational methods are being developed to assist specialists in the task of analyzing ocular images to provide greater precision to the diagnosis. In this paper, we present a methodology for automatic classification of glaucoma using Capsule Network (CapsNet), a recent model of deep learning that analyzes the hierarchical spatial relationships between characteristics to represent images, so that it requires fewer training samples than traditional CNNs to achieve efficient classification. Before the execution of CapsNet, we applied a pre-processing step to the images, in order to highlight the characteristics. Our results were promising, with 90.90% accuracy, 86.88% recall, 94.64% precision, 90.59% f1-score, 0.904 AUC and 0.801 kappa index. The main contribution of our method is the fact that we have achieved promising results without the need to apply data to increase and segment the region of the optical disc. Thus, our study showed the potential of the capsules in identifying the relationships between the characteristics, even in the face of a reduced set of training.

The demand response (DR) aims to balance the purveyance and demand of electricity to maximize the reliability and efficiency of the energy supply process in the electrical power system (EPS). However, one of the main impediments to the insertion of DR in the residential context is the need of programming the use of various electrical appliances and the scheduling of renewable resources and storage system in the same time interval, that requires a range of specific knowledge and time availability of the consumer to handle the various home appliances. This article presents a preference-based multi-objective optimization model based on real-time electricity price to solve the problem of optimal residential load management. The proposal’s purpose is to minimize both the electricity consumption associated cost and the inconvenience caused to consumers. The proposed model was formalized as a nonlinear programming problem subject to a set of constraints associated with the consumption of electrical energy and operational aspects related to the residential appliance categories. The proposed multi-objective model was solved computationally by the Constrained Many-Objective Non-Dominated Sorted Genetic Algorithm (NSGA-III) to determine the new scheduling of residential appliances, renewable energy resources, and energy storage system utilization for the entire time horizon, considering consumer preferences. The results show that the multi-objective DR model proposed using the NSGA-III technique can minimize the total cost associated with energy consumption as well as reduce the inconvenience of consumers, besides helping consumers to take advantage of DR’s benefits without requiring manual intervention.

Inefficient parking solutions in metropolitan areas contribute to the creation of traffic jams that lead to stress for the drivers and contribute toward greenhouse emissions. Since parking issues affect most of the citizens, it is hard to imagine future smart cities with efficient parking solutions. With the introduction of the Internet of things (IoT), objects are able to connect to the Internet and may collect data in real time. The potential of these data is enormous and may allow cities to move from estimates to concrete data, which enable precise predictive models to be created and improves urban area planning and scaling. For this reason, parking solutions would benefit from the introduction of the IoT concept of connecting everything. The paper overviews the most common parking solutions, analyzing their pros and cons. Then, it introduces an IoT-oriented smart parking system that can be deployed indoor and/or outdoor. With the proposed solution, drivers can remotely monitor the vacancy status of parking spots in real time through a mobile app (or even Web). Furthermore, the solution acts as a repeater, extending the wireless signal to parking slots that are distant from the primary access point. This solution is evaluated, demonstrated, and validated through a prototype, and it is ready for use.

A significant challenge to invest in the stock market is to make complex decisions and evaluate different scenarios and trends in financial reports. An alternative capable of attending these needs is the use of algorithms for trading. This strategy may involve the use of artificial intelligence approaches to make effective decisions. Hence, this paper proposes the development of a solution to support decision making in the purchase and sale of stocks from the application of machine learning techniques. For this purpose, the performance of a recurrent neural network architecture, also known as long short-term memory (LSTM), was evaluated concerning its efficiency in the short-term prediction. The generated models were evaluated based on the root-mean-square error and mean absolute percentage error metrics. Results show the LSTM algorithm was capable of predicting stock prices with a low error approximately.

The significant increase in energy consumption by the growth of the population or by the use of new equipment has brought big challenges to the energy security as well as the environment. There is a need that consumers can track their daily use and understand consumption standards for better organizing themselves to obtain financial and energetic efficiency. With the improvement of smart networks technology for better energy supply, a smart meter is not just a simple measurement gadget anymore, but it has additional functions including smart equipment control, bidirectional communication that allows integration of users and networks, and other functionalities. Smart meters are the most fundamental components in smart power grids. Besides, the meters used with a management system can be utilized for monitoring and controlling home appliances and other gadgets according to the users' need. A solution of an integrated and single system should be more efficient and economical. Smart measurement systems allow monitoring the energy consumption of the final consumers while providing useful information about the energy quality. The information provided by these systems is used by the operators to enhance the energy supply, and different techniques can be also applied for this end, such as charge scheduling, management from the demand side, and non-intrusive load monitoring. The Internet of Things (IoT) is becoming a great ally in the management of smart distribution and energy consumption in smart systems scenarios. To address these issues, this paper proposes and demonstrates a new smart energy meter following an IoT approach and its associated costs and benefits. The developed device incorporates several communication interfaces. In order to easily integrate with any monitoring software solution, the meter has a multi-protocol connection. Finally, the provided solution is validated and demonstrated in real-life environments and it is also under use.

