With a Little Help From My Car: Sharing Automated Vehicle’s Situation Awareness Reduces Driver-Initiated Automation Disengagement Without Delaying Takeover Response Time

Abstract

Objective

This study investigated the effect of sharing vehicle situation awareness (VSA) on driver takeover behavior in complex urban environments.

Background

As automated vehicles (AV) expand their operational design domain, little is known about driver interactions with driving automation in complex urban settings. Drivers often become either overly reliant on automation or fail to rely on it even when capable, leading to misuse or disuse. Sharing VSA information could enhance drivers’ awareness of the AV system and response when AVs request manual control in complex situations.

Methods

A driving simulator tested sharing VSA information via augmented reality head-up displays (AR HUDs) during takeover scenarios. Participants were assigned to control or experimental groups that received different combinations of VSA elements: perception (object highlighting), comprehension (confidence assessment), and projection (trajectory information). Two urban driving scenarios (parking lane and intersection) were tested.

Results

Sharing VSA information reduced driver-initiated automation disengagement before takeover requests without delaying response times. Perception information alone showed no significant difference from baseline, but adding egocentric projection information significantly reduced driver-initiated overrides, while allocentric projection did not. Adding confidence assessment further enhanced effectiveness. The parking lane scenario was associated with quicker responses, fewer full takeovers, and softer braking.

Conclusion

Specific combinations of VSA information reduced driver-initiated disengagement from automation without compromising response times. The type and presentation of shared information significantly affect human-automation interaction.

Application

These findings can guide the design of AV systems that better support driver-vehicle interaction in complex urban environments.