Robots evolved by exchanging genetic materials for generations

"Our work is the first to show that epigenetic factors are used in improving the physical system of robots," the scientists say.

Researchers from Vassar College conducted a series of experiments for robots competing with each other in different tasks. This is to evaluate growth over 10 generations when they have exchanged genetic genetic material in a process similar to mating.

Research for the first time shows that development factors play a role in the evolution of robots as in biological systems. The study also points to Darwin's evolutionary principle that makes sense for robots and marks an important step in understanding the interplay between evolution and development that contributes to these systems.

After 10 generations, the physical manifestation of the robot has evolved to a new step with the way of changing wiring - thought to be the result of building a new genome - (Photo: Frontiersin.org).

The robot in the study is assigned two tasks: collecting light and avoiding obstacles. The researchers used random pairing algorithms to allow them to exchange genetic material from gene pool.

These genes are made up of binary code, which regulates different conductors in the robot's hardware. When the robots "reproduce" over 10 generations, the physical expression of the genome has changed by changing the wire according to the new genetic set. The researchers measured the robot's progress by implementing and comparing this change with simulation.

Although evolutionary robots have been known for some time, researchers say this is the first time epigenetic factors are considered. Jake Brawer and Aaron Hill said: "For robot researchers, the challenge of Evo-devo (evolution and development: development) is to create systems that represent the material outside the robot combining all three. The scale of time and processes inherent in each behavior, development and evolution Because of the complexity of building and developing the physical part of robots, this is a difficult challenge to find. " evolution of things. "The first step towards this goal, we created a physical representation system that allows us to systematically test such evolutionary and evolutionary processes. "

Researchers use random pairing algorithms that allow the exchange of genetic material from gene pools made up of binary code, regulating different conductors in the robot's hardware, similar to mating in organisms. - (Photo: Frontiersin.org).

The researchers said: "It is important to note that our goal is not to show adaptive evolution but to test the hypothesis that epigenetic factors can alter dynamics. evolution of expressing the outside body of some robots According to our knowledge, our work is the first to show the epigenetic factor used in improving the physical system of robots ".