An Unsupervised Drowsiness Detection Architecture Using Fuzzy Classifier

Accidents occur predominantly due to the loss of alertness of the driver or due to the drowsiness of the driver.In order to avoid such mishaps the continuous monitoring of the physiological signals like drowsiness and fatigue would provide the solution to avert the accidents before occurrence. The proposed Short Time Fourier Transform-based Improved Particle Swarm Optimization using Fuzzy Classifier (STFTIPSF) methodology makes use of STFT for yielding a good number of feature, fuzzy classifier for the classification as it is a rule-based one and IPSO procedure for focusing on the features and discovering the best threshold values. The Receiver Operating Characteristics (ROC) presents the classification accuracy.