AntBot Demonstrates Amazing Navigation Ability, No GPS Required

Humans still have much to learn from nature. It was determined long ago that ants have tremendous navigational abilities. Depending on the type of ant and the ant’s environment it will either depend on pheromone trails to find its way home or something called path integration. The common ant will generally depend on pheromone trails to assist in navigation. This technique involves tracing their smells to backtrack. On the other hand, the ants that tend to have exceptional path integration abilities are desert ants, living in an environment where pheromone trails cannot exist due to excessive heat.

Path integration is a process that entails counting the number of strides, knowing the speed of movement, being aware of the sun’s location and remembering visual cues. Amazingly, ants utilize all this information to reliably find their way back home after venturing out over large distances.

Equipped with this knowledge, French scientists set out to determine if they can mimic the extraordinary navigational skills of the desert ants. The first step was to build a robot with the necessary sensors; some of which include a celestial compass, optic flow sensor and Pololu MinIMU. Of course, there has to be a brain to drive the bot and read the sensors. In this case, the Raspberry Pi was the board of choice.

The celestial compass is made up of two UV light sensors along with sheet polarizers. The UV sensors are used to detect the light, and the sheet polarizes are implemented to help derive the precise location of the sun through Earth’s atmosphere. While the ideal atmospheric conditions are clear, it can include clouds and overcast skies. Through testing, the scientist found that the AntBot is able to determine its angle of direction with a median error of less than one-tenth of a percent accuracy when there is a small amounts of cloud in the sky. With an overcast sky, it is able to determine its angle of direction with a median error of less than six-tenths percent accuracy.

Next, algorithms were created on the Raspberry Pi to test the bot’s capabilities of using its sensors to find its way back to a starting point after wandering out. Current tests show very good results. After traversing a distance of 14 meters, AntBot was able to return to its starting location within an impressive 6.5cm. However, the scientists are quick to point out that much more testing needs to be done before they can adequately compare these results against desert ant’s abilities. Desert ants can wander out over a distance of more than 700 meters and precisely find their way back. For an accurate comparison, AntBot would have to traverse a distance greater than 30km and still find its way back.

The concepts applied in the AntBot experiments can have several applications to autonomous robots. It can provide a navigational system for agricultural or geographical analysis. More importantly, it can serve as a means of navigation in environments where communication and GPS may not be reliable. This can include exploration missions in space or emergency situations during the time of natural disasters. Amazingly, this may all be possible due to inspiration from the way a tiny desert ant navigates.