Supporting Drivers of Partially Automated Cars through an Adaptive Digital In-Car Tutor

Boelhouwer, Anika; van den Beukel, Arie Paul; van der Voort, Mascha C.; Verwey, Willem B.; Martens, Marieke · 2020 · OpenAlex-citations

DOI: 10.3390/info11040185

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of helping drivers safely understand and use complex Advanced Driver Assistance Systems (ADAS) in partially automated vehicles. Drivers often struggle with system capabilities and limitations due to inconsistent naming conventions, lack of dealer training, and poor human-machine interfaces. Traditional training methods, such as driving schools or simulators, are costly and time-consuming. The authors propose a Digital In-Car Tutor (DIT) as a low-cost, situated learning solution that guides drivers through automation features during regular drives using adaptive audio and augmented reality feedback. The researchers conducted a between-subjects driving simulator study with 38 participants divided into an experimental group (DIT) and a control group (Information Brochure, IB). The simulated vehicle featured Level 2 automation, including Adaptive Cruise Control and Lane Keeping. Participants completed three driving sessions involving scenarios where they had to decide whether to engage or disengage automation based on safety. The DIT group received real-time, situation-adaptive tutoring during the first session, while the IB group read a static brochure beforehand. Sessions 2 and 3 occurred immediately after and two weeks later, respectively, without DIT support, to assess retention. Performance was measured by appropriate automation use (correct reliance vs. incorrect reliance/take-over) and take-over quality (time to collision, deceleration, lateral acceleration). Results indicated that the DIT group demonstrated significantly more correct automation use and fewer collisions than the IB group during the first session. Specifically, the DIT effectively reduced inappropriate reliance on automation in critical scenarios, such as those involving obscured pedestrians or missing road markings. While the DIT group also showed better take-over quality initially, statistical significance for vehicle control metrics varied. In subsequent sessions without the tutor, the performance gap narrowed, though the DIT group maintained lower rates of incorrect reliance throughout all sessions. Interestingly, the DIT group exhibited some under-trust (unnecessary disengagement) in the final session. Participants reported high acceptance of the DIT, citing ease of use and usefulness. The study concludes that a Digital In-Car Tutor is a viable, efficient method for supporting drivers in learning to use partially automated systems safely. By providing situated, adaptive feedback, the DIT helps establish accurate mental models of system capabilities and limitations more effectively than static information. This approach offers a scalable alternative to expensive simulator training, potentially enhancing traffic safety by reducing misuse of automation. The findings suggest that integrating adaptive tutoring into vehicle interfaces can facilitate safer driver-automation interaction, particularly during the critical initial learning phase.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• trust calibration
• in vehicle coaching
• automation
• automation surprise

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Methodological Resource: tool software
• Theoretical Contribution: computational model, conceptual framework