Rommel Jair Gutiérrez Yépez

Mr. Rommel Jair Gutiérrez Yépez

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 October 2024 in IEEE Access
Verified
Journal Article
Open access
William Eduardo VillegasWilliam Eduardo Villegas
William Eduardo Villegas
RYRommel Jair Gutiérrez Yépez
Rommel Jair Gutiérrez Yépez
ISIván Sánchez-Salazar
Iván Sánchez-Salazar
AMAracely Mera-Navarrete
Aracely Mera-Navarrete
The increasing use of Internet of Things (IoT) devices in critical sectors has increased exposure to security threats, making protecting these systems a priority challenge. Based on static configurations, traditional security approaches have proven ineffective in the face of the dynamic nature of emerging threats, as they cannot adapt in real time to changes in the environment or new attack vectors. This work proposes an adaptive security framework for Internet of Things (IoT) systems capable of autonomously detecting, mitigating, and adapting to various threats, improving precision and response times, and optimizing energy consumption. The framework was implemented in a distributed Internet of Things environment, using adaptive architectures based on the Robot Operating System (ROS) and microservices orchestration with Kubernetes. The results showed a significant improvement in response time, with a reduction of 44%, reaching an average of 250 milliseconds, compared to 450 milliseconds for static approaches. Furthermore, a 92% precision in threat detection was achieved, reducing false positives to 4% and false negatives to 6%. Power consumption was controlled, reaching a maximum of 160 milliamp-hours after facing multiple threats, confirming the system’s efficiency in resource-limited environments. These results demonstrate that the proposed adaptive framework is a robust and efficient solution for security in Internet of Things environments, overcoming the limitations of traditional approaches and ensuring adequate protection without compromising energy efficiency.

5 December 2024 in IEEE Access
Verified
Journal Article
Open access
William Eduardo VillegasWilliam Eduardo Villegas
William Eduardo Villegas
JGJaime Govea
Jaime Govea
RYRommel Jair Gutiérrez Yépez
Rommel Jair Gutiérrez Yépez
ANAlexandra Maldonado Navarro
Alexandra Maldonado Navarro
AMAracely Mera-Navarrete
Aracely Mera-Navarrete
Cybersecurity is characterized by its dynamism and complexity, with malicious actors continually developing new strategies to evade traditional intrusion detection systems. These systems, while once robust, now often need help to adapt to evolving threat tactics, resulting in a high incidence of false positives and inadequate response capabilities. This scenario presents a critical challenge: how can defenders stay one step ahead of threats without compromising operational efficiency and security effectiveness? The problem is the need for intrusion detection systems that react to known threats and anticipate and adapt to emerging threats in real-time. This work examines the implementation and effectiveness of an adaptive intrusion detection system using deep learning algorithms to strengthen cybersecurity. The research focused on evaluating the system’s ability to identify and neutralize cyber threats more efficiently and accurately than traditional methods. Quantitative analysis showed that AIDS significantly improved in several key metrics: precision increased by 12.5%, reaching 90%, while recall enhanced by 13.3%, reaching 85%. Furthermore, the F1-score experienced an increase of 12.9%, settling at 87.5%. Qualitative evaluations complemented these results through case studies and testimonials from IT staff, which corroborated the improvement in the detection and response to security incidents. The results reveal that the adaptive intrusion detection system, with its machine learning approach, not only improves threat detection and management but also optimizes operational efficiency, reducing false positives and accelerating response times.

