Most people consider the coast to be the edge of the city. A team of MIT researchers envisions it as a dynamic, Lego-like construction site.
Their new system, called “Floatform,” is a swarm of small square robotic boats that assemble themselves into larger formations on water, break up, and reassemble to form something new with minimal human direction.
Each robot, about the size of a 21 centimeter square dinner plate, is a self-contained vessel with its own thrusters, sensors and magnetic latches. Together, they point to a future in which floating infrastructure could become more adaptive: a floating platform after an emergency, a market on a canal, or a platform that pops up for a festival and dissolves when the crowd goes home.
“Our FloatForm projects envision a future where the coast will become a programmable extension of the city, where autonomous boats can self-organize into bridges, platforms, and other useful structures on demand,” says Daniela Ruse, Panasonic Professor of Electrical Engineering and Computer Science at MIT and director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL). “Such distributed robotics open up new possibilities for mobility, emergency response, public space, and infrastructure on water.”
“With Floatform, we are essentially turning static water surfaces into dynamic, programmable spaces,” says Wei Wang, lead author of a new paper on the project and a former MIT research scientist who now leads the Marine Robotics Lab at the University of Wisconsin at Madison. “Imagine an urban environment where public space is not fixed, but can expand, contract or reconfigure autonomously on demand.”
“We see it as building infrastructure on water, using modular systems to build a larger system,” says Alejandro Gonzalez-Garcia, a former researcher at MIT CSAIL and the Senseable Cities Lab. “If there’s an emergency, you can build a new bridge to ease traffic in the city. Or you can build floating markets and floating stages. If you want a more livable city, you might also want to use water.”
Open-access work, published today nature communicationComes from the laboratories of Roush and Carlo Ratti, professor of the practice of urban technologies and planning at MIT and director of the Senseable City Lab, and evolves from their joint project Robot with the Amsterdam Institute for Advanced Metropolitan Solutions, which places full-sized autonomous ships on Amsterdam’s canals. Those canals once carried the city’s goods; Today, they mostly carry tourists.
“We explored whether canals could be used for waste collection, or for transportation, to unload some of the stress on the streets back into the water,” says Niklas Hagemann, an MIT graduate student in architecture, CSAIL affiliate, and former Sensible City Lab researcher, who has worked on the project from its early stages. “As urban areas become denser, can you expand public space on water that is currently underutilized?”
Floatform shrinks that vision down to tabletop scale to answer a tough question: How do you get dozens, and eventually thousands, of floating robots to organize themselves?
Lessons from the Ant Fleet
The team found the answer in biology. The fire ants escape the flood by tying their bodies into living boats, and no leader choreographs the gathering. Each ant follows simple local rules, and a flexible structure emerges.
“Each ant is a free agent,” says Gonzalez-Garcia. “We wanted each robot to have its own capabilities, in the same way that colonies of ants form a raft.”
Most existing self-assembling robot systems, in water and elsewhere, rely on a central computer directing every movement. This approach is sensitive to single points of failure and scales poorly: the planning math balloons as robots are added, and the swarm must assemble sequentially, while most robots remain idle while waiting their turn. Floatform overturns the balance. A lightweight central planner steps only sparingly, providing each robot with a final state to correct the mesh to, a level of geometric precision that purely distributed methods struggle to guarantee. Everything else, including navigating toward a target shape, avoiding collisions, and adapting to disturbances, is handled by the robots themselves, which coordinate by exchanging positions with their nearest neighbors. The whole herd moves together.
That similarity is what sets the works apart. The planning complexity of the floatform approach depends only on the robot’s local neighbors, not on the total size of the swarm. “What we’re trying to do is have minimal central intervention, and they will all move together at the same time,” says Gonzalez-Garcia.
In experiments at MIT, a fleet of eight robots repeatedly assembled from random locations into a target shape, stuck into a rigid structure, disassembled on command, reassembled in a new configuration, and then moved across the pool as a single ship, with each run taking four to eight minutes. In that last mode, called mass transportation, a planner charts a trajectory for the entire structure and each robot calculates its contribution. “Every robot becomes an actuator,” explains Gonzalez-Garcia. The simulation showed the framework scaling smoothly up to a swarm of 64.
