The College of Engineering at Al-Iraqia University Discusses a Master’s Thesis on Traffic Congestion Classification Using Audio Analysis and Machine Learning

The thesis submitted by researcher Omar Abdullah Hassan included presenting a new model designed to identify and classify traffic volume as congested (abnormal flow) and non-congested (normal flow) based on an audio dataset. First, the dataset was collected using signal processing techniques and the machine learning model was used to classify the audio data into normal and abnormal traffic flow.

The thesis aims to demonstrate the effectiveness of the proposed method in accurately classifying traffic conditions. From these results, it was noted that LSTM provides a better accuracy of 98.25% on the RTD dataset and 98.69% on the IDMT- Traffic dataset in evaluating traffic congestion classification.