(a) Two Alice robots facing each other, with (left) and without (right) the additional module for light detection. (b) Three Alice robots pursuing a light trail. (c) Typical time course of an experiment with three loops (access to a fourth loop that is visible had been blocked) and symmetrical bifurcations. The letter S indicates the starting area of the network where the robots are placed at the beginning of the experiment. The letter T indicates the target area. The top three pictures represent 3 snapshots of an experiment where a group of 10 robots selects the shortest path. (via Simon Garnier)
The highway traffic system can appear to be a bit of a pain at times - the dreaded rush hour which results in a seemingly endless line of vehicles slowly tip-toeing their way home. The solution can not possibly be to continue forth designing similar wide-lane traffic transportation systems that lead to traffic inducing, bottleneck effects. A way of achieving optimal travel at any time of day must exist, and it appears that ants have it down pat.
On daily foraging missions, ant colonies efficiently travel great distances with the help of several navigational aids. These typically include: visual cues, pheromone based route tracking, celestial navigation, and even step/body rotation counting - not unlike many fitness enthusiasts of today. But when faced with a fork in the road, the possibility of ant utilizing the geometric properties of their environment to navigate comes into question. This hypothesis has been tested in a recent study contributed to by Simon Garnier, Maude Combe, Christian Jost, and Guy Theraulaz at the New Jersey Institute of Technology’s Swarm Lab.
The sugar-cube sized ant robots were equipped with two light sensing antennae for the study. In addition to testing the geometric variable present in their route navigation system, the ant method of using pheromones for navigation was simulated using light; each individual robot would leave a light-trail on its path down the maze. A previous study explained that the angle created by a fork in the road determined which path an ant would take: a small incident angle led to the food source, while a larger incident angle generally led back to the nest. Thus, the study was performed utilizing two distinct mazes: one with symmetrical bifurcations - perfectly symmetrical forks in the road - and one with asymmetric bifurcations.
Programmed to like ants, whom generally follow a relatively straight path with little desire for exploration, the 10 robots were allowed to navigate the maze galleries as close attention was paid to the relationship between the movement of the robots, the light-trail path, and the maze bifurcations.
The maze galleries were carved out of PVC and white cardboard and presented the robotic ants with a network of 9cm wide, 2.5cm high corridors. The robo-ants, named Alices by the research group, were built at the EPFL in Switzerland. These 22mm x 21mm x 20mm bots were equipped with two watch motors with a max travelling speed of 40mm/s, four infrared sensors for target and obstacle detection, and a photodiode equipped module to detect changing light gradients.
The results of the experiment demonstrated remarkably similar patterns of movement between the ants and the Alices. The findings showed that the physical angle of the bifurcations were not as important in the individual case: by monitoring the bot behavior with respect to the “pheromone” light-trails it was found that the ant-bots preferred to take the path that had been previously travelled by another bot. In this method, once the shortest route was found, more and more ant-bots would take that same route due to the increased light “pheromone” presence. At the collective behavior level, the group of Alices were more likely to choose the shortest path to the target destination when traversing an assymetric network.
Simon Garnier, the research team leader, wasn’t too surprised that the experimented explained ant navigation behavior as dependant on both physical layout and the presence of pheromones. Ants have demonstrated a preference to navigate through angled bifurcations that require the least bodily movement - hence the small incident angle - and that also contain a stronger pheromone scent; this results to the development of a positive feedback loop traffic system. The research will further help scientists study how physical layouts and pheromone trails in the environment can affect the travel of other insects. Subsequent work hopes to stem into the study of how similar navigation systems can be used to alter human-made environments for optimal travel.
Also, can we have Ant-Bots for our holiday 2013?
C
See more news at: