Sentiment Matters: An Analysis of 200 Human-SAV Interactions

2025-10-09

AI Summary

This study investigates the affective mechanisms underlying human interactions with shared autonomous vehicles (SAVs) to enhance user acceptance and perceived service quality. Method: We introduce the first open-source, multimodal dataset specifically designed for SAVs—comprising 2,136 real-world human-vehicle dialogues annotated with corresponding psychological feedback—and propose an integrated analytical framework combining random forest modeling, chord diagram visualization, and zero-shot large language model (LLM)-based sentiment analysis, benchmarked against lexicon-based baselines (e.g., TextBlob). Contribution/Results: Empirical findings identify response sentiment polarity as a critical determinant of user experience; zero-shot LLMs demonstrate superior alignment with users’ subjective affective perceptions compared to supervised baselines, achieving statistically significant improvements in sentiment classification accuracy. This work establishes an empirically grounded foundation and reusable data-method infrastructure for affect-aware human–vehicle interaction design in intelligent mobility systems.

Technology Category

• Cognitive Modeling & Cognitive Systems: Affective Computing
• Humans and AI: Interaction Techniques and Devices
• Intelligent Robots: Human-Robot Interaction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Shared Autonomous Vehicles (SAVs) are likely to become an important part of the transportation system, making effective human-SAV interactions an important area of research. This paper introduces a dataset of 200 human-SAV interactions to further this area of study. We present an open-source human-SAV conversational dataset, comprising both textual data (e.g., 2,136 human-SAV exchanges) and empirical data (e.g., post-interaction survey results on a range of psychological factors). The dataset's utility is demonstrated through two benchmark case studies: First, using random forest modeling and chord diagrams, we identify key predictors of SAV acceptance and perceived service quality, highlighting the critical influence of response sentiment polarity (i.e., perceived positivity). Second, we benchmark the performance of an LLM-based sentiment analysis tool against the traditional lexicon-based TextBlob method. Results indicate that even simple zero-shot LLM prompts more closely align with user-reported sentiment, though limitations remain. This study provides novel insights for designing conversational SAV interfaces and establishes a foundation for further exploration into advanced sentiment modeling, adaptive user interactions, and multimodal conversational systems.

Problem

• Analyzing sentiment's impact on autonomous vehicle acceptance
• Benchmarking LLM versus traditional sentiment analysis methods
• Developing conversational interfaces for shared autonomous vehicles

Innovation

• Dataset with 200 human-SAV conversational interactions
• Random forest modeling identifies key acceptance predictors
• LLM-based sentiment analysis outperforms traditional methods