Judging Intelligent Robotics (30.119) at SUTD 🤖

I had the opportunity to support the judging session for the course 30.119 Intelligent Robotics at the Singapore University of Technology and Design (SUTD) on 11 December 2025. Participating as a volunteer with the James Dyson Foundation, I assisted in evaluating the students’ final robotics projects.

The session, held from 11:30 AM to 2:00 PM, marked the culmination of several weeks of hands-on learning, where students designed, built, and deployed autonomous robotic systems using ROS 2. It was a showcase of their technical depth, problem-solving skills, and growing confidence in real-world robotics development.

Course Overview
The Challenge & Presentations
Technology Stack
Judging Criteria
My Experience

Course Overview

The 30.119 Intelligent Robotics course is designed to give students practical exposure to real-world robotic systems, with a strong emphasis on autonomy and software integration.

Key details:

Learning Objective: Build an autonomous robot capable of overcoming a mini challenge, with a focus on ROS 2 capabilities such as navigation, mapping, and exploration.
Students & Teams: 26 students divided into 7 teams.
Student Profile: Term 7 undergraduate students enrolled in 30.119 Intelligent Robotics.
Venue: Dyson–SUTD Innovation Studios
Point of Contact: Asst. Prof. U-Xuan

The Challenge & Presentations

The live demo was split into two parts.

Teams had to demonstrate the navigation capability of a TurtleBot in a structured indoor environment:

Navigate through a sequence of predefined waypoints
Stop navigating when encountering stop signs
Enter different rooms and count the number of intruders
Intruders were represented by photos of people pasted on the walls
Identify and localize intruders, visualizing them in RViz
Return the robot safely to the start pose

Part 2 (Bonus): Autonomous Exploration & Mapping

As a bonus task, teams attempted to:

Autonomously explore the environment
Build a map using exploration strategies
Optimize both map coverage and time taken

Each team was given 10 minutes to complete both the main and bonus tasks.

Technology Stack

The students worked with a solid and industry-relevant robotics stack:

Navigation: ROS 2 Nav2
Perception: YOLO for intruder (person) detection
Exploration & Mapping: Frontier-based exploration using https://github.com/AniArka/Autonomous-Explorer-and-Mapper-ros2-nav2

It was great to see students integrate multiple subsystems—navigation, perception, and mapping—into a cohesive autonomous pipeline.

Judging Criteria

The evaluation focused on both performance and robustness:

Successful navigation through all waypoints
Correct detection and counting of intruders
Accurate localization of intruders in RViz
Ability to return to the start pose
Time taken to complete the tasks

For the bonus round:

Amount of map coverage achieved
Time taken to build the map autonomously

My Experience

The atmosphere during the judging was a mix of nervous energy and excitement. Some teams executed their pipelines smoothly, while others faced last-minute glitches—very much reflective of real-world robotics deployments.

What stood out was the competitive yet enthusiastic spirit. Students were eager to beat each other’s completion times and climb the leaderboard, cheering on successful runs and learning quickly from failures. Despite the pressure of the 10-minute limit, the teams showed strong problem-solving skills and resilience.

Overall, it was a fun and rewarding experience to judge these projects. Sessions like this are incredibly motivating for students and serve as a strong foundation for taking on more advanced robotics courses and research in the future. It was inspiring to see how far they’ve come in building intelligent, autonomous systems within a single course.