13 March 2025 in IEEE Access
Verified
Journal Article
Open access
William Eduardo VillegasWilliam Eduardo Villegas
William Eduardo Villegas
RYRommel Jair Gutiérrez Yépez
Rommel Jair Gutiérrez Yépez
ANAlexandra Maldonado Navarro
Alexandra Maldonado Navarro
AMAracely Mera-Navarrete
Aracely Mera-Navarrete
The Internet of Things (IoT) expansion has exposed connected devices to significant security vulnerabilities, particularly in terms of authentication and authorization. Traditional solutions, such as centralized servers or Proof of Work (PoW)–based blockchain, are unfeasible due to the resource limitations of IoT devices, such as their low processing capacity and dependence on batteries. This study proposes a lightweight blockchain system based on a simplified Proof of Stake (PoS) consensus mechanism designed to optimize energy consumption and improve resilience to attacks in IoT networks. The system implements a hierarchical network topology that improves data propagation and transmission times, significantly reducing latency compared to distributed topologies. In addition, it uses lightweight cryptographic algorithms such as ECDSA for authentication and AES-128 for authorization, ensuring transaction security without compromising the efficiency of IoT devices. The results show that the system reduces energy consumption by 54% compared to PoW solutions in high-load scenarios while maintaining an average latency below 30 ms. Furthermore, the system achieved a 92.5% attack detection rate under low malicious load, demonstrating its effectiveness in high-threat environments. This work offers a scalable and efficient solution that optimizes security and performance in IoT networks, opening new possibilities for its application in critical infrastructures and real-time sensor networks.

1 January 2025 in IEEE Access
Verified
Journal Article
Open access
RYRommel Jair Gutiérrez Yépez
Rommel Jair Gutiérrez Yépez
William Eduardo VillegasWilliam Eduardo Villegas
William Eduardo Villegas
ANAlexandra Maldonado Navarro
Alexandra Maldonado Navarro
Sergio Luján-MoraSergio Luján-Mora
Sergio Luján-Mora

20 November 2024 in IEEE Access
Verified
Journal Article
Open access
William Eduardo VillegasWilliam Eduardo Villegas
William Eduardo Villegas
RYRommel Jair Gutiérrez Yépez
Rommel Jair Gutiérrez Yépez
ANAlexandra Maldonado Navarro
Alexandra Maldonado Navarro
AMAracely Mera-Navarrete
Aracely Mera-Navarrete
Federated learning is presented as an effective solution to train artificial intelligence models on the Internet of Things networks without centralizing data, thus preserving privacy and minimizing security risks. However, its implementation in low-power devices, such as Tiny Machine Learning, faces significant challenges due to the processing, memory, and unstable connectivity limitations that characterize these environments. This study addresses these issues by developing a federated system optimized for Tiny Machine Learning devices, integrating differential privacy and encryption techniques adapted to their constraints. The methodology employed includes the evaluation of model precision and energy consumption in variable communication scenarios, as well as heterogeneous workloads. The results show that federated learning reduces energy consumption by 33% compared to the centralized approach, reaching an average of 100 mW. Furthermore, implementing differential privacy-maintained precision with a loss of only 1.2% demonstrates the feasibility of protecting sensitive data without substantially affecting performance. However, precision dropped by 8% in high communication loss scenarios, underlining the importance of stable connectivity for optimal system performance.

8 July 2024 in IEEE Access
Verified
Journal Article
Open access
William Eduardo VillegasWilliam Eduardo Villegas
William Eduardo Villegas
JGJaime Govea
Jaime Govea
RYRommel Jair Gutiérrez Yépez
Rommel Jair Gutiérrez Yépez
AMAracely Mera-Navarrete
Aracely Mera-Navarrete
The current education landscape is marked by the growing integration of artificial intelligence technologies, which seek to improve interaction and efficiency in educational environments. However, many of these tools operate on separate platforms, complicating their use and reducing their potential effectiveness. This study introduces InteractiveClass, an innovative tool that fully integrates with Microsoft Teams, facilitating automated assessment and student interaction without requiring multiple interfaces. InteractiveClass encourages active student participation through real-time questions, quizzes, discussion forums, and other activities. In addition, it uses AI capabilities to offer immediate and personalized feedback and perform automatic evaluations. This integration improves usability and promotes student engagement in hybrid or fully online environments. Study results reveal significant improvements in assessment efficiency and student satisfaction. Students who used the tool showed a 30% increase in class participation and improved their grades by 25% compared to those who did not use it. Additionally, the tool demonstrated high consistency in assessments, with a precision of 95% compared to manual assessments. These findings underline the potential of InteractiveClass to transform education through AI technology, offering a practical and effective solution to the challenges of modern education.