Does training improve users' mental models about adaptive cruise control?

Hungund, Apoorva; Pai, Ganesh; Pradhan, Anuj K. · 2024 · Crossref

DOI: 10.55329/aqze5695

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 investigates whether and how driver training improves mental models regarding Adaptive Cruise Control (ACC), a common Advanced Driver Assistance System (ADAS). The research is motivated by evidence that drivers often possess incomplete or inaccurate mental models of vehicle automation, leading to mode confusion, miscalibrated trust, and potential safety errors. While prior literature suggests training can enhance driver knowledge, results have been mixed, and the specific impact of different training modalities on mental models and trust remains under-explored. The study aims to determine if training improves ACC mental models, whether visualization-based training is more effective than text-based methods, and if improved mental models correlate with increased trust in the system. The researchers conducted an online, between-subject experiment with 36 licensed drivers who were naïve to ACC. Participants were randomly assigned to one of three groups: Text-Based Training (streamlined owner’s manual content), System Visualization Training (state diagrams illustrating ACC states and transitions supplemented by text), or a Sham Control group (training on unrelated ADAS features). Mental models were measured using the Completeness and Accuracy of Mental Models Survey (CAMMS), which assessed general knowledge (completeness) and specific operational nuances (accuracy) before and after training. Trust was measured using a standardized Trust Survey. Statistical analyses included ANCOVA to compare post-training scores while controlling for pre-training baselines, and Pearson’s correlation tests to examine relationships between mental model scores and trust. The results indicated that training generally improved drivers’ mental models, particularly regarding the accuracy of system limitations and operational parameters. Pre-training accuracy scores were largely incorrect, whereas post-training scores shifted significantly toward correctness across all groups. The text-based group achieved the highest post-training overall mental model scores, outperforming both the visualization and control groups; however, these differences were not statistically significant. ANCOVA revealed no main effect of training method on post-training overall mental model scores after adjusting for pre-training levels. Furthermore, the study found no significant correlation between post-training mental model scores and overall trust scores, suggesting that improved knowledge did not necessarily translate to calibrated trust in this context. The findings provide evidence that training can enhance users’ mental models of ACC technology, correcting misconceptions about system capabilities and limitations. However, the study did not find that visualization-based training was superior to text-based methods, nor did it establish a link between improved mental models and driver trust. These results imply that while training is a viable strategy for improving driver knowledge of ADAS, the format of the training may not be the primary determinant of learning outcomes. Additionally, the lack of correlation with trust suggests that improving mental models alone may not be sufficient to calibrate driver trust, highlighting the need for further research into how training influences both cognitive understanding and affective responses to automated systems.

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-nvidia on 2026-08-10; verification: verified.

Topics

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

• situational awareness
• trust calibration
• mental model of traffic
• mode awareness
• automation surprise
• automation

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).

• Applied Guidance: countermeasure evaluation
• Empirical Findings: self report data
• Theoretical Contribution: computational model