Large flocks of bats and birds have long fascinated scientists: How do such large numbers of creatures fly so close together without hitting each other?
Now, scientists have presented findings PLOS Computational Biology that uncover a simple rule of 'wing' that make this possible for one species of bat called ‘Daubenton’s bat’.
Bats use echolocation to 'see' the world around them. They produce short chirps, at higher frequencies than humans can hear, and listen for any echoes. Based on the direction and delay of an echo, the bat can work out what objects are nearby.
The researchers used microphones to eavesdrop on these echolocation chirps, as the bats foraged for insects. They were then able to work out exactly what each bat was hearing.
Interpreting the movements of thousands of bats is very complicated, so the scientists simplified the problem by observing just two bats. They were also able to reduce the complexity from three dimensions to two, by studying a specific species, which forages close to the surface of bodies of water.
This gave the researchers enough information to develop a mathematical model that described the bats' behaviour:
If a bat hears that its neighbour has found some prey, it heads towards them and follows a mere five wingbeats behind. This makes sure the chasing bat has just enough time to react to whatever its friend does.
This is the first time that this apparently complex bat behaviour has been shown to arise from such a simple set of rules.
Dr. Marc Holderied, one of the authors of the study, contrasts this with how drivers behave at a roundabout. When several cars arrive at once, they will often simply stop until one driver tentatively moves first.
Using their simple 'Highway Code', the bats are not only more likely to find prey, but can also safely fly close to each other without risking collision.
It came as a surprise to the team that even complicated tandem and formation flights they had observed with real bats could be explained by such simple rules.
“All the behaviours just pop out of the system,” Holderied remarked.
But can such a simple model really apply to a colony of bats? The team tested their theory by creating a 3D model of a real forest.
“If we let our little software bats lose in this forest, we can see whether they crash, or whether they find a safe route through. And we can see whether these are similar to the routes that the real bats are using,” explained Holderied.
The team believe their findings could be important in the design of teams of autonomous robots, for applications such as search and rescue.
“It's a bit like the Holy Grail to get close to what animals like the bats can do.”