Personalized Voice-Based eHMI for Autonomous Personal Mobility Vehicles

Duration: Nov 2023 – Sep 2025
Affiliation: Human Robotics Laboratory, Nara Institute of Science and Technology (NAIST) · Ikoma, Japan

Note: Full write-up coming soon. Related publication: Personality-Driven eHMI for APMVs.

Overview

A three-stage research pipeline for personalizing the eHMI of autonomous personal mobility vehicles (APMVs) based on passenger personality — developed as part of the NAIST MEXT-funded robotics program.

Three-Stage Pipeline

1. Empirical Study — Survey-based study to understand how personality traits influence passenger preferences for robot communication style (tone, speed, verbosity)
2. Causal Discovery & Recommendation — Causal graph modeling (GNN) between personality factors and eHMI preferences; recommendation system to predict preferred voice profile per user
3. Voice Cloning & Personalized eHMI — Neural voice cloning to synthesize personalized robot voices; integration into the APMV eHMI interface

Technologies

Python Voice Cloning TTS Causal Discovery Recommendation System HRI ROS 2