MIT and Amsterdam jointly develop unmanned ships

2024-03-15

MIT has been collaborating with the Amsterdam Advanced City Solutions Institute to develop autonomous ships. The final project is a full-size, fully automated robot driven vessel, ready to be deployed into the canal in Amsterdam, capable of carrying multiple passengers.

Initially, the MIT project began in a pool at the university, with a small-scale prototype. In 2020, MIT produced a half scale model, measuring two meters in length, to showcase its autonomous navigation skills. Now, the project has produced a pair of full-size robot autonomous boats, which can carry five people comfortably on the canals of Amsterdam. However, carrying passengers is not the only use provided by robots for autonomous navigation of ships.

They are also designed to collect garbage floating on the water surface and transport goods. In addition, when providing on-demand infrastructure, this type of robot autonomous boat can gather together and overlap with each other, forming a floating platform for countless purposes. The robot's autonomous boat is powered by electricity and charged wirelessly, capable of running for 10 hours per charge. In addition, the latest version has also improved the close proximity mode, aimed at providing overlap for infrastructure needs.

MIT also provides improved dynamic positioning for fleets, enabling them to navigate safely in water. The control system can also be adjusted based on the number of people on board. The navigation system on the ship uses GPS to automatically select a safe route between two locations. The ship can also continuously scan its environment to avoid obstacles and other objects such as ships, bridges, and pillars.

The combination of LiDAR and cameras arranged around the robot boat enables it to observe the surrounding environment 360 degrees. The designer refers to these sensors and cameras as "perception kits" that can capture unseen objects that ships cannot recognize, which are marked as unknown. Then, researchers can sort out the data collected by the ship at the end of the day and label items that cannot be identified during the ship's operation. This process enables the system to learn new objects in the real world and safely avoid objects that it cannot recognize in real-time. The control algorithm used for autonomous ships is very similar to that used for autonomous vehicle.

Scientists use thin film nanophotonic devices to generate "ultra wideband" entangled photo