Driver Behavior State Recognition based on Silence Removal Speech

Numerous researches have linked driver behavior to the cause of accident and some studies are concentrated into different input providing practical preventive measures. Nonetheless speech has been found to be a suitable input source in understanding and analyzing driver's behavior state due to the underlying emotional information when the driver speaks and such changes can be measured. However, the massive amount of driving speech data may hinder optimal performance of processing and analyzing the data due to the computational complexity and time constraint. This paper presents a silence removal approach using Short Term Energy (STE) and Zero Crossing Rate (ZCR) prior to extracting the relevant features in order to reduce the computational time in a vehicular environment. Mel Frequency Cepstral Coefficient (MFCC) feature extraction method coupled with Multi Layer Perceptron (MLP) classifier are employed to get the driver behavior state recognition performance. Experimental results demonstrated that the proposed approach is able to obtain comparable performance with accuracy ranging between 58.7% and 76.6% to differentiate four driver behavior states, namely; talking through cell telephone phone, out-burst laughing, sleepy and normal driving. It is envisages that such engine can be extended for a more comprehensive driver behavior identification system that may acts as an embedded warning system for sleepy driver.