Enhancing Traffic Systems

The automobile industry is witnessing a global surge in the development and deployment of autonomous vehicles. These vehicles are expected to safely navigate highways, recognize obstacles, and interpret both temporary and permanent traffic signs, making Traffic Sign Recognition (TSR) essential for enhancing safety and efficiency. While significant advancements have been made in recognizing traffic signs in English, Arabic Traffic Sign (ArTS) recognition remains underexplored. This study addresses this gap by optimizing seven deep learning (DL) architectures—MobileNetV2, EfficientNetB0, DenseNet121, ResNeXt50, InceptionV3, NASNetLarge, and ResNet-50—on a recently compiled Arabic Traffic Sign Recognition (ArTS) dataset. Our experimental results demonstrate the effectiveness of these optimized models, suggesting their potential to improve the efficiency and safety of autonomous systems in Arabic-speaking regions.