At MIT, scientists have presented a model that generates the best shapes for robots based on the parts and terrain where they will be used. The devices are very unusual, but they cope well with movement.
Researchers at MIT presented a way in which a model can automatically generate a robotic body. It is based on the available details and the terrain on which the device will operate. The system also understands what obstacles the robot needs to overcome. The article’s lead author, Allan Zhao, told MIT News that despite the variety of tasks robots are used for, their designs tend to be very similar in shape or design.
Robots often mimic humans, animals, or vehicles’ bodies, with wheels and tracks to move. But researchers note that it is not always the most useful and effective form. Therefore, the only limitations of the RoboGrammar model are the practical ones associated with building robots. For example, in one of the simulations, where the terrain was rough with many turns, the robot’s design looks more like a crocodile. In contrast, in the simulation, the device handles the movement really well.
So far, the system cannot create robots without human participation. But the researchers note that this is the first step towards making the devices more efficient. The team’s next step is to present the prototypes and test them in real life. The scientists added that the system could benefit both engineers and developers of generated video games who quickly and efficiently create crowded environments.
Earlier, scientists from the United States presented an AI-based robot that can rearrange items on a shelf. In the simulator, its efficiency was very high, but it will be about 80% in real life. The system can predict the location of an object, even when only part of it is visible. It also uses Google’s technology, COntext Inference (COCOI), an online contact interface that injects physical objects’ properties into a simple system.