Driver Re-Engagement with AutonoMy (DREAM) – An Innovative Measure of Cognitive Function in the Human-Autonomy Partnership

Abstract

Autonomous vehicles present a significant opportunity for the automotive industry to drastically improve road safety and efficiency whilst providing economic benefits to manufacturers and end users. However, the legal and ethical ramifications of simply removing the driver from the control loop means that autonomous vehicle technologies must be slowly integrated, with a key focus on the human-autonomy partnership (HAP). This means that the design and development of these technologies must account for the interactions between the human driver and autonomous system. The inevitability that, at some point during the journey the primary driving responsibilities will shift between the driver and vehicle means a deeper understanding of the impact these interactions have on the driver is integral in ensuring the HAP functions safely and effectively. Assessment of the human factors (HF) associated with human-autonomy interaction may combine several subjective, performance-based, and physiological measurement techniques to provide relevant information for system design. However, the complexity of the HAP means that a deeper understanding of the cognitive impact of these interactions are imperative in ensuring that they are integrated safely. There is a growing interest from HF researchers within the domain around the use of innovative, portable neuroimaging devices to assess cognitive function in real-world driving situations and provide valuable information on the HAP. The aim of this research was to explore the potential for a novel measure of cortical activity called functional Near Infrared Spectroscopy (fNIRS) to assess driver cognition in a real-world driving environment, with a key focus on validating fNIRS as a reliable measure of cognitive function within the HAP. This research employed subjective, objective, and physiological techniques to help validate fNIRS, through a series of studies designed to assess human autonomy interaction and map regions of the prefrontal cortex associated with different aspects of driver cognition in both non-autonomous and autonomous scenarios. By creating a cortical map of cognitive functions within the HAP, this information can be used to inform the design of autonomous vehicle technologies such as the human-machine interface (HMI). The development of robust methodologies helped to validate fNIRS as a reliable measure of cognitive function against established HF assessment techniques. Findings from the studies within this thesis demonstrate the importance of understanding the underlying cognitive processes involved in human-autonomy interaction and present a significant opportunity to apply fNIRS as an innovative measure of driver cognition. In addition, the development of a robust, validated, and real-world methodology detailed in this thesis demonstrates the potential for fNIRS to inform the HAP and presents an exciting opportunity for researchers and manufacturers to apply these measures to the development of safer vehicle technologies. Indeed, the importance of understanding the impact of human autonomy interaction presents several challenges for researchers and manufacturers that will need to integrate innovative measures of cognitive function into the HF “toolkit” to ensure the development and integration of a safe and effective HAP.

Date of Award	Jan 2022
Original language	English
Awarding Institution	Coventry University
Sponsors	HORIBA MIRA Ltd.
Supervisor	Graham Shelton-Rayner (Supervisor), Dale Richards (Supervisor) & Kevin Vincent (Supervisor)