“The beauty of this massively decentralized approach is that the computation doesn’t bog down as the swarm grows,” says Wang. “Whether you’re working with eight boats or 80 boats, the entire fleet coordinates and moves together. Because the overall assembly time does not increase significantly in principle, the system remains highly scalable.”
There is also a physical benefit to living together. “If you have waves or currents, our boats join together like an ants’ boat and become more stable,” says Hageman.
an origami handshake
The robots connect via a latching mechanism completely hidden inside each hull. A single servo motor in the center drives an origami-inspired auxetic structure, a geometry that contracts equally in all directions at the same time, pulling inward to release permanent magnets all around, or pushing them outward to capture neighbors at 10 to 15 centimeter intervals. The magnets are arranged with alternating poles, so the boats click into the neat square mesh reliably.
The beautiful part is what the mechanism does not: consume (much) electricity. A 3D-printed gearbox holds the latch in any position when the motor is stopped. “It uses energy to latch and latch, but between those positions, it doesn’t use any energy,” Hageman says. For infrastructure that can retain configuration for hours, this matters. “Because robots are so small, you can only have so big a battery,” says Gonzalez-Garcia. “If they use less energy on latching on, they can use more energy on computation, or actually on walking.”
It took some neat engineering to get there. Four miniature thrusters arranged in an “X” give each robot omnidirectional motion, including spinning in place, but they pack large forces relative to the robot’s small inertia, making early prototypes slippery and prone to aggressive spins at low speeds. The team added static wings to increase hydrodynamic drag and tuned the controllers to remain robust across all robots, which are never the same at this scale. The magnets presented their own problem: They held on so well that de-latching sometimes required the robots to free themselves.
from pond to canal
In 10 trials, the system completed its mission without human intervention 90 percent of the time with four robots and 70 percent of the time with eight robots. When things went wrong, the architecture showed its resilience: a robot that briefly lost its balance could rejoin the structure on its own, without stopping the entire swarm, and robots stuck in formation impasses learned to free themselves and try again.
Going from a controlled indoor tank to a real canal or harbor will require more than confidence. “There is always a relationship between the size of the boat and the magnitude of the disturbance it can handle,” says Gonzalez-Garcia. “These boats are very small, so they can’t operate in very rough waters.” Increasing scaling would mean strengthening the latches, possibly with mechanical interlocking such as those used in full-sized robots, and trading the laboratory’s ultrasonic indoor position for GPS or vision-based sensing. Helpfully, the coordination algorithm was designed to be sensor-agnostic: swap sensors, keep the logic.
The team envisions applications beyond city canals, from building floating platforms for offshore inspection and maintenance to adaptive sensor networks for studying migratory species to reconfigurable docking stations for emergency response in hard-to-reach areas. There is also the possibility of offshore and remote operations, from temporary construction platforms to environmental monitoring and scientific expeditions.
And the geography is very open. “Venice, the Netherlands, Belgium, the fjords and lakes of Norway, really any city with a river can take advantage of this,” says Gonzalez-Garcia. “The project makes use of places where water is already important, but it also raises the question: where else could water be used for something else?”
“This is an exciting step toward realizing distributed collective behavior on water,” says Steven Cerone, assistant professor at the University of Michigan, who was not involved in the research. “Combination, self-reorientation and collective motion are difficult enough in dry environments, but achieving these behaviors in a predominantly distributed fashion on water represents a serious additional challenge, and this team has credibly overcome it. By shifting the computational burden onto the robot, they have created a more flexible system that may enable such robot groups to be deployed in open water environments for search tasks, environmental monitoring and reconfigurable marine infrastructure in the near future.”
Gonzalez-Garcia, Hagemann and Wang co-wrote the paper with senior author Ratti, who is also a professor at Politecnico di Milano and Russa. Gonzalez-Garcia is additionally affiliated with the MECO Research Team at KU Leuven. The research was supported by a grant from the Amsterdam Institute for Advanced Metropolitan Solutions, with additional support from the University of Wisconsin at Madison. The team thanks MIT Sea Grant and Professor Michael Triantafillou for providing the test tanks.
