TAR UMT Team, Civility Circuits at the Global Competition on Design for Futures (GCD4F) 2026 Southeast Asia Regional Finals

Project title: Socially Aware Navigation of Robots and Drones

1. Introduction

The 9th Global Competition on Design for Futures (GCD4F) is an annual international event co-organised by Beijing Normal University and the UNESCO Institute for Information Technologies in Education (UNESCO IITE), and is a flagship initiative of the World Digital Education Alliance (WDEA). University students are challenged to develop AI-driven solutions to real-world problems aligned with the United Nations Sustainable Development Goals (SDGs) under three themes: AI for Science, AI for Education and AI for Society. The Southeast Asia Regional Finals were held on 20–21 July 2026 at Menara OUM, Malaysia, and featured 17 finalist teams from Malaysia, China and Vietnam.

2. The Team and Project

The TAR UMT team, Civility Circuits, had four members from the BSc in Data Science programme: team leader Wong Wai Luo, and team members Wong Wai Ming, Chin Yong Zhao and Toh Ming Yang. The project idea originated from the Machine Learning subject they took, with the main goal of training a robot, using machine learning algorithms, to move through a human environment with socially acceptable behaviour while still reaching its goal.

3. The Proposed Solutions

Congratulations to the team for successfully advancing to and competed in the Southeast Asia Regional Finalists round. The proposed solutions made four key contributions:

• Solving the freezing robot problem: a dynamic reward system penalises the robot for staying in the same position, or when its previous and next positions are similar.

• Respecting human social space: a proxemic zone is built around each person, and a high penalty is applied to any risky route that could lead to a collision with a human.

• Ranking the best decisions: a function ranks the penalty-reward ratio of the possible actions so that the robot always chooses the safest one.

• Finding the best route to the goal: the robot predicts a human’s next movement from their momentum and their distance to the robot, and plans its route accordingly.