Driver Engagement Adrià Roig, Francesco Deiana*, James Jackson, Clara Cabutí, Cristina Periago, Fiona Azcarate, Elena Castro, IDIADA Automotive Technology SA, Spain KEYWORDS – Engagement, distraction, interaction, behaviour, assistance systems. ABSTRACT With the emergence of driver assistance systems, it is crucial to ensure that drivers have a clear understanding of the technology's capabilities and can operate their vehicle safely. This understanding can be influenced by various factors, including vehicle instructions, user interface, warnings, and system control behavior. To comprehensively evaluate safety performance for real-world use and overall system design implementation, it is important to validate the system accounting for these factors. This paper presents a specific test methodology for application in an automotive test facility. It captures pre-use information and driver-vehicle interaction during assisted driving in relation to the user interface and system control performance. Data collection was defined around the quantification of driver engagement with the driving task using subjective metrics to analyse the progressive effects of system use, and objective metrics, which consider the driver's behaviour and ability to respond to an emergency scenario. In a pilot evaluation, using two vehicles with different assisted driving concepts, a between-subjects test was conducted. A sample of 39 naive drivers was recruited and, after a customer-focused description of the system functions, they were instructed to drive on a test track in a continuous highway driving scenario with the longitudinal and lateral driver assistance functions active. This was followed by a critical "cut-out" that required a driver response to avoid an obstacle in the lane. Our results demonstrate that there is variability in how drivers interact with the system during 'normal driving', with subjective measures showing differences in engagement metrics. Objective measures for driver reaction to the critical event also showed differing levels of driver vigilance associated with perceived functionality of individual systems. Due to the automatic braking system of the vehicles, it wasn’t possible to analyze the Time to Collision (TTC) based on the braking time, therefore it is intended to do so in future implementations, as well as setting an additional camera above the brake pedal. Also, it is planned to extend the study to other vehicles and automation systems such as L3. The distinguishing feature that characterizes this paper is the fact that it compares the systems of different vehicles and how they affect the driver. In addition, it allows the systems to be evaluated from a combined subjective and objective point of view. It should be noted that it offers an experimental methodology that can be extrapolated to any type of vehicle and any type of autonomy level. Our approach has implications for system design verification with high replicability, accounting for use by representative drivers. From this test we can extract that concrete results have been obtained that allow us to individualize between the different systems. Besides that, the analysis of subjective and objective metrics related to the mental load perceived by the drivers and the level of confidence, revealed interesting differences between the two types of L2 systems. This experimental test is a step forward in the development of test methods for global assistance system assessment and provides a platform for further refinement and progression of tests. Additionally, it has possible applications in consumer and regulatory testing with representative drivers.
Ms. Cristina Periago, Human Factors, Applus+ IDIADA