CSIRO researchers offer a glimpse into the future of robots


An artist's impression of an amphibian robot based on the ocean, the coast or the river. He would travel in water like an eel, but would have legs to crawl and climb. Photo: CSIRO.

Researchers at Australia's national science agency, CSIRO, offered a glimpse of what the robots of the future would look like.

In an article published in Nature Machine IntelligenceCSIRO's Future Science Platform for Integrated Active Matter (AIM FSP) indicates that robots may soon be taking their evolutionary engineering tips.

This concept, known as Multilevel Evolution (MLE), argues that today's robots fight in complex, unstructured environments because they are not specialized enough and must emulate the incredibly diverse adaptation that animals have undergone to survive in their environment.

Earlier in January, lead author of the study, Dr. David Howard, said that evolution does not care about the appearance of something.

READ: New course UNSW prepares people to work in robotics and AI world

"He researches a much wider design space and presents effective solutions that would not be immediately obvious to a human designer.

"An animal like a manta ray or a kangaroo may seem unusual to human eyes, but it's perfectly calibrated for its environment," Howard said.

The article argues that in just 20 years, cutting-edge technologies such as discovery and characterization of high productivity materials, advanced manufacturing and artificial intelligence may allow robots to be designed from the molecular level to perform their mission in extremely challenging circumstances.

Natural evolution-based algorithms would automatically design robots combining a variety of materials, components, sensors, and behaviors.

Advanced computer-based modeling could then quickly test prototypes in simulated "real-world" scenarios to decide which works best.

The end result would be simple, small, highly integrated, highly specialized and highly economical precision robots designed for your task, environment and terrain. They adapt on their own and automatically improve their performance.

An example would be a robot designed for basic environmental monitoring in extreme environments.

It would need to traverse difficult terrain, collect data and eventually degrade completely so as not to pollute the environment.

MLE's approach to designing the robot would depend entirely on terrain, climate, and other factors.

A robot designed to work in the Sahara desert would have to use materials that survive heat, sand and dust. It could be solar powered, sliding through sand dunes and using severe UV light as a trigger to eventually degrade.

The thick, low vegetation of the Amazon would be a totally different challenge.

A robot designed for this environment can crawl around fallen trees and trunks, be fed by biomass, like vegetable matter that covers the jungle floor, and degrade with moisture.

In both cases, the MLE would automatically select the appropriate materials and components in a high performance robot design, based on how well the robot performs a given task.

An infinitely more scalable process than the current approaches that require teams of engineers to design just one robot.


Source link