ENGAGE project

Limitations and discrepancies between driver behaviour and driving aiding systems

Concerns have arisen regarding driver inattention and inappropriate behaviour when using Level 2 systems, as drivers may not fully understand the system's limitations and continue to be responsible for vehicle control. Therefore, it is crucial for drivers to maintain focus when using these systems.

Features such as Adaptive Cruise Control (AAC), Lane Change Assist (LCA) and Automatic Emergency Braking (AEB), provide the driver with an assisted and comfortable driving experience. The system assists the driver in performing the driving task, but at the same time, they must convey to the driver that they are not driving in their place, and that they must remain focused and attentive.

When implemented correctly, these systems can provide significant benefits for drivers, but if they don't guarantee the right level of engagement, they can be a safety risk by creating false expectations about the vehicle's abilities and overreliance.

Euro NCAP Vision 2030 recognizes driver engagement as a fundamental pillar to ensure the safe deployment of assisted driving systems and considers it a key factor.

New test methodology to assess driver engagement with ADAS systems

To address this challenge, IDIADA Human Factors multidisciplinary team has developed a specific test methodology for evaluating driver engagement in assisted driving scenarios. This methodology captures pre-use information and driver-vehicle interaction during assisted driving, considering factors such as the user interface, warnings, and system control behaviour.

In a pilot evaluation using two vehicles with different assisted driving concepts, IDIADA recruited 39 naive drivers and instructed them to drive on a test track in a continuous highway driving scenario with the longitudinal and lateral driver assistance functions active.

The results of this evaluation demonstrated variability in how drivers interacted with the systems during normal driving, with subjective measurements showing differences in engagement metrics.

Objective measurements for driver reaction to a critical event also showed differing levels of driver vigilance associated with the perceived functionality of individual systems.

Towards a global assessment methodology to test and validate driving assistance systems

This test methodology offers a way to compare users’ driver engagement on the systems of different vehicles and evaluate them from a combined subjective and objective point of view. It can be applied to any type of vehicle and any type of autonomy level, making it a valuable tool for system design verification, consumer testing, and regulatory testing.

Furthermore, this methodology marks progress in the development of assessment methods for global assistance systems and offers a platform for further refinement and advancement of testing procedures. It also holds potential applications in consumer and regulatory testing with representative drivers.

Currently, the study is in a phase of further development. The same methodology will indeed be tested in the dynamic simulator, replicating the experiment conducted on the track with the implementation of additional subjective and objective metrics.