Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{201339,
        author = {Vinisha A and Shashank G and Shivaparasad and Hemanth Kumar D and Divyaraj G N},
        title = {Synthetic-to-Smart: Enhancing EV Battery Insights Using TimeGAN and Temporal Fusion Transformer},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {4648-4654},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=201339},
        abstract = {Electric vehicles (EVs) are changing the contempo-rary transportation, yet the mass utilization of the vehicles is largely conditional upon the trustworthy functioning of battery systems and their correct monitoring. The ability to predict the most important battery health indicators: State of Charge (SOC), State of Health (SOH), and Remaining Useful Life (RUL) is a problematic issue because the real-world data is usually not accessible, and such data is usually noisy and imbalanced. Con-ventional methods of battery degradation analysis find it difficult to reflect the nonlinear, time-varying, and complicated nature of battery behavior. This work suggests a hybrid framework, which integrates Time-series Generative Adversarial Networks (TimeGAN) to generate synthetic data with a Temporal Fu-sion Transformer (TFT) to predictive model to overcome these difficulties. TimeGAN is a neural network that is utilized to create realistic multivariate time-series data that both maintains temporal patterns and feature-relationships, and it can overcome the problem of data scarcity and distribution. The t-SNE visual-ization and statistical similarity measure is used to evaluate the quality of the generated data and make sure that it is consistent with actual battery signals. The TFT model is then trained on the augmented dataset and it utilizes attention mechanisms and gated architectures to learn long-term dependencies and nonlinear patterns of degradation. Consequently, the model can make correct predictions of SOC, SOH and RUL. On the whole, the offered framework increases the accuracy of prediction, as well as the reliability of battery monitoring systems.},
        keywords = {Battery management system, TimeGAN, Tempo-ral fusion transformer, State of charge, State of health, Remaining useful life, Synthetic data generation},
        month = {May},
        }