This project involves a swarm of robots (e-pucks with special extension boards) that are being used to emulate the foraging behaviour of ants, specifically pheromone based ant-trails. Ant trails usually consist of pheromones that are being "sprayed" on the ground by foraging ants on their way to the nest from a potential food source. Other foraging ants can follow these traces to quickly find the most efficient path to the food source.
AntBots is an experimental setup to model the pheromone trail based foraging behaviour of ants using a special phosphorescent glowing paint. We have built two custom addons for the e-puck robot that allow for trail laying and following on the glowing floor, as well as a way for the robots to mimic the ants capability of using polarization patterns as a means of navigation. Experimental results show that our trail and sun compass add-on boards are accurate enough to allow a single robot to lay and follow a trail repeatedly.
This optimization of transportation paths is a good example of "swarm intelligence" and therefore quite interesting for swarm researchers. Since ants use special pheromones to generate ant trails, the emulation of such behaviours using robots would usually require complicated and expensive chemical dispensers and chemical sensors. However, researchers at the Artificial Life Lab found a novel method that closely emulates and impressively visualizes the pheromone characteristics of an ant trail.
Here you can see a simple simulation showing the dynamics of ant-trails. Ants (red circles) which found food (blue) lay a pheromone trail (green) on their way to the nest (orange). This pheromone evaporates over time which eliminates outdated information about depleted food sources.