Marine Robotics Dataset Collection

Reliable evaluation of marine localisation algorithms requires carefully collected real-world datasets. This project involved designing and conducting field experiments using instrumented kayaks equipped with Doppler Velocity Logs (DVL), LiDAR sensors, GPS, and onboard computing systems.

The resulting datasets were collected at Pandan Reservoir in Singapore and used for evaluating probabilistic localisation and mapping algorithms.

Research Contributions

• Designed a marine robotics data collection platform.
• Integrated multiple navigation sensors.
• Produced real-world datasets for algorithm validation.
• Supported future marine autonomy research.

Demonstration Video

Future Directions

Current marine robotics research extends these datasets towards autonomous environmental monitoring and long-term marine autonomy.