As robots become more commonplace and begin to play an even bigger role in industrial applications, maintaining them and ensuring that there is a continual evolution of them has led to a whole new industry in maintenance and development. But, what if they could be designed to improve and maintain themselves. Then, their uses can be further increased and extended to work in even more remote locations.
Researchers at the University of Cambridge have been studying Charles Darwin and Alfred Russell Wallace and looking at the principles of natural selection. As you know, the basic principle of natural selection is "survival of the fittest", meaning those who are better in their roles will be chosen to carry on the family genes therefore, their offspring will perform even better in the future.
The team at Cambridge created a "mother" robot (a robotic arm) which in turn, could build different sized "child" robots using plastic cubes with a motor inside. Without any human intervention, beyond the initial command, "She" built and tested five generations of ten children and then observed how they performed at various tasks. At the end of each task, she selected the one which performed best and used its information to build another generation of ten children so that the best or fittest would move forward.
According to the report, "the researchers found design variations emerged and performance improved over time: the fastest individuals in the last generation moved at an average speed that was more than twice the average speed of the fastest individuals in the first generation. This increase in performance was not only due to the fine-tuning of design parameters, but also because the mother was able to invent new shapes and gait patterns for the children over time, including some designs that a human designer would not have been able to build."
It may be many years before we're anywhere close to humaniod self-replicating robots, but evolutionary robotics is a growing field and in time, they will have potential to replace humans in many tasks where currently robots aren't a practical alternative to humans.
To listen to the team leader discussing the future of bio-inspired robots, watch this video.
Speaking of robots replacing humans, could we soon see boy bands being replaced by automation and would we notice the difference? At the University of Arizona, a team led by Kelland Thomas is trying to teach robots how to improve music and get them to play real instruments.
By sending a database of jazz tunes through the learning software, it's imagined that the robtos will be able to learn the music so well that it's able to improve based on that knowledge.
According to Thomas, "A jazz musician improvises, given certain structures and certain contraints and certain basic guidelines that musicians are all working with. Our goal is going to be an improvisational system. So yeah, it will be able to jam.".
It's not all fun and games and because this research is funded by the US Defense Advanced Research Projects Agency (DARPA), you'd expect it to have a military purpose and tha's exactly the plan. By feeding robots with all the information available, drones and other miliatry robots could be able to decide their own movement patternes and strike targets with the same certainty as their human counterparts.
But until that happens, let's have some music.