A fleet of "robots" could collect garbage or self-assemble in floating structures in the many canals of Amsterdam – ScienceDaily



[ad_1]

The city of Amsterdam envisions a future where autonomous fleets of boats will criss-cross its many canals to carry goods and passengers, collect garbage or self-assemble to form floating bridges and scenes. To realize this vision, MIT researchers have equipped their fleet of robotic boats, developed as part of an ongoing project, with new features that allow them to focus and attach to each other and to try in case failure.

About a quarter of Amsterdam's surface is water, with 165 canals along busy streets. Several years ago, MIT and the Institute of Advanced Metropolitan Solutions of Amsterdam (AMS Institute) collaborated on the "Roboat" project. The idea is to build a fleet of autonomous robotic boats – rectangular hulls equipped with sensors, thrusters, microcontrollers, GPS modules, cameras and other equipment – offering intelligent mobility on the water to reduce congestion in the busy streets of the city.

One of the objectives of the project is to create Roboat units that provide on-demand transport on the waterways. Another goal is to use Roboat units to automatically build "instant" structures, such as foot bridges, performance stages or even food markets. The structures could then automatically disassemble at specific times and become target structures for different activities. In addition, the robots could be used as agile sensors to collect data on the city 's infrastructure, air quality and water, among others.

In 2016, MIT researchers tested a prototype Roboat that crisscrossed the Amsterdam canals, moving forward, backward and sideways along a preprogrammed course. Last year, researchers developed more economical and agile versions, on a scale of one quarter, printed in 3D, low cost, and equipped with advanced algorithms trajectory tracking.

In an article presented at the International Conference on Robotics and Automation, researchers describe robot units that can now identify and connect to docking stations. Control algorithms guide robots to the target, where they automatically connect to a custom locking mechanism with millimeter precision. In addition, Roboat sees it when it has missed the connection, saves and tries again.

The researchers tested the locking technique in a pool at MIT and the Charles River, where the water is more agitated. In both cases, Roboat units were generally able to connect in about 10 seconds, from about 1 meter, or after several unsuccessful attempts. In Amsterdam, the system could be particularly useful for day-to-day garbage collection. Roboat units could bypass a channel, locate and lock platforms containing garbage cans and return them to collection facilities.

"In Amsterdam, canals used to be used for transport and other uses of roads, roads near canals are now very crowded – and are subject to noise and pollution – so the city wants add more features to the channels ", The first author, Luis Mateos, is a graduate student from the Department of Urban Studies and Planning (DUSP) and a researcher from the MIT Senseable City Lab. "Self-driving technologies can save time, money and money, and improve the city in the future."

"The goal is to use robots to give life to new capabilities on the water," adds co-author Daniela Rus, director of the Computer and Computer Lab. Artificial Intelligence (CSAIL) and Professors Andrew and Erna Viterbi of Electrical and Computer Engineering. Science. "The new locking mechanism is very important for creating contextual structures.Roboat does not need locking for standalone water transport, but you need the lock to create any." which structure, whether mobile or fixed. "

Wei Wang, joint post-doctoral fellow at CSAIL and Senseable City Lab; Banti Gheneti, postgraduate student in the Department of Electrical and Computer Engineering; Fabio Duarte, researcher at DUSP and Senseable City Lab; and Carlo Ratti, director of the Senseable City Lab, principal investigator and professor of DUSP.

To make the connection

Each Robot is equipped with locking mechanisms, including ball-and-socket components, at the front, rear and on the sides. The ball component looks like a badminton shuttlecock – a cone-shaped rubber body with a metal bullet at the end. The component of the socket is a large funnel that guides the component of the ball into a receiver. Inside the funnel, a laser beam acts as a security system that detects when the bullet enters the receiver. This activates a three-armed mechanism that closes and captures the ball, while sending a feedback signal to both Roboats to indicate that the connection is complete.

With respect to software, Roboats use custom computer vision and control techniques. Each robot has a LIDAR system and a camera, which allows them to move independently from one point to another of the channels. Each docking station – usually a stationary robot – carries a sheet of paper with an augmented reality tag, called AprilTag, that looks like a simplified QR code. Commonly used for robotic applications, the AprilTags allow robots to detect and calculate their exact 3D position and orientation relative to the tag.

The AprilTags and cameras are located in the same places in the center of the Roboats. When a travel robot is about one or two meters from the AprilTag, it calculates its position and orientation relative to the label. Generally, this would generate a 3D map for the boat's movements, including roll, pitch and yaw (left and right). But an algorithm removes everything except the lace. The result is an easy-to-calculate 2D plane that measures the distance of the robotic camera and the left and right distance of the beacon. Using this information, the Roboat moves towards the beacon. By keeping the camera and tag perfectly aligned, the Roboat is able to connect accurately.

The funnel compensates for any misalignment of the robot's steering (swinging up and down) and lifting (vertical up and down) because the channel's waves are relatively small. If, however, the Roboat exceeds its calculated distance and does not receive a feedback signal from the laser beam, it knows that it has failed. "In difficult waters, lifeboats on the current scale of one quarter are sometimes not powerful enough to withstand gusts of wind or strong currents," said Mateos. "A logical component of the robot says," You missed, so save, recalculate your position and try again. "

Future iterations

Researchers are now designing robots that are about four times the size of current iterations, so they are more stable on the water. Mateos is also working on a funnel update featuring tentacle-like rubber clips that tighten around the spit – like a squid grabbing its prey. This could help give more control to the robot units when, say, they tow platforms or other robots through narrow channels.

In the works is also a system that displays the AprilTags on an LCD monitor that modifies the codes to signal multiple Roboat units to assemble in a given order. At the beginning, all Roboat units will receive a code allowing them to stay exactly one meter away. Then the code changes to direct the first robot to lock. Then, the screen changes the code so that the next Roboat locks, and so on. "It's like the phone game.The changing code transmits a message to one robot at a time, and this message tells them what to do," says Mateos.

The research was funded by the AMS Institute and the city of Amsterdam.

[ad_2]

Source link