A New Study by Dario Floreano
Dario Floreano and his team at the Laboratory of Intelligent Systems in Lausanne, Switzerland, have recently used robots to study the evolution of social behaviour. For years, a question that has plagued evolutionary biologists has been the case of how and why altruistic behaviour has evolved in groups. It’s obvious that altruistic behaviour normally implies risk to the actor. In the worst case scenario it can even mean death. So, on the surface, altruism appears to go against the primal force that drives life at large: An individual’s drive to survive and pass on their genes to the next generation.
Evolutionary biologist W.D. Hamilton came up with the theory of kin selection, also known as Hamilton’s rule. This theory demonstrates that altruistic behaviour towards a relative does not contradict an individual’s drive to pass on their genes because relatives share some of our genes. Therefore, by helping their evolutionary fitness, we are also helping our own.
Dario and his team equipped robots with a rudimentary nervous system as well as cameras for eyes, and special sensors to help them find ¨food¨ disks. Researchers had the robots wheel around an arena in a classic game of survival of the fittest. The robots that got the most disks won and were able to ¨live¨ in the next generation.
Researchers then used computer simulations to mimic what may happen in social evolution. Would they see evolution of altruistic behaviour in these intelligent robots? To make the tests more realistic some robots were programmed with mutations. Just as in nature, mutations can serve us both disadvantages and benefits. Due to these mutations some robots became either better or worse at finding food disks. After several trials, the robots that remained in the arena were proficient food finders. At this point, the researchers made clones of some of the robots and during the trials they began observing that robots that were highly related were more likely to share food with each other.
The amazing thing is that the robots were only programmed to have the capability to share food with each other. They were programmed in a way that allowed the robots to ¨decide¨ whether or not to do this, much the way animals do in nature. While this may seem like a very artificial representation of a real life phenomenon, it is a very innovative way of a testing a theory that is so difficult to study in real life. Dario is also a member of the Coordination Action for The Robot Companions for Citizens FET Flagship Initiative.
Click HERE to read the full research published in PLOS Biology