Research on Multimodal Human-Machine Interface for Takeover Request of Automated Vehicles

WANG, Junfeng; WANG, Yue; CUI, Yin · 2025 · Crossref

DOI: 10.7307/ptt.v37i5.958

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 the effectiveness of multimodal human-machine interfaces (HMI) for takeover requests (TOR) in Level 3 automated vehicles, specifically addressing the safety risks associated with drivers engaged in non-driving related tasks (NDRTs). As automation liberates driver attention, engagement in NDRTs can impair situation awareness and delay response times during critical handovers. While prior research has largely focused on unimodal prompts, this work evaluates four multimodal combinations—visual-auditory (V-A), auditory-tactile (A-T), visual-tactile (V-T), and visual-auditory-tactile (V-A-T)—to determine their impact on takeover success, time, and quality under medium-to-high cognitive loads. The researchers conducted a driving simulator experiment with 30 participants who performed NDRTs (watching video content) while the vehicle operated in automated mode. The experimental route included four distinct takeover scenarios, such as pedestrian crossings and sudden lane obstructions, triggered when the time-to-collision reached seven seconds. A Latin square design randomized the presentation of the four prompt types to mitigate sequence effects. The visual prompt consisted of hierarchical pop-up warnings, the auditory prompt used a standardized female voice, and the tactile prompt utilized seat vibration. Performance was evaluated based on takeover success rate, reaction and execution times, and vehicle dynamics including longitudinal deceleration, lateral acceleration, and minimum time-to-collision. The results demonstrated that multimodal interfaces significantly influence takeover performance. The V-A-T prompt yielded the shortest reaction time (0.99 seconds) and the lowest longitudinal deceleration, indicating a smoother and more controlled handover. In contrast, the V-T prompt resulted in the longest reaction time (2.82 seconds), the highest failure rate, and the most severe braking maneuvers, leading to the highest probability of near-collision events. Although V-A and A-T prompts showed similar reaction times, the inclusion of tactile cues in the A-T condition significantly increased longitudinal deceleration compared to V-A, suggesting that tactile vibrations may induce excessive urgency or reflexive braking. Lateral acceleration remained stable across all conditions, indicating that none of the prompts caused significant loss of lateral vehicle control. The study concludes that multimodal prompts, particularly those combining visual, auditory, and tactile channels, enhance driver response speed and takeover quality by providing redundant sensory information that compensates for attentional distraction. However, the specific modality combination matters; tactile-only or visual-tactile combinations proved less effective than those including auditory cues. These findings provide practical guidelines for designing HMI systems in automated vehicles, emphasizing the superiority of tri-modal prompts for safety-critical takeovers while cautioning against the potential for tactile cues to provoke overly aggressive braking responses.

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.

StageOutcomeToolModelPromptAttemptsCompleted
discover success Crossref 1 2026-08-09
archive success canonical_url 1 2026-08-09
extract success pdftotext 4 2026-08-10
clean success clean 2 2026-08-10
chunk success chunk 2 2026-08-10
embed success embed Qwen/Qwen3-Embedding-8B 2 2026-08-10
promote success 1 2026-08-09
summarize success llm qwen3.6-27b-nvidia summ-v5 2 2026-08-10
tag success vector_similarity 17 2026-08-11
verify partial 2 2026-08-10

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified_with_issues.

Topics

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

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).