MIT’s Autonomous Roboat II Gets Several Upgrades

Now two meters long, the robotic boat is capable of carrying passengers.

The latest Roboat revision incorporates a new algorithm for SLAM (Simultaneous Localization and Mapping), a nonlinear predictive controller, and a moving horizon estimator, an optimization state estimator. (📷: MIT)

MIT’s CSAIL (Computer Science and Artificial Intelligence Laboratory) has been developing their Roboat autonomous floating vessel for five years now, which is designed to deliver goods and passengers in the city of Amsterdam.

Roboat II navigates autonomously using algorithms similar to those used by self-driving cars, but now adapted for water,” states MIT professor Daniela Rus, director of CSAIL. “We’re developing fleets of Roboats that can deliver people and goods and connect with other Roboats to form a range of autonomous platforms to enable water activities.”

The two-meter boat can ferry a couple people (at a COVID-19 safety distance) utilizing four propellers down Amsterdam’s canals and waterways. The first Roboat prototype could move forward, backward, and laterally using a preprogrammed path along the canals. The second revision emerged three years later and was capable of reconfiguring itself — autonomously connecting/disconnecting with other Roboats in a myriad of different configurations depending on the application.

The latest iteration, Roboat II, has been scaled up to accommodate a pair of passengers to explore transportation applications. The engineers have outfitted the autonomous boat with a new algorithm for SLAM (Simultaneous Localization and Mapping), a model-based optimal controller known as nonlinear model predictive controller, and an optimization-based state estimator, called moving horizon estimation.

The boat works by first getting a passenger pickup task from a user at a specific location. A system coordinator will then assign the task to an unoccupied Roboat that’s closest to the passenger’s position. When the boat arrives to pick up the passenger, it creates a feasible path to the desired drop off location based on current traffic conditions.

The process uses the new SLAM algorithm, LIDAR, and an IMU for localization, pose, and velocity. The controller tracks the reference trajectories issued from the planner and provides updates to the path to avoid obstacles and potential collisions. The latest upgrades to the Roboat II are another step forward for autonomous boats, and the team hopes to refine their platform so that other, non-smart boats can be easily enhanced.

Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles