MIT has announced that its fleet of robotic boats has been updated so that it can now “shapeshift” by autonomously disconnecting and reassembling into various configurations to form floating structures in Amsterdam’s many canals.

Described as “rectangular hulls equipped with sensors, thrusters, microcontrollers, GPS modules, cameras, and other hardware,” the autonomous boats are being developed by MIT and the Amsterdam Institute for Advanced Metropolitan Solutions (AMS Institute), as part of an ongoing project known as “Roboat.”

Amsterdam’s long-term goal is to have the roboats cruise its 165 winding canals, conducting a variety of tasks such as transporting goods and people, collecting trash, or self-assembling into “pop-up” platforms such as bridges and stages to help relieve congestion on the busy streets of the city. 

MIT’s work on roboats dates back a few years, as in 2016, the university’s researchers tested a roboat prototype that could move forward, backward, and laterally along a preprogrammed path in the canals. Last year, researchers designed low-cost, 3-D-printed, one-quarter scale versions of the boats, which were “more efficient and agile,” MIT says, and were also equipped with advanced trajectory-tracking algorithms.

Earlier this summer, researchers developed an autonomous latching mechanism that allows the boats to target and clasp onto each other, and keep trying if they fail.

In a new paper that was presented during the IEEE International Symposium on Multi-Robot and Multi-Agent Systems, researchers talk about an algorithm that allows the robots to seamlessly reshape themselves as efficiently as possible. The algorithm handles all the planning and tracking that allows groups of roboat units to unlatch from one another in one set configuration, travel a collision-free path, and reattach to their appropriate spot on the new set configuration.

During demonstrations in an MIT pool and in computer simulations, groups of linked roboat units rearranged themselves from straight lines or squares into other configurations, such as rectangles and “L” shapes. While experimental transformations took just a few minutes, MIT notes that more difficult shapeshifts could take more time, depending on the number of moving units and differences between the two shapes.

“We’ve enabled the roboats to now make and break connections with other roboats, with hopes of moving activities on the streets of Amsterdam to the water,” explains Daniela Rus, one of the MIT professors leading the project, and the director of the Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science.

“A set of boats can come together to form linear shapes as pop-up bridges, if we need to send materials or people from one side of a canal to the other. Or, we can create pop-up wider platforms for flower or food markets.”

Researchers had to tackle several challenges during the course of their work, including those related to autonomous planning, tracking, and connecting groups of roboat units. An example is how it would require “complex communication and control techniques that could make movement inefficient and slow” to give each unit unique capabilities to locate each other, agree on how to break apart and reform, and then move around freely. 

The researchers developed two types of units— coordinators and workers—to facilitate smoother operations. To form a single entity known as a “connected-vessel platform” (CVP), one or more workers connect to one coordinator. All coordinator and worker units are equipped with four propellers, a wireless-enabled microcontroller, and several automated latching mechanisms and sensing systems that enable them to link together.

Coordinators are also equipped with GPS for navigation, and an inertial measurement unit (IMU), which computes localization, pose, and velocity, while workers only have actuators that help the CVP steer along a path. Each coordinator is aware of and can wirelessly communicate with all connected workers. Structures make up multiple CVPs, and individual CVPs can latch onto one another to form a larger entity.

During shapeshifting, all connected CVPs in a structure compare the geometric differences between its initial shape and new shape. Each CVP then determines if it stays in the same spot and if it needs to move. Each moving CVP is then assigned a time to disassemble and a new position in the new shape.

According to MIT, each CVP uses a custom trajectory-planning technique to calculate a way to reach its target position without interruption, while maximizing the route for speed. To accomplish this, each CVP “precomputes all collision-free regions around the moving CVP as it rotates and moves away from a stationary one,” MIT explains.

After precomputing those collision-free regions, the CVP then finds the shortest trajectory to its final destination, which still keeps it from hitting the stationary unit. To make the whole trajectory-planning process “very efficient,” optimization techniques are used, as the precomputation takes just over 100 milliseconds to find and refine safe paths. Data from the GPS and IMU is used to allow the coordinator to estimate its pose and velocity at its center of mass, and wirelessly controls all the propellers of each unit and moves into the target location.

During their experiments, the researchers tested three-unit CVPs, made up of one coordinator and two workers, in several different shapeshifting scenarios. Each scenario involved one CVP unlatching from the initial shape and moving and relatching to a target spot around a second CVP.

For instance, three CVPs rearranged themselves from a connected straight line into a straight line connected at front and back, as well as an “L.” During computer simulations, up to 12 roboat units rearranged themselves from, for example, a rectangle into a square or from a solid square into a Z-like shape.

The researchers hope to use the roboats to form into a dynamic “bridge” across a 60-meter canal between the NEMO Science Museum in Amsterdam’s city center and an area that’s under development. The idea, known as RoundAround, is to employ roboats to sail in a continuous circle across the canal, picking up and dropping off passengers at docks and stopping or rerouting when they detect anything in the way. Right now, it takes about 10 minutes to walk around that waterway, but the bridge could potentially cut that time to around two minutes. This is still an “explorative concept,” MIT notes.

“This will be the world’s first bridge comprised of a fleet of autonomous boats,” Ratti says.

“A regular bridge would be super expensive, because you have boats going through, so you’d need to have a mechanical bridge that opens up or a very high bridge. But we can connect two sides of canal [by using] autonomous boats that become dynamic, responsive architecture that float on the water.”

The researchers will continue to develop the roboats to make sure that they can safely hold people, and are capable of operating in all weather conditions, such as heavy rain. They’re also making sure the roboats can effectively connect to the sides of the canals, which can vary greatly in structure and design.