Demand-side load management methods need to find a compromise between utility and consumer interests. However, oftentimes these entities have conflicting objectives, particularly in price-based demand response programs. This paper considers the demand-side management problem from the perspective of both the energy service provider and its consumers in two steps: a local and decentralized load schedule optimization, followed by a system-wide demand profile optimization. A multi-objective optimization model is formulated for the load scheduling problem to minimize consumer expenses and the cost associated with inconvenient load shifting. Then, the proposed approach to improve the aggregated demand profile by reducing the peak-to-average ratio is described. Ant colony optimization algorithms were implemented and compared with other heuristics. Numerical results based on simulated load profiles show that ant algorithms were able to improve the aggregated demand load profile.

Computational Intelligence (CI) is a sub-branch of Artificial Intelligence (AI) that focus on studying adaptive mechanisms to enable intelligent behavior in complex environments. CI techniques have been successful in solving complex problems in many different knowledge areas. However, despite their usefulness, developing solutions based on CI techniques is not a trivial activity, since it involves the codification/adaptation of algorithms to specific context and problems. In this paper, we present and validate through a quasi-experiment a new paradigm to develop CI-based solutions using a more mature version of Athena (2.0): Computational Intelligence as a Service (CIaaS). Using this tool, both researchers and practitioners can design and evaluate CI-based solutions by dragging and dropping components in a visual environment, in a cloud-based platform. The results of the quasi-experiment suggest that our approach can help researchers to design and evaluate CI-based systems in a simple, reliable and fast way.

Internet of Things (IoT) has emerged as a new paradigm that allows different objects interconnection to create new smart services and applications. In this sense, key features such as traceability, unique identification, energy efficiency, heterogeneity of devices, scalability and ubiquity, are necessary at the network structure for efficient performance of IoT applications. As a single technology, it is not capable to provide all these features, then, the integration of Wireless Sensor Networks (WSNs) and Radio Frequency Identification (RFID) arises as an important approach for these solutions. In this paper, a framework aiming to provide the network features required for IoT applications is proposed, considering the challenges that comes from the integration of these two technologies. In this framework, two components are considered, the Fuzzy Q-Algorithm and the Fuzzy System-Based Route Classifier. The former comprises a fuzzy system to enhance an anti-collision protocol for RFID, and the latter is a fuzzy system that classifies routes and assists routing protocols in applications that use WSNs and RFID. Experiments show that framework provides the following benefits: improves the performance of RFID tags identification, lowers packet loss rates, decreases the nodes energy consumption, and shows improvements in network load balancing. Moreover, these benefits provide greater energy efficiency and quality of service for IoT applications.

Voltage sags are among the most relevant power quality disturbances. Furthermore, they also have high occurrence rates. Their stochastic nature makes monitoring difficult and causes significant losses to power utilities and customers. This paper presents an approach to overcome the problem of allocating power quality monitors. To do so, our approach accounts for topological coverage, unmonitored voltage sags, and the total cost of required equipment. We used NSGA-II to build our approach due to its efficiency in dealing with combinatorial problems. We also used the Monte Carlo simulation method to model the time series in our approach due to the random nature of power quality disturbances. To evaluate our approach, we simulated the IEEE 13-, 34- and 37-bus distribution systems using the DigSILENT Power Factory 15.1 software. The evaluation results show that our approach supported cost reduction associated with the installation of power quality monitors, both in terms of identifying adequate number and position of the performance monitors.

Robot soccer is a robot soccer category competition at FIRA. Robot soccer teams lacks though a formal programming procedure in order to standardize and properly model game strategies. Such lack of standardization causes several limitations on the competition as a science promoter, namely: (1) the difficulty of understanding other's code, (2) absence of abstraction for the conceiving of game strategies, (3) need for high programming skills in order to produce effective game strategies and (4) lack of clear separation between robots moving primitives and specific application domain concerns. In a previous paper, we have proposed a two-level semi-formal specification to provide game strategies modeling to robot teams in SimuroSot category. This paper presents the SimuroSot Development Kit, a visual tool which implements such specification and enables complete visual programming of robot strategies and subsequent automatic code generation in Lingo programming language. Experimentation scenarios with volunteers of different knowledge levels and formation have been settled. Results show that both specification and tool improve the learning curve of robot teams programming.

COVID-19 is a respiratory disease that, as of July 15th, 2021, has infected more than 187 million people worldwide and is responsible for more than 4 million deaths. An accurate diagnosis of COVID-19 is essential for the treatment and control of the disease. The use of computed tomography (CT) has shown to be promising for evaluating patients suspected of COVID-19 infection. The analysis of a CT examination is complex, and requires attention from a specialist. This paper presents a methodology for detecting COVID-19 from CT images. We first propose a convolutional neural network architecture to extract features from CT images, and then optimize the hyperparameters of the network using a tree Parzen estimator to choose the best parameters. Following this, we apply a selection of features using a genetic algorithm. Finally, classification is performed using four classifiers with different approaches. The proposed methodology achieved an accuracy of 0.997, a kappa index of 0.995, an AUROC of 0.997, and an AUPRC of 0.997 on the SARS-CoV-2 CT-Scan dataset, and an accuracy of 0.987, a kappa index of 0.975, an AUROC of 0.989, and an AUPRC of 0.987 on the COVID-CT dataset, using our CNN after optimization of the hyperparameters, the selection of features and the multi-layer perceptron classifier. Compared with pretrained CNNs and related state-of-the-art works, the results achieved by the proposed methodology were superior. Our results show that the proposed method can assist specialists in screening and can aid in diagnosing patients with suspected COVID-19.