ISEP – GECAD – Artigos
Permanent URI for this collection
Browse
Recent Submissions
- SoK: Realistic Adversarial Attacks and Defenses for Intelligent Network Intrusion DetectionPublication . Vitorino, João; Praça, Isabel; Maia, EvaMachine Learning (ML) can be incredibly valuable to automate anomaly detection and cyber-attack classification, improving the way that Network Intrusion Detection (NID) is performed. However, despite the benefits of ML models, they are highly susceptible to adversarial cyber-attack examples specifically crafted to exploit them. A wide range of adversarial attacks have been created and researchers have worked on various defense strategies to safeguard ML models, but most were not intended for the specific constraints of a communication network and its communication protocols, so they may lead to unrealistic examples in the NID domain. This Systematization of Knowledge (SoK) consolidates and summarizes the state-of-the-art adversarial learning approaches that can generate realistic examples and could be used in real ML development and deployment scenarios with real network traffic flows. This SoK also describes the open challenges regarding the use of adversarial ML in the NID domain, defines the fundamental properties that are required for an adversarial example to be realistic, and provides guidelines for researchers to ensure that their future experiments are adequate for a real communication network.
- Constrained Adversarial Learning and its applicability to Automated Software Testing: a systematic reviewPublication . Vitorino, João; Dias, Tiago; Fonseca, Tiago; Maia, Eva; Praça, IsabelEvery novel technology adds hidden vulnerabilities ready to be exploited by a growing number of cyber-attacks. Automated software testing can be a promising solution to quickly analyze thousands of lines of code by generating and slightly modifying function-specific testing data to encounter a multitude of vulnerabilities and attack vectors. This process draws similarities to the constrained adversarial examples generated by adversarial learning methods, so there could be significant benefits to the integration of these methods in automated testing tools. Therefore, this systematic review is focused on the current state-of-the-art of constrained data generation methods applied for adversarial learning and software testing, aiming to guide researchers and developers to enhance testing tools with adversarial learning methods and improve the resilience and robustness of their digital systems. The found constrained data generation applications for adversarial machine learning were systematized, and the advantages and limitations of approaches specific for software testing were thoroughly analyzed, identifying research gaps and opportunities to improve testing tools with adversarial attack methods.
- Towards Adversarial Realism and Robust Learning for IoT Intrusion Detection and ClassificationPublication . Vitorino, João; Praça, Isabel; Maia, EvaThe internet of things (IoT) faces tremendous security challenges. Machine learning models can be used to tackle the growing number of cyber-attack variations targeting IoT systems, but the increasing threat posed by adversarial attacks restates the need for reliable defense strategies. This work describes the types of constraints required for a realistic adversarial cyber-attack example and proposes a methodology for a trustworthy adversarial robustness analysis with a realistic adversarial evasion attack vector. The proposed methodology was used to evaluate three supervised algorithms, random forest (RF), extreme gradient boosting (XGB), and light gradient boosting machine (LGBM), and one unsupervised algorithm, isolation forest (IFOR). Constrained adversarial examples were generated with the adaptative perturbation pattern method (A2PM), and evasion attacks were performed against models created with regular and adversarial training. Even though RF was the least affected in binary classification, XGB consistently achieved the highest accuracy in multi-class classification. The obtained results evidence the inherent susceptibility of tree-based algorithms and ensembles to adversarial evasion attacks and demonstrate the benefits of adversarial training and a security-by-design approach for a more robust IoT network intrusion detection and cyber-attack classification.
- Evolutionary Algorithms applied to the Intraday Energy Resource Scheduling in the Context of Multiple AggregatorsPublication . Almeida, José; Soares, João; Lezama, Fernando; Canizes, Bruno; Vale, ZitaThe growing number of electric vehicles (EVs) on the road and renewable energy production to meet carbon reduction targets has posed numerous electrical grid problems. The increasing use of distributed energy resources (DER) in the grid poses severe operational issues, such as grid congestion and overloading. Active management of distribution networks using the smart grid (SG) technologies and artificial intelligence (AI) techniques by multiple entities. In this case, aggregators can support the grid's operation, providing a better product for the end-user. This study proposes an effective intraday energy resource management starting with a day-ahead time frame, considering the uncertainty associated with high DER penetration. The optimization is achieved considering five different metaheuristics (DE, HyDE-DF, DEEDA, CUMDANCauchy++, and HC2RCEDUMDA). Results show that the proposed model is effective for the multiple aggregators with variations from the day-ahead around the 6 % mark, except for the final aggregator. A Wilcoxon test is also applied to compare the performance of the CUMDANCauchy++ algorithm with the remaining. CUMDANCauchy++ shows competitive results beating all algorithms in all aggregators except for DEEDA, which presents similar results.
- Intraday Energy Resource Scheduling for Load Aggregators Considering Local MarketPublication . Almeida, José; Soares, João; Canizes, Bruno; Razo-Zapata, Ivan; Vale, ZitaDemand response (DR) programs and local markets (LM) are two suitable technologies to mitigate the high penetration of distributed energy resources (DER) that is vastly increasing even during the current pandemic in the world. It is intended to improve operation by incorporating such mechanisms in the energy resource management problem while mitigating the present issues with Smart Grid (SG) technologies and optimization techniques. This paper presents an efficient intraday energy resource management starting from the day-ahead time horizon, which considers load uncertainty and implements both DR programs and LM trading to reduce the operating costs of three load aggregator in an SG. A random perturbation was used to generate the intraday scenarios from the day-ahead time horizon. A recent evolutionary algorithm HyDE-DF, is used to achieve optimization. Results show that the aggregators can manage consumption and generation resources, including DR and power balance compensation, through an implemented LM.
