New Method of Driver Drowsiness Detection System Based on Fuzzy Logic and Artificial Neural Networks

According to experts, around a quarter of serious accidents on highways are due to driver drowsiness, underlining the crucial importance of early detection of this dangerous phenomenon. In this paper, we present an innovative approach to creating an intelligent system for detecting driver drowsiness in real time, by combining artificial neural networks (ANNs) with fuzzy logic (FL). Artificial neural networks, in particular the Multilayer Perceptron Model (MLP), are used to perform deep learning and predict the onset of drowsiness by exploiting data acquired by the system. However, due to variations in brightness, noise and blur factors, ANN can be less reliable under certain conditions. In such situations, fuzzy logic is activated to compensate and provide robust drowsiness detection. This is done by extracting the driver's facial and ocular symptoms, thus generating an appropriate alert signal to indicate the state of drowsiness. In addition, advanced image processing techniques are implemented where necessary to enhance the system's accuracy. This comprehensive approach offers an innovative solution for the early detection of driver drowsiness, with the potential to save lives and reduce accidents associated with this dangerous condition.