Development of an Evaluation Method for a Driver's Cognitive Workload Using ECG Signal

Abstract

High cognitive workload decreases a driver's ability of judgement and response in traffic situation and could result in a traffic accident. Electrocardiography (ECG) has been used for evaluation of drivers' cognitive workload; however, individual differences in ECG response corresponding to cognitive workload have not been fully considered. The present study developed an evaluation method of individual driver's cognitive workload based on ECG data, and evaluated its usefulness through an experiment in a driving simulator. The evaluation method developed by the present study determined the optimal ECG evaluation condition for individual participant by analysis of area under the receiver operating characteristic curve (AUC) for various conditions (total number of conditions = 144) in terms of four aspects (ECG measure, window span, update rate, and workload level). AUC analysis on the various conditions showed that the optimal ECG evaluation condition for each participant was significantly different. In addition, the optimal ECG evaluation condition could accurately detect changes in cognitive workload for 47% of the total participants (n = 15). The evaluation method proposed in the present study can be utilized in the evaluation of individual driver's cognitive workload for an intelligent vehicle. †

Classification of a Driver's cognitive workload levels using artificial neural network on ECG signals

“…The participants were instructed to drive (primary task) the driving simulator while performing an N-back task (secondary task). The N-back task was to recall the N step's earlier number when an experiment instructor presented a sequence of arbitrary numbers (Hong et al, 2014;Son et al, 2010).…”

“…On the other hand, there is an individual variation in heart response. Many studies have reported that heart responses to tasks show significant differences between individuals (Hong et al, 2014;Lee et al, 2010;Lal and Craig, 2001). First, an effective ECG measure varies noticeably among individuals.…”

Development of a Classification Model for Driver's Drowsiness and Waking Status Using Heart Rate Variability and Respiratory Features

“…Existing studies mainly used EEG and ECG sensors to measure drivers' physiological responses (Hong et al, 2014;Lee et al, 2010;Rogado et al, 2009), which may cause participants' repulsion on the sensor-attached location and the restrictions on participants' actions, due to sensor attachment. The PPG and respiration sensors are considered to be more easily embodied for actual use of a car system, compared to other bio signal sensors.…”