Smiles and Angry Faces vs. Nods and Head Shakes: Facial Expressions at the Service of Autonomous Vehicles

Rouchitsas, Alexandros; Alm, Håkan · 2023 · Crossref

DOI: 10.3390/mti7020010

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 communication gap between autonomous vehicles (AVs) and pedestrians, specifically focusing on how external human–machine interfaces (eHMIs) can convey vehicle intention. As AV drivers will not be engaged in driving, they cannot use traditional cues like eye contact or gestures to negotiate right-of-way. The authors evaluate the efficiency of anthropomorphic eHMIs that use facial expressions to signal whether an AV intends to yield or not yield to a pedestrian. The research compares emotional expressions (smile for yielding; angry face for non-yielding) against conversational expressions (nod for yielding; head shake for non-yielding). The primary aim was to determine which type of expression allows pedestrians to make crossing decisions more efficiently, defined by lower response latency, thereby supporting both traffic safety and flow. The researchers employed a 2 × 2 × 2 within-subject experimental design with independent variables for virtual human character (VHC) gender (male/female), vehicle intention (yielding/non-yielding), and expression type (emotional/conversational). Participants performed a speeded, two-alternative forced-choice crossing intention task in a laboratory setting. They viewed 3D animated sequences of VHCs displaying the specific facial expressions or head movements, presented out of context to isolate the communicative cue from vehicle kinematics. Participants were instructed to respond as quickly and accurately as possible by pressing a button to indicate whether they would cross the street. The study measured both the accuracy and the latency of these decisions to assess the effectiveness and efficiency of the different expressions. The results demonstrated that emotional facial expressions communicated vehicle intention more efficiently than conversational expressions. Specifically, participants exhibited significantly lower response latency when interpreting emotional cues (smiles and angry faces) compared to conversational cues (nods and head shakes). While previous research had established that both types of expressions are highly accurate for inferring intention, this study highlighted that emotional expressions facilitate faster decision-making. The findings suggest that the universal recognition of emotional states allows for quicker mental processing and anticipation of behavior compared to the cognitive interpretation of conversational gestures like nodding or shaking the head. The significance of these findings lies in the design of anthropomorphic eHMIs for autonomous vehicles. The study concludes that emotional expressions are superior for communicating safety-critical intentions in time-sensitive traffic scenarios. By enabling pedestrians to decide more rapidly whether to cross, emotional eHMIs can reduce ambiguity and improve traffic flow without compromising safety. This supports the broader argument for using anthropomorphic elements in AV design to leverage existing social cognition, thereby enhancing public trust and acceptance of autonomous technology. The results provide empirical evidence for prioritizing emotional facial expressions over conversational gestures in the development of external communication systems for AVs.

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_with_issues.

Topics

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

• ehmi external hmi
• vru facing ehmi
• signaling behavior