The study aims to address traffic congestion in urban areas by using a modified Mobilenet-v2 convolutional-neural-network (CNN) model and optical-character-recognition (OCR) for emergency vehicle identification. The model, chosen for its high accuracy and low computational complexity, was trained on a locally obtained vehicle dataset of 243 samples. The system prioritizes emergency vehicle movement in congested traffic and regulates vehicle flow in designated lanes. The model demonstrated an average recognition accuracy of 99.69% in emergency vehicle identification, outperforming existing models in terms of precision, recall, and F1-score for bus, car, and emergency vehicle identification. The modified MobileNet-v2 achieved perfect precision, recall, and F1-score on the validation dataset under the defined experimental conditions. The study suggests that using a larger dataset in future work could improve the model's generalizability. This innovative approach to automatic traffic control, incorporating MobileNet-v2 and OCR, offers a solution to delayed emergency response time and improves overall traffic management efficiency.
Lawal et al. (Mon,) studied this question.