RECOMMENDED FOR YOU:X
Scientists Create Robot Ants That Imitate Colony Behavior and Navigate Mazes
Scientists from the New Jersey Institute of Technology and the Research Center on Animal Cognition have used robot technology to shed light on how ants forage and navigate. By creating sugar cube-sized machines called “Alices”, the researchers successfully replicated the movement patterns of Argentine ants without having to program the robots. The Alices left light trails that were able to be detected by the other units via sensors – a process similar to the way ants leave chemical markers. By choosing the path that deviated the least from their trajectory through a maze, the robots mimicked the behaviors of the insects.
At the start of the experiment, each branch of the maze contained no light trails, forcing the robots to express an “exploratory behavior”. Moving this way, they randomly foraged for signals in the same general direction. This caused the Alices to choose paths that stayed close to their trajectory when reaching a fork in the maze. If they detected light, they would turn towards that path. It was found that the robots did not need to be programmed to navigate the maze, but were able to find their way using only the light trail and random explorations. This allowed them to find a relatively direct route from their starting place to their target area on the outside of maze.
The fact that the robots were able to solve the maze in the same fashion as the Argetintine ants suggests that it does not take a complex cognitive process for the insects to establish foraging trails. It also shows that the overall geometry of transport networks has a great deal of influence on the flow of information. Humans have begun to imitate ants in how they establish trucking routes and delivery chains. By using the wisdom of animal behavior, businesses have been able to create shipping networks that are faster and more cost-effective. When it comes to efficiency, we can look to the collective intelligence of a colony.
Via Science Daily
Browse by Keyword