Research Title

Overall Project Title

• Development of Mobility and Connectivity Platform for Autonomous Delivery

Sub-project Title

• Development of Integrated Control System for Multi-mobility and Autonomous Delivery Service Demonstration

Sponsor

• Ministry of Trade, Industry and Energy 

Participating Institutions

• KATECH (lead), Yonsei University, Seoul National University, Kakaomobility, Minisoft, MondrianAI etc

Research Period

• Total Period: 2023. 04. 01 - 2026. 12. 31

• Participation Period: 2024. 06. 01 - Present

Role

• Practical Manager (Contribution: 50%)     * PI: Prof. Yong Gu Ji

• Conducted as a Ph.D. researcher at Yonsei University

Keywords

• eHMI, AV–Pedestrian Communication, Pedestrian crossing decision

Research Goal

• The goal of this project is to ensure the smooth coexistence of delivery robots and pedestrians in shared spaces (such as indoor areas, sidewalks, and parks) by minimizing pedestrians’ discomfort, perceived risk, and unnecessary hesitation or pauses during interactions. 

Research Introduction

To achieve this goal, we attach an external Human–Machine Interface (eHMI) to delivery robots (mobility systems) to facilitate interaction with pedestrians.

Our current approach focuses on movement-based eHMI, which displays only the robot’s physical motion states.

Unlike most existing eHMI designs, this approach avoids directive or advisory messages such as “Walk,” “Do not Cross,” or “Safe to cross” which raise ethical and legal concerns.

Instead, the movement-based eHMI communicates only the robot’s own state, such as “Driving” or “Decelerating” leaving the final decision to the pedestrian.

This design is grounded in prior research showing that pedestrians ultimately rely on their own judgment based on vehicle movement and past experiences, rather than depending solely on eHMI messages.

To evaluate this approach, we developed a virtual environment in Unity and conducted interaction experiments with pedestrians. Participants observed the movement-based eHMI and made crossing decisions, measured through Crossing Decision Time (CDT) as a key indicator to validate its effectiveness.

As a next step, we are exploring path-based eHMI.

This design communicates the robot’s intended trajectory, for example by projecting arrows onto the road surface, displaying turn directions on a screen, or using a light band to indicate turning movements.

Research Achievements

• Journal

  • Jongwoo Park, Suhwan Jung, Nakyung Kim, Taewoong Kim & Yong Gu Ji. (2025). "Effects of Autonomous  Vehicle Braking Behavior and Kinematic-Based External Human–Machine Interface on Pedestrian Crossing Decisions" (In Preparation) 

• Conference

  • Park, J., Kim, N., Park, S., Park, G., Jung, S., & Ji, Y. G. (2025). I’m Going to Stop, Will You Cross?: The Impact of Vehicle’s eHMI on Pedestrian’s Decision-Making. In International Conference on Human-Computer Interaction (pp. 181-193). Springer, Cham. (Gothenburg, Sweden) - Oral [PDF]