Decision time? Check out our latest product comparisons

Researcher uses Cyber Rodents to study evolution

By

April 15, 2014

The study used robots known as Cyber Rodents to study mating behaviour over more than 1,00...

The study used robots known as Cyber Rodents to study mating behaviour over more than 1,000 generations (Photo: OIST)

Image Gallery (2 images)

A study has used rodent-like robots to look at the evolutionary development of different mating strategies over an extended period of time. In contrast to direct studies of nature, the observation of robots allows researchers to avoid inherent time-based difficulties of studying evolution, with the results suggesting something a little more complex than the classic one-beats-all natural selection hypothesis.

According to conventional evolutionary theory, a single, optimal phenotype, or mating behavior, should predominate all others, with natural selection taking care of the less efficient strategies. However, in nature we witness a great many populations where this is not the case, and instead see a variety of successful behaviors co-existing. Due to our short life spans, developing an informed theory on why this is the case is decidedly problematic when looking directly at long-term evolutionary cycles in nature.

The study, conducted by Dr. Stefan Elfwing of the Okinawa Institute of Science and Technology, was designed to tackle this long-standing problem of evolutionary theory through the use of robots. Known as Cyber Rodents, the wheeled robots were equipped with cameras for visual detection of energy sources (colored blue) and the rail lamps of other robots (colored green), and infrared communicators for the exchange of genotypes. In biological terms, the Cyber Rodents were hermaphrodites, with all robots in the test able to produce virtual offspring.

During their 288-second life spans, the robots were able to execute two basic tasks – searching for a partner with which to mate, or searching for batteries. The probability of successfully producing offspring was determined by the robot’s internal energy level, thus creating a trade-off between foraging for energy and moving directly to mate.

Dr. Elfwing was able to avoid the inherent time frame difficulties associated with observing evolution in nature, using computer simulation to study over 1,000 generations in each experiment. The results detail the emergence of two distinct behaviors, or phenotypes, within the experiment – the Forager and the Tracker.

The study found that two distinct phenotypes were able to efficiently co-exist (Photo: OIS...

The Forager phenotype would actively search for batteries, only mating when it saw the face of another Cyber Rodent, and never waiting for them to turn around. Conversely, the Tracker would wait for other robots to turn around for mating, with the length of time waited being determined by its current internal energy level.

The experiment was conducted some 70 times with varying results, but it was the experiments where more than one phenotype emerged that successful reproduction rates tended to be highest.

By conducting the experiment with different ratios of phenotypes, Stefan was able to show that the two behavior types could efficiently co-exist within a single population, with the stable ratio being 25 percent Foragers to 75 percent Trackers.

"In this experiment, our robots were hermaphrodites, all robots mate and can produce offspring," said Dr. Elfwing. "In the next stage, we want to see if the robots will take on male and female roles, by taking different risks and costs in reproduction.”

His findings are described in a paper that was recently published in the journal PLOS ONE.

Source: OIST

About the Author
Chris Wood Chris specializes in mobile technology for Gizmag, but also likes to dabble in the latest gaming gadgets. He has a degree in Politics and Ancient History from the University of Exeter, and lives in Gloucestershire, UK. In his spare time you might find him playing music, following a variety of sports or binge watching Game of Thrones.   All articles by Chris Wood
Tags
Post a Comment

Login with your gizmag account:

Or Login with Facebook:


Related Articles
Looking for something? Search our 29,166 articles