Dr. Santos Gracia Villar, president of the Iberoamerican University Foundation (FUNIBER), is participating in a study that reviews and proposes approaches based on machine learning and deep learning to enhance early fault detection in new energy vehicles.
The transition to cleaner transportation has become a global priority, driven by the need to reduce the carbon footprint and move toward sustainable energy sources. In this context, new energy vehicles (NEVs)—such as electric vehicles, plug-in hybrids, and hydrogen fuel cell vehicles—are gaining prominence as an alternative to combustion-engine cars. However, their widespread adoption still faces technical and operational challenges, particularly related to battery performance, the availability of charging infrastructure, and the reliability of electronic systems under real-world conditions.
In line with these challenges, the study, titled “Enhancing fault detection in new energy vehicles via novel ensemble approach,” focuses on how the analysis of large volumes of sensor data—collected during vehicle operation—can help anticipate failures, optimize battery health management, and improve energy consumption in real time. These capabilities are key to reducing downtime, preventing costly repairs, and increasing user confidence in sustainable mobility technologies.
One of the study’s central contributions is the proposal of a comprehensive fault detection framework that combines traditional machine learning models with advanced deep learning architectures. Among the models evaluated are approaches such as logistic regression, passive-aggressive classifiers, Ridge classifiers, and perceptrons, along with deep learning techniques such as convolutional neural networks (CNNs), artificial neural networks (ANNs), and gated recurrent units (GRUs), which are particularly well-suited for sequential data.
A hybrid model to improve accuracy and interpretability
The study also introduces a hybrid model called GRULogX, which integrates gated recurrent units with logistic regression. This combination aims to leverage, on the one hand, the ability of gated recurrent units to extract temporal patterns in time series and, on the other, the probabilistic interpretation offered by logistic regression. To validate its performance, the team used a real-world fault diagnosis dataset, allowing the results to be applied to practical scenarios.
Among the evaluated models, the proposed GRULogX approach stands out for achieving 99% accuracy, in addition to reporting high accuracy and completeness scores. This model design strengthens early fault detection, facilitating predictive maintenance strategies and contributing to improved availability and safety of electric vehicles. More broadly, these advances align with the need to support clean transportation solutions, reducing operational disruptions and costs associated with faults not detected in time.
If you want to learn more about this study, click here.
To read more research, check out the UNEATLANTICO repository.
The Iberoamerican University Foundation (FUNIBER) offers scholarships to study master’s programs in the field of technology, such as the Master’s in Strategic Management in Software Engineering. This program is designed to train experts capable of integrating advanced technologies, such as machine learning and software development, into business strategy, contributing to the improvement of critical processes such as early fault detection, operational efficiency, and data-driven decision-making.
Take the next step in your professional career. Enroll in this master’s program and become a leader in the technological innovation of the future.