VEHICLE SOUND DETECTION USING DEEP LEARNING MODELS

Identifying and categorizing vehicle-related sounds plays an essential role in areas such as traffic surveillance, autonomous driving technologies, and the assessment of urban noise levels. Recent advances in deep learning have significantly improved audio-based detection systems, particularly under complex and noisy conditions. In this work, we analyze and compare the effectiveness of advanced deep learning approaches, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and combined hybrid models, for vehicle sound recognition. A carefully prepared dataset containing various vehicle audio samples is used to evaluate model performance based on precision, recall, and F1-score. The experimental findings reveal that the hybrid CNN–LSTM architecture delivers superior results, achieving an F1-score of 96.3% and an accuracy of 95.8%, thereby surpassing individual deep learning models. These outcomes demonstrate the strong capability of deep learning techniques to support intelligent transportation applications and urban acoustic monitoring systems