SAUVC 2025 Autonomous Underwater Vehicle

AUV development for the Singapore AUV Challenge

Developed an Autonomous Underwater Vehicle (AUV) within a tight one-month timeframe and passed the competition video qualification on the first attempt for HKUST.

I participated heavily in the algorithm and hardware design, which included placing the electronics in a waterproof housing for the first time. I also developed a new communication protocol between the AUV and the topside control station, utilizing a brand-new Pixhawk flight controller and an NVIDIA Jetson computer.

For my first AUV project, I designed a robust navigation algorithm for the qualification task that guides a robot through a gate and back to its starting point. Using computer vision, the algorithm tracks thick lines on the pool floor to reach the target. Once it registers a designated visual signal, it executes a U-turn and returns. By relying on deterministic image processing rather than machine learning, the system bypasses common vulnerabilities related to poor training data quality and environmental inconsistencies.

Demo video

Skills Earned: Rapid Development, Hardware Design, Computer vision.