Research Proves Drones Sound Like Bees, Which Is Good News for Elephants

There isn't much that scares an elephant, but drones designed to sound like bees could help save them

3 min read

Evan Ackerman is IEEE Spectrum’s robotics editor.

Close up of a bee
Photo: iStock Photo

There isn’t much that scares a fully grown African elephant. Lions? Maybe a little bit. Humans? Not really. Mice? Nah, that’s kind of a myth. It turns out that what really scares elephants is something much smaller, although it can fly in large swarms: honeybees.

And it makes sense, because an elephant’s bulk doesn’t do much to protect it from bees, which can find all kinds of unpleasant nooks and crannies to sting. Elephants reliably flee from bees, which has led some communities to create fences made of beehives to keep elephants from raiding their crops. It works pretty well.

The Duke researchers ran a series of tests to compare bee sounds with drone sounds, and were able to show that especially at the high frequency ranges, the similarities are significant

A few years ago, researchers from Duke University brought a quadcopter to Wonga Wongue National Park in Gabon, in West Africa, to develop a system for monitoring African forest elephants. The elephants did not appreciate being monitored by a drone one bit, and would run away from it or even throw dirt at it with their trunks.

Is it a coincidence that elephants that don’t like bees also don’t like drones? No. It isn’t. And that’s information we can definitely use.

I don’t know whether the average drone sounds more like a swarm of angry bees or just one gigantic angry bee, but elephants don’t make much of a distinction between the two, and they react similarly to drones and bees both, as the Duke researchers describe in a paper from last year:

Upon testing the [drone] with the park rangers, known as Ecoguards, and the researchers observed signs of disturbance amongst the elephants when using the UAV system. Interestingly, some elephants seemed unaware of the UAV, while others seemed distressed based on the presence of the UAV. Some signs of distress involved throwing dirt upon hearing the UAV and quickly retreating, as well as spraying dirt with their trunks, a behavior known as dusting. In many of these cases the elephants would not have been able to visually perceive the UAV based on the terrain, which suggests that the elephants were exhibiting these behaviors upon only hearing the UAV.

For humans trying to track or study elephants, this is bad news, because (as the researchers put it) “the usefulness of UAVs applied to conservation is limited if the animals are distressed by the technology implemented to help them.” On the other hand, for humans trying to keep elephants away from other humans for their own good, this is good news, according to this PBS NewsHour report:

The question, though, is whether elephants are really mistaking the sound that a drone makes for the sound that bees make, and if so, how that could be more effectively leveraged for conservation. The Duke researchers ran a series of tests to compare bee sounds with drone sounds, and were able to show that especially at the high frequency ranges, the similarities are significant.

Specifically, a 3DR Iris+ sounds the least like bees, while a DJI Phantom sounds the most like bees, although it’s worth noting that the researchers were only able to test a small handful of drones. This is not to say that an elephant would only mistake a DJI Phantom for bees and not the 3DR Iris+, but it does provide a way to score the absolute “bee-ness” of a drone, potentially helping conservationists choose the most appropriate platform for their work. 

Using this metric, it may also be possible to make targeted modifications to drones to either increase or decrease their bee-ness, and this technique could be applied to other sources of stress as well, depending on what animal the drone will be interacting with:

It should be noted that this data analysis is based on the assumption that there is an aural source of discomfort displayed in a particular species, in this case elephants. Researchers wishing to implement this approach to justify what platforms are better than others for wildlife conservation must initially determine a potential auditory source of discomfort for a species of interest from which recordings could be obtained. 

For humans, there’s peer-reviewed research on the sounds that annoy us the most. Or, you could just leave the drones as-is, since most drones are pretty effective at annoying humans anyway.

[ “Small UAV Noise Analysis” (pdf) ] via [ Unmanned Systems ]

The Conversation (0)