Design & analysis of k-means algorithm for cognitive fatigue detection in vehicular driver using oximetry pulse signal

Most of the fatal injuries and the loss of lives occur due to lack of timely and quick action to be taken by the vehicular drivers. The difficulty in determining the incidence of fatigue-related accidents is due to the difficulty in identifying fatigue as a causal or causative factor in accidents. In most instances, one or more indirect or circumstantial pieces of evidence are used to make the case that fatigue was a factor in the accidents. In this paper we propose an unconventional approach for fatigue detection in vehicular drivers using Oximetry Pulse (OP) signal to detect cognitive fatigue of the driver thereby reducing the loss of the lives and vehicular accidents. This method includes implementation of basic K-means and modified K-means for detection of fatigue state of drivers. Oximetry Pulse signal has been recorded from vehicular drivers for Pre and Post driving states and were processed using various wavelet functions to extract typical set of features. The K-means classifiers were trained and tested for these datasets. Each of the features extracted was treated as single decision making parameter. From the test results it could be found that some of the wavelet features could fetch 100 % classification accuracy with modified K-means while few others with basic K-means classifier.