A Hybrid Model Utilizing Principal Component Analysis and Artificial Neural Networks for Driving Drowsiness Detection

The detection of drowsiness while driving plays a vital role in ensuring road safety. Existing methods need to reduce external interference and sensor intrusiveness, their algorithms must be modified improve accuracy, stability, timeliness. In order realize fast accurate using physiological data that can collected non-intrusively, hybrid model with principal component analysis artificial neural networks was proposed this study. Principal used remove the noise redundant information from original data, were classify processed data. Three other models designed for comparison, including classic machine learning algorithms, single networks, algorithms. results indicated average accuracy exceeded 97%, training time lower than 0.3 s, standard deviation model’s 0.7%, indicating could detect more accurately quickly comparison stability. Thus, help detection. This method applied active warning systems (AWS) intelligent vehicles future.