A driver authentication system integrated to stress-level determination for driving safety

Abstract

This paper analyzes the effects of electrocardiogram, electromyogram, respiration, hand and foot galvanic skin response signals on biometric identification using the features extracted through a similar process, which the author has previously introduced for drivers' distress recognition. This process consists of time-domain statistical feature extraction and a subspace-based feature selection. The utilization of subspace features is observed to increase the identification accuracy and decrease the execution time. Besides, the foot galvanic skin response signal is found to be more sensitive to identification among other signals, and increased stress level causes increased false negativity and hence decreased accuracy. These outcomes show that driver identification can be achieved by a process similar to the previously proposed for distress recognition, even using just one wearable technology without the need for any other equipment, and thus give the possibility to realize driver-customized infotainment and vehicle security applications using a single system.