- Boosting the Usage of Green Energy for EV Charging in Smart Buildings Managed by an Aggregator Through a Novel Renewable Usage IndexPublication . Guzman, Cindy P.; Bañol Arias, Maria Nataly; Franco, John Fredy; Soares, João; Vale, Zita; Romero, RubenThe growing trend of electric vehicles (EVs) and building integrated photovoltaics (BIPVs) is a promising means to reduce related climate change issues. EV loads can be managed via an aggregator to maximize the usage of green energy produced by photovoltaic units (PV) through smart charging strategies that exploit controllable EV demand connected to BIPV. Previous works have focused on the EV charging coordination in a smart BIPV, although without an optimization that encourages EV charging with the energy produced by the PV units. This paper proposes an aggregation strategy that maximizes a green energy index (GEI) for the smart charging coordination of EVs, which takes advantage of periods with high PV availability to charge the EV batteries; moreover, a post-processing stage for the GEI provides EV owners with information about the percentage of charged energy, period by period, that comes from PV generation. The results for a case study with 510 EVs integrated with 17 smart BIPVs show that the strategy effectively optimizes the usage of the energy produced by the PV units to charge the EVs, contributes to reduce non-renewable energy consumption of the building sector, and satisfies the EV owners’ energy requirements for transportation.
- Joint Optimal Allocation of Electric Vehicle Charging Stations and Renewable Energy Sources Including CO2 EmissionsPublication . Lima, Tayenne Dias de; Franco, John F.; Lezama, Fernando; Soares, João; Vale, ZitaIn the coming years, several transformations in the transport sector are expected, associated with the increase in electric vehicles (EVs). These changes directly impact electrical distribution systems (EDSs), introducing new challenges in their planning and operation. One way to assist in the desired integration of this technology is to allocate EV charging stations (EVCSs). Efforts have been made towards the development of EVCSs, with the ability to recharge the vehicle at a similar time than conventional vehicle filling stations. Besides, EVs can bring environmental benefits by reducing greenhouse gas emissions. However, depending on the energy matrix of the country in which the EVs fleet circulates, there may be indirect emissions of polluting gases. Therefore, the development of this technology must be combined with the growth of renewable generation. Thus, this proposal aims to develop a mathematical model that includes EVs integration in the distribution system. To this end, a mixed-integer linear programming (MILP) model is proposed to solve the allocation problem of EVCSs including renewable energy sources. The model addresses the environmental impact and uncertainties associated with demand (conventional and EVs) and renewable generation. Moreover, an EV charging forecast method is proposed, subject to the uncertainties related to the driver's behavior, the energy required by these vehicles, and the state of charge of the EVs. The proposed model was implemented in the AMPL modelling language and solved via the commercial solver CPLEX. Tests with a 24-node system allow evaluating the proposed method application
- Local Electricity Markets for Electric Vehicles: An Application Study Using a Decentralized Iterative ApproachPublication . Faia, Ricardo; Soares, João; Fotouhi Ghazvini, Mohammad Ali; Franco, John F.; Vale, ZitaLocal electricity markets are emerging solutions to enable local energy trade for the end users and provide grid support services when required. Various models of local electricity markets (LEMs) have been proposed in the literature. The peer-to-peer market model appears as a promising structure among the proposed models. The peer-to-peer market structure enables electricity transactions between the players in a local energy system at a lower cost. It promotes the production from the small low–carbon generation technologies. Energy communities can be the ideal place to implement local electricity markets as they are designed to allow for larger growth of renewable energy and electric vehicles, while benefiting from local transactions. In this context, a LEM model is proposed considering an energy community with high penetration of electric vehicles in which prosumer-to-vehicle (P2V) transactions are possible. Each member of the energy community can buy electricity from the retailer or other members and sell electricity. The problem is modeled as a mixed-integer linear programing (MILP) formulation and solved within a decentralized and iterative process. The decentralized implementation provides acceptable solutions with a reasonable execution time, while the centralized implementation usually gives an optimal solution at the expense of reduced scalability. Preliminary results indicate that there are advantages for EVs as participants of the LEM, and the proposed implementation ensures an optimal solution in an acceptable execution time. Moreover, P2V transactions benefit the local distribution grid and the energy community.
- Preliminary results of advanced heuristic optimization in the risk-based energy scheduling competitionPublication . Almeida, José; Lezama, Fernando; Soares, João; Vale, Zita; Canizes, BrunoIn this paper, multiple evolutionary algorithms are applied to solve an energy resource management problem in the day-ahead context involving a risk-based analysis corresponding to the proposed 2022 competition on evolutionary computation. We test numerous evolutionary algorithms for a risk-averse day-ahead operation to show preliminary results for the competition. We use evolutionary computation to follow the competition guidelines. Results show that the HyDE algorithm obtains a better solution with lesser costs when compared to the other tested algorithm due to the minimization of worst-scenario impact.
- Rating and Remunerating the Load Shifting by Consumers Participating in Demand Response ProgramsPublication . Silva, Cátia; Faria, Pedro; Vale, ZitaEffective and active consumers providing flexibility through Demand Response (DR) programs have three important aspects: rating each consumer according to previous participation, remuneration of that participation, and determining the rebound effect of consumption after the event. In this paper, the authors design a rate to classify and select the proper participants for a DR event considering the context in which the event is triggered. The aggregator estimated the shifting of consumption to periods after the event is modeled, and the respective remuneration is estimated under different scenarios. This shifting can be done in several time frames in the future. The scenarios are developed to test the acceptable time range in which the load should be allocated according to the rebound effect. The results show that a higher time range can avoid huge peak consumption, optimizing the system operation with benefits for consumers, DSO, and the aggregator.