Traffic sign Identification acts as a key role in the production of automated vehicles and intelligent driver assistance systems by accessing vehicles to understand road environments and making safe decisions through the identification of signs. To improve robustness against real-world variations, Rotation, zoom, brightness variation, and shifting were some of the data augmentation techniques used and class weight balancing was applied to mitigate dataset imbalance and enhance performance. The model was trained using image preprocessing, normalization, and evaluated through accuracy metrics and a confusion matrix. The proposed model gained a test accuracy of 98.38% and a validation accuracy of 99.87%, showing enhanced stability, reliability, and generalization. When compared to the traditional approaches, this CNN model shows clear advantages which are; the earlier method required significantly more computational sources, which involved more rigid training methods, and lacked efficiency for real-time deployment. In contrast, the Sequential CNN is lightweight, computationally efficient, and better suited for real-time and edge-device applications without compromising accuracy. Overall, the Sequential CNN model effectively improves computational and architectural challenges, which makes it a better solution for real-time transportation network intelligence and autonomous vehicle applications.