<p>This study investigates the ways in which machine vision, deep learning (DL), and sophisticated algorithms are transforming road safety, with a special emphasis on the innovative application of Convolutional Neural Networks (CNNs) to comprehend the conduct of human drivers. By analyzing frames captured by in-car cameras, these models adeptly interpret subtle patterns, enabling the anticipation of a diverse range of driving behaviours. The paper underscores the promise of DL in shaping a considerate future for car safety, marking a paradigm shift in preventing road accidents through the convergence of cutting-edge technologies. This paper introduces a novel Hybrid Convolutional Neural Network (HCNN) for detecting human driver behaviour to enhance road safety. It concludes with a comprehensive review of current approaches, recognizing the challenges presented by subtle early models. It identifies that the proposed method is the most effective for human driver behaviour detection, achieving a detection accuracy of 98%. The experimental results have also proven that the proposed method achieves superior performance in terms of evaluation metrics (i.e. Precision = 0.878, Recall = 0.93) compared to some existing models.</p>