"Parasitic robot" tosses treats to steer a cyborg turtle
While the most powerful robots are still learning to walk, millions of years of evolution has made animals masters of tricky terrain. To take advantage of this scientists are exploring ways to augment living animals with robotic systems, and now a team from the Korea Advanced Institute of Science and Technology (KAIST) has used a "parasitic robot" to control the movements of a live turtle by feeding it treats.
Cockroaches, dragonflies, and locusts have all gotten the cyborg treatment in the past, allowing scientists to take advantage of their natural talents. After all, why build a drone from scratch that can hover and zip around at high speeds when a dragonfly can already do just that? By the same logic, a turtle is already adept at crawling over rocky ground and swimming, so it's more straight forward to just build a system to control the animals.
A few months ago, the KAIST team was experimenting with having a human user send instructions to steer the turtle just by thinking "left" or "right," but this new iteration instead allows the creatures to autonomously move towards a target. It makes use of what the team calls a "parasitic robot" attached to the turtle's shell and programmed to move to a given waypoint. Turtles naturally move towards sources of light, so to direct the animal the robot flashes an LED in front of its face. When the turtle moves in the right direction, the robot reinforces the behavior by feeding it a little treat.
After a training period of five weeks, the team says the dynamic duo of robot and turtle were able to move around a water tank to reach changing waypoints with ease. In future, other animals like mice, birds and fish could be fitted with similar systems for a range of exploration and reconnaissance missions.
"This hybrid animal-robot interaction system could provide an alternative solution to the limitations of conventional mobile robot systems in various fields, and could also act as a useful interaction system for the behavioral sciences," says Dae-Gun Kim, first author of the study.
The research was published in the Journal of Bionic Engineering.