3D printed Robots adapt themselves to different kinds of terrain


Researchers from the University of Oslo have designed self-instructing robots on 3D-printers, capable of adapting to unforeseen obstacles. In future robots are supposed to operate in hazardous areas such as in deep mines, distant planets, radioactive sites, dangerous landslip areas and on the sea bed beneath the Antartic. These places are too extreme for human involvement. Everything needs to be automatically controlled.

Imagine a robot entering a wrecked nuclear plant. It encounters a staircase that no one had expected. The robot captures the picture, understands the situation and using a 3D printer that is attached to one of its arms prints a new robot or a new part that can pass the stairs. The process is completed without human involvement and without delaying the mission.

The research team has been working on these possibilities for years, and has reportedly developed three generations of self-learning robots. The first generation robot, nicknamed “Henriette,” designed by Professor Mats Hovin learnt itself how to walk and jump over hurdles. When it lost a leg, it had learnt how to manage with just one leg.

The second generation robot was introduced a couple of years later by Maters student Tonnes Nygaard, and was paired with a simulation program that could figure out how the robot’s body should be shaped, how many legs are required, the length of the legs and the standard distance between them. The robot self-designed all its key parts.

The third generation robots are more flexible. Researchers can mention their desired characteristics in the simulation program such as what are the robot’s tasks, how fast it should move, its size and energy consumption.

Thus the simulation program understands the requirements of individual robots and its responsibilities and ran thousands of simulations to come up with the best possible model, which included the shape of the body and number of legs. The sophisticated robots were all created using 3D printer and tested by real-world standards. Check out the video below to see how the robot adapts to changing environment.

Via: Spectrum/Engadget