Scientists repurpose hoverfly vision to detect drones by sound
It's a bug's life
Video After some careful study, it turns out the brain of an insect is pretty good at separating signal from noise.
Researchers from the University of South Australia, Flinders University, and Australian defense company Midspar Systems found that to be the case when they teamed up to reverse-engineer the visual systems of hoverflies. Why? To improve acoustic drone detection software.
Specifically, they wanted to use a bug's visual pathways to detect acoustic signals. It's the first time this particular approach has been taken, though insect vision has been used to improve detection systems in the past.
"We have now shown we can pick up clear and crisp acoustic signatures of drones, including very small and quiet ones, using an algorithm based on the hoverfly's visual system," said Anthony Finn, UniSA Professor of Autonomous Systems.
The result of their research is a drone detection system that can hear a drone, and pick it out among a lot of background noise, from nearly four kilometers away. Bio-inspired algorithms, compared to traditional techniques, improve detection ranges up to 49 percent and detection rates by up to 50 percent, the researchers said.
The reason for choosing a hoverfly brain, and its visual system in particular, comes from the insect's ability to see and navigate in poorly lit conditions with a high degree of visual noise.
- Chinese drone-maker DJI denies aiding Russia's Ukraine invasion
- Alphabet's Wing drone unit inks supermarket delivery deal
- India bans drone imports to help local manufacturers take off
- Heart attack victim 'saved' by defibrillator delivery drone*
Dr Russell Brinkworth, associate professor of autonomous systems at Flinders University, said that the team worked under the assumption that "the same processes which allow small visual targets to be seen amongst visual clutter could be redeployed to extract low volume acoustic signatures from drones buried in noise."
Scanning acoustic signals with an insect's visual pathways required converting it into a two-dimensional spectrogram. That image is fed to the hoverfly "brain," which is used to suppress noise and unrelated signals, which in turn makes the drone, however quiet it may be, stick out like a sore thumb. Here's a video here about the tech:
There are immediate implications for this work, both in war zones like Ukraine where drones have been used to deliver explosives, and in airports and other restricted air spaces where drones can be a hazard.
"Developing the capacity to actually monitor small drones when they're active near our airports or in our skies could be extremely beneficial towards improving safety," Brinkworth said.
As it was funded by the Australian Department of Defense's Next Generation Technologies Fund, this research will go directly toward addressing the use of drones in warfare. The researchers' findings are also available freely online, and published this year in the The Journal of the Acoustical Society of America, which opens the door to further study and development. ®