A group of researchers from Colorado State University (CSU) has developed flexible robots that resemble an elephant's trunk in shape and mobility. The development is described in the journal Nature Machine Intelligence. In the future, such machines could help in rescue operations, manufacturing or space missions.
Unlike most modern robots, which are assembled from rigid blocks, the new design uses the tensegrity principle - tensional integrity. In it, rigid rods are held together not by welding or bolts, but by a network of tensioned cables. Thanks to this, the robot is able to bend, compress and change shape. It has no traditional joints, but at both ends of its tubular body there is a pair of gripper claws. With them, the robot can latch onto objects or connect to other robots.
"Usually a robot's physical structure is fixed, but our approach allows flexible systems to adapt their shape to the environment and changing tasks - to compress, stretch or grasp," explained the project leader, Jiannguo Zhao, an associate professor in the Department of Mechanical Engineering. According to him, the robots are modular: each works as a building block, on its own or in a bundle with others, forming larger and more functional structures.
The double claws make it possible to do without magnets, which sometimes fail during connection. Zhao noted that the combination of grippers and a flexible body opens up many possibilities: the robots can wrap around and lift objects of various sizes, crawl over complex terrain like a snake, and also connect into rings, chains and branching structures.
The first author of the paper, doctoral student Mahmoud Hasan Saikot, refined the original design created by Sidney Spiegel. The team faced the task of selecting motors small enough for various movements, yet powerful and reliable during prolonged operation. The researchers spent a lot of time testing configurations and tested the robots not only in the laboratory but also in nature.
Currently the systems are controlled by humans. However, it has already been shown that two robots can independently find each other and connect. In the future, the team hopes to create semi-autonomous groups of dozens or even hundreds of such machines that will reach the desired location and perform tasks with minimal human intervention. There are also plans to increase the speed of movement.
In addition to Zhao, Saikot and Spiegel, the authors of the paper are listed as Sudira Akalanka Kariyawasam, Andrew Stefka and Josh Krisler.