Authors : Sami Shadman Sakib Md; Arick Md Abdul Mahed; Hassan Md Hasibul; Mehedy Abdul Mukit Al; Wu Zhong

Volume/Issue : Volume 11 - 2026, Issue 5 - May

Google Scholar : https://tinyurl.com/4zy2pxzt

Scribd : https://tinyurl.com/bderuuxf

DOI : https://doi.org/10.38124/ijisrt/26may1386

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Traffic Sign Recognition (TSR) is a key technology for self-driving cars and smart driver assistance systems, helping vehicles understand road signs in real time to keep roads safe. This study presents a simple Convolutional Neural Network (CNN) that we built from scratch using the German Traffic Sign Recognition Benchmark (GTSRB) dataset. Our model reached a test accuracy of 96.53%, which is impressive for its small size. We made this model lightweight so it can work on basic devices like those in cars, balancing speed and accuracy. We used smart ways to prepare the data, like resizing images and adding variety, and stopped training early to avoid mistakes. Our results, shown through accuracy and loss graphs, a confusion matrix, and comparisons with bigger models, prove that a small model can work just as well as complex ones. This makes it great for real-world use and learning purposes, showing how simple AI can make driving safer.

Keywords : Traffic Sign Recognition, Convolutional Neural Network, GTSRB Dataset, Lightweight Model, Autonomous Driving, Edge Computing, Deep Learning.

References :

1. J. Chen, L. Zhang, and Y. Wang, “Multiscale feature fusion for efficient traffic sign recognition,” IEEE Trans. Intell. Transp. Syst., vol. 22, no. 8, pp. 5123–5132, Aug. 2021.
2. S. He, X. Li, and Q. Chen, “Lightweight CNN for edge-based traffic sign recognition,” IEEE Access, vol. 10, pp. 32456–32465, 2022.
3. Y. Liu, Z. Wang, and H. Zhou, “Optimized YOLOv5 for traffic sign detection and classification,” J. Real-Time Image Process., vol. 19, no. 3, pp. 789–799, 2022.
4. W. Zhang, J. Lin, and K. Yang, “Attention-enhanced CNN for robust traffic sign recognition,” IEEE Trans. Veh. Technol., vol. 72, no. 5, pp. 6456–6466, May 2023.
5. C. Lin, Q. Zhao, and L. Wu, “Dual-attention CNN for traffic sign recognition,” Comput. Vis. Image Underst., vol. 230, pp. 103–115, 2023.
6. A. Singh and R. Verma, “Ensemble learning for traffic sign classification,” IEEE Intell. Syst., vol. 37, no. 4, pp. 89–97, 2022.
7. P. Patel and S. Roy, “Hybrid EfficientNet-CNN for traffic sign recognition,” IEEE Trans. Neural Netw. Learn. Syst., vol. 34, no. 7, pp. 4123–4132, Jul. 2023.
8. S. Rahman and M. Arif, “Data-driven augmentation for robust TSR,” IEEE Trans. Image Process., vol. 33, pp. 234–245, 2024.
9. M. Rahimi and S. Amini, “EfficientNet-Lite for mobile TSR,” IEEE Embed. Syst. Lett., vol. 15, no. 2, pp. 67–70, Jun. 2023.
10. L. Chan and Y. Zhou, “Ethical considerations in autonomous driving AI,” IEEE Technol. Soc. Mag., vol. 42, no. 1, pp. 45–53, 2023.
11. A. Rahman, S. Lee, and J. Kim, “GAN-based augmentation for TSR,” IEEE Access, vol. 11, pp. 56789–56798, 2023.
12. M. Arif, T. Khan, and R. Ali, “Domain adaptation for global TSR,” IEEE Trans. Intell. Veh., vol. 9, no. 3, pp. 1234–1243, Mar. 2024.
13. H. Xu, J. Zhang, and L. Chen, “Real-time TSR for embedded systems,” IEEE Internet Things J., vol. 11, no. 4, pp. 7890–7900, Apr. 2024.
14. R. Sharma, P. Kumar, and S. Jain, “Explainable AI for TSR,” IEEE Comput. Intell. Mag., vol. 18, no. 2, pp. 34–43, May 2023.
15. S. Ghosh and M. Alam, “Accessible AI education for TSR,” IEEE Educ. Rev., vol. 39, no. 1, pp. 56–64, 2023.
16. T. Kim, H. Park, and J. Lee, “Edge computing for real-time TSR,” IEEE Trans. Comput., vol. 73, no. 6, pp. 1456–1467, Jun. 2024.
17. J. Wang, S. Chen, and M. Yang, “Multimodal TSR systems,” IEEE Trans. Multimedia, vol. 26, no. 3, pp. 2345–2356, Mar. 2024.
18. H. Li, R. Zhang, and C. Liu, “Synthetic data for TSR enhancement,” IEEE J. Sel. Topics Signal Process., vol. 17, no. 4, pp. 789–800, Aug. 2023.
19. Q. Zhao, Y. Li, and X. Wang, “Robust TSR under adverse conditions,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 47, no. 1, pp. 89–102, Jan. 2025.
20. S. Roy and M. Alam, “Behavioral modeling with TSR,” IEEE Trans. Veh. Technol., vol. 72, no. 10, pp. 12345–12356, Oct. 2023.
21. A. Mahmud, S. Rahman, and T. Islam, “Training TSR on degraded datasets,” IEEE Trans. Intell. Transp. Syst., vol. 24, no. 9, pp. 10234–10245, Sep. 2023.
22. R. Gupta, S. Sharma, and V. Kumar, “Active learning for efficient TSR,” IEEE Trans. Neural Netw. Learn. Syst., vol. 35, no. 5, pp. 6789–6800, May 2024.