Adaptability to be key to future robots, argues the study



[ad_1]

An artist's impression of an Antarctic-based robot. Image Credit: csiro.au

An artist's impression of an Antarctic-based robot. Image Credit: csiro.au

New research explores the possibility of making robots that emulate evolutionary adaptation so that they are better able to respond to their environments.

The article, authored by researchers from CSIRO's AIM FSP (Future Science Platform of Integrated Active Matter) and published in Nature: Machine Intelligence, argues that robots could be designed to display Multi-Level Evolution (MLE), which would allow them to become more adapted to unstructured and complex environments.

"Evolution does not care what it looks like. It looks for a much broader design space and offers effective solutions that would not be immediately obvious to a human designer," said lead author Dr. David Howard.

"An animal like a stingray or a kangaroo may seem unusual to human eyes, but it is perfectly calibrated for its environment."

Within 20 years, technologies such as the discovery and characterization of high productivity materials, advanced fabrication and artificial intelligence could enable the design of robots from the molecular level, argues the article.

Natural evolution-based algorithms would automatically design robots combining a variety of materials, components, sensors, and behaviors. Advanced computer-based modeling could quickly test prototypes in simulated "real-world" scenarios to decide which works best.

The goal would be to create simple, highly integrated, specialized and cost-effective robots that are precisely responsible for the specific tasks they must perform in certain environments and terrains. They would also adapt on their own and improve their own performance automatically.

"CSIRO is committed to leading scientific thinking and has initiated this collaboration with international researchers to understand the nature of the future in robotics. CSIRO has collaborated with researchers from the University of Vrije in the Netherlands, the University of Lorraine in France and the La Trobe and Monash universities in Australia, "said AIM FSP Director Dr. Danielle Kennedy.

"The future scientific platform program is part of CSIRO's investment in creating future industries for Australia and helps train the next generation of researchers. The focus of AIM is not limited to robots, we are also exploring the future of the food industry, manufacturing, environmental monitoring and industrial design. "

[ad_2]

Source link