Steffen Wischmann / University of Lausanne
Researchers used a simulated version of these two-wheeled robots with flashing lights to show how randomness in the occurrence of genetic traits can drive evolution of language.
Even if everything about different groups of animals is identical down to the level of their genes and physical surroundings, they can develop unique ways to communicate, according to an experiment done with robots that use flashing lights to "talk."
The Swiss researchers used the robots to get handle on why there is such diversity in communication systems within and between species, something that is difficult to do in living animals.
The answer, they found, "is contingencies in evolutionary history, i.e. stochasticity (randomness) in the occurrence order of new ... traits," Steffen Wischmann, a researcher in the department of ecology and evolution at the University of Lausanne, told me in an email.
He and his colleagues started with 20 populations of identical two-wheeled robots each equipped with a camera, a food detection sensor, a simple information processing program, and a ring that could emit a blue or green light.
These robots, grouped in populations of 20 individuals, were placed in an arena containing a food source. The team ranked each robot according to how long they spent at the food source.
They then used a "standard roulette-wheel selection algorithm" to select 100 robots' programs, or genes, for reproduction, according the paper published this week in the Proceedings of the National Academy of Sciences.
"Because the 'genes' — which encoded specifications of the robots neural controllers, responsible for processing sensory information and producing motor actions — were initially set to random values, the robots behaved unpredictably at first," the journal explains in a news advisory.
"But after 1,000 generations, all 20 populations emitted light to indicate food location. In approximately half the populations, the robots emitted a signal only in the presence of food, while the other populations also emitted a different color light in areas without food."
It turned out that the one-signal robots were the most efficient communicators — they found the food faster — but they were also the weakest competitors when pitted against other groups of robots who communicated with two flashing lights.
In other words, there's a tradeoff between communication efficiency and competitive robustness, the researchers note. And, randomness in evolutionary history can affect the outcome of competition between populations.
"Since the two-signal populations use both signals they can also utilize the signals of other populations independent of which signal this other population uses to signal the presence of food," Wischmann explained to me.
Further analysis of the data gleaned from the robots shows that the signaling differences occurred early in the robots' evolution.
This randomness in the occurrence of mutations can drive the evolution of language and "might also be involved in speciation processes," the team concludes.
More on language evolution:
- Robots invent their own spoken language
- Grow a new language in your head
- A baby's babble leads to language
- Mice given 'human' version of speech gene
Ten years of war have given robot developers a chance to refine and improve their bots. Now the robots are finding all sorts of new jobs on the homefront.