Detecting transient changes in heart rate and heart rate variability during experimental simulated autonomous driving scenarios can indicate participant arousal and cognitive load, providing valuable insights into the relationship between human and vehicle autonomy. Successfully detecting such parameters unobtrusively may assist these experimental situations as well as naturalistic driver monitoring systems within an autonomous vehicle. However, non-contact sensors must collect reliable and accurate signals. This study aims to compare the in-seat, non-contact Plessey EPIC sensor to the gold standard, contact Biopac sensor. Thirty participants took part in five-minute simulated autonomous vehicle journeys in a city environment and a rural environment, and a five-minute resting condition. To ensure the seat sensor was sensitive to elevated heart rate values, heart rate was also collected following the energetic Harvard Step Test. Lin concordance coefficients and Bland-Altman analyses were employed to assess the level of agreement between the non-contact Plessey EPIC sensor and the contact Biopac sensor for heart rate and heart rate variability. Analyses revealed a high level of agreement (rc > 0.96) between both sensors for one-minute averaged heart rate and five-minute averaged heart rate variability during simulated autonomous driving and rest, and one-minute averaged heart rate following the Harvard Step Test. In addition, the non-contact sensor was sensitive to significant differences during tasks. This proof of principle study demonstrates the feasibility of using the non-contact Plessey EPIC sensor to accurately detect heart rate and heart rate variability during simulated autonomous driving environments.
Keywords: Automated vehicles; Driver monitoring; Heart rate; Heart rate variability; Non-contact recording.
Copyright © 2021 Elsevier B.V. All rights reserved.