Dragonflies are incredible little critters. These unfathomably agile fliers are able to travel at speeds of up to 60 kilometers (37 miles) per hour during a high-speed pursuit – and, as highlighted by the University of Adelaide’s (UoA) recent research, they are able to catch their prey 95 percent of the time, even if they fly into a swarm.
That makes dragonflies one of the most successful hunters in the world. Indeed, it’s this biological prowess that’s inspired a team of bioengineers at the UoA to create a mechanical version of their own. Although the final model has yet to emerge in a peer-reviewed publication, a 2016 blog post by the university claims they are succeeding.
“We’re delighted to say that we’ve been able to build an autonomous robot that, using computational models bio-inspired by the dragonfly’s neuronal processing, can effectively and efficiently pursue targets in unstructured environments,” Dr Steven Wiederman, a senior lecturer at the the university’s medical school and the project's lead researcher, said in the blog post.
This is quite the claim, and it’d be wonderful to see a video of the robo-dragonfly in action. Still, it’s perfectly plausible. After all, if you can make mechanical and autonomous bats, why not do the same for the far more stealthy dragonflies?
Far from just studying the biomechanical intricacies of dragonflies in order to replicate them in the laboratory, the team had to determine how their neurology manages to deal with the high-speed shenanigans in the first place.
Using cutting-edge equipment, the researchers managed to detect a new type of neuron that exhibits what they refer to as an “attentional spotlight”. This neuron, unlike others, manages to devote its electrical impulses to highlighting a single target amidst a swarm of distractors.
Additionally, the researchers found a selection of neurons that are designed to “predict the future”, in that they work out the most likely possible trajectory of the target’s movement. The combination of these two neuron types allow dragonflies to capture their prey with a ludicrously high hit rate.
Remarkably, the team then replicated these neurons using computational algorithms. When placed into a virtual reality environment, the virtual dragonfly was able to take out its prey just as successfully as it did in real life.
Only time will tell if the same behavior can be replicated in the robo-dragonfly yet to make its debut, but if it can, the implications are near-endless. Such a small bio-inspired drone could be used for all kinds of surveillance and search and rescue activities, from pursuing hostile targets through busy streets to looking for people trapped beneath rubble.
2
