My friend Steve Peck is a novelist and a professor of evolutionary ecology, whose work on mathematical models of insect populations led him to the topic of one of his short stories, about a tiny flying robot, modelled after a dragonfly. The robots would cruise the countryside, looking for insect pests, which it killed, while sparing the beneficial insects. The story seemed pretty far-fetched when it was published in 2012.
I recently told the dragonfly robot story to another friend, Keith Andrews, an entomologist with years of experience in pest control in Central America. Keith immediately seized on the robot idea. âHow did it run, on photovoltaic cells? Did it collect its prey in a kind of stomach or just kill them?â
The story doesnât say what powered the dragonflies, just that they snipped off bits of their prey to store in a mechanical stomach, so that researchers could identify the bugs later from their DNA.
Real life dragonflies do hunt and kill other insects, to eat. But once a predatory insect is full it rests. A robot wouldnât need to pause and digest, and could be programmed to just keep up the slaughter all day.
âA robot would be great for that plague of locusts in Africa,â Keith said. âIt wouldnât have to damage an insect much to disable it. A good zap right between the eyes or even to the thorax or abdomen would put a grasshopper out of business.”
A pest control robot could be instructed to target only the pest species of interest, and not kill anything else. It would be the ultimate ecological pest control strategy.
Since Steve published his story eight years ago, the pieces for a dragonfly robot have started to come together.
For starters, flying robots are getting better.
In The Fate of Food, Amanda Little writes that inventors already have a prototype weed-killing robot called See & Spray, that uses a large set of digital photos to distinguish cotton seedlings from weeds. As a tractor pulls See & Spray across the field, the device spots the weeds and squirts them with herbicide. (No doubt future generations of the technology may invent alternatives to herbicide; the point is that the robot can recognize weeds).
Little also describes a robot, already in commercial use, that kills sea lice, parasites of farmed salmon, by zapping the pests with a laser (in the recent blog The Fate of Food).
If youâre wondering if digital software could work to identify pests on small farms, itâs already being done. Researchers at IITA (International Institute of Tropical Agriculture) in Kenya have invented an app called Nuru (Swahili for âlightâ) that instantly compares thousands of photos of diseased and healthy plants to distinguish between cassava brown streak disease, cassava mosaic disease and cassava green mite damage. The app is already being tested by 28,000 farmers in Kenya.
Art can inspire technology ahead of its time. Novels fueled the idea of space travel, but engineers made it happen. I can only hope that some young robotics designers will read Steve Peckâs story.
Further Reading
Little, Amanda 2019 The Fate of Food: What Weâll Eat in a Bigger, Hotter, Smarter World. New York: Harmony Books. 340 pp.
Peck, Steven L. 2012 Dragonfly Miscalculations. The Journal of Unlikely Entomology.
RTB 2019 Smarter farming: Using apps to diagnose crop health problems. In RTB 2019 Building for better science. Annual Report 2019. Lima, Peru. CGIAR Research Program on Roots, Tubers and Bananas. Available online at: www.rtb.cgiar.org/2019-annual-report
