It buzzes, and it’s got a keen interest in flowers. But the tiny creature that just landed on a blooming lily, ready to pick up pollen, is not a bee. It’s a drone.
The miniature machine is Japanese roboticist Eijiro Miyako’s solution to take some of the workload off honeybees and increase crop yields. Worldwide, bees are dwindling, falling prey to a combination of a deadly parasitic mite called Varroa destructor, pesticides, climate change and deforestation. More and more bees are abandoning their hives, in what is known as colony collapse disorder. According to a recent study, US beekeepers lost a shocking 44% of their hives in 2015-16.
This is very much a human problem too, because bees and other insects are needed to cross-pollinate about a third of the world’s food crops. Without bees, warns a UN report, plants including carrots, onions, apples, oranges and even coffee will struggle to thrive and reproduce. Ultimately this could cripple £465 billion worth of annual food production globally.
That’s why Miyako, who works at Japan’s National Institute of Advanced Industrial Science and Technology in Tsukuba, and many other researchers around the world are looking for technologies to either save or replace our buzzy bees.
This is not just about getting honey, of course; cross-pollination also boosts genetic diversity and improves the quantity and quality of crops.
Miyako plans for the worst-case scenario: what if bee numbers collapse to a point where there are simply not enough insects to carry pollen from flower to flower. After all, whether this is done by a living or a mechanical bee is secondary. So Miyako has developed drone bees, with their bellies covered in horsehair that has been coated with a sticky gel to pick up the pollen. When the robot lands on the next flower, the pollen rubs off – resulting in the world’s first artificial pollinator.
Miyako stumbled upon his drone idea by accident. He recalls how in 2015, while cleaning his office, he rediscovered jars of gel he had created a decade ago. “I kept the gels in a drawer and forgot about their existence completely,” he says. When a jar slipped from his hand, a few drops of the gel landed on the floor and absorbed dust. That made him think of pollen grains stuck to a bee.
But robotic bees were not his first option. Miyako started off by using ants – real ants – that had a small drop of gel attached to their backs. He put them in a container with flowers for three days. While the ants could absorb pollen from the flowers, they didn’t have the tiny hairs bees use to pollinate flowers. And of course ants being alive, and being ants, were rather impossible to control. So last year he turned to robots.
At the moment, he buys toy drones from an online retailer, the latest ones made by Japanese toy company G-Force: 1.5 by 1.5cm quadcopters that weigh just 15g. They have already done some useful work and cross-pollinated Japanese lilies (Lilium japonicum) in his lab, says Miyako, adding that he “saw actual growing pollen tubes on the female flower, to show that the pollination was complete”.
Now, the drones are still controlled by an operator, and their batteries have to be recharged for 15 minutes for about three minutes of flying: not ideal for work on a farm. But Miyako and his team have recently got in touch with roboticists, who are looking into building autonomous robotic ‘bees’ that would be equipped with GPS, artificial intelligence and high-resolution cameras to independently – and accurately – hop from flower to flower.
One challenge for the drones will be to recognise the flowers they’ve already been to – something that real bees achieve thanks to an electrical charge that serves as a reminder, says Richard James MacCowan, founder of Biomimicry UK. “How the drones can be designed to do this at a low-cost and low-tech solution will be key, especially in developing nations,” adds MacCowan. Miyako is confident that his future robots will get there at some point, though, eventually creating “a swarm of robotic bees that are going to self-think” and recognise images of flowers they’ve been to.
It’s not the first time that scientists have worked on creating robotic bees. Researchers at Harvard University in Massachusetts developed a similar device back in 2013, which used static electricity to latch onto surfaces, but they never used their drones for pollination. And there are other efforts going on around the world to save bees – and help our crops.
For instance, US semiconductor giant Intel together with Australian scientists are focusing on the bigger picture, hoping to understand what causes the bees’ decline. They recently equipped almost 10,000 bees on the island of Tasmania with tracking chips to follow their trails. Using radio-frequency identification (RFID) tags mounted on the bees’ backs, they monitored whether the insects diverged from their usual behaviour. The tags “work just like an electronic badge used by a staff member of a company,” says Paulo de Souza, science leader at Australia’s Commonwealth Scientific and Industrial Research Organisation.
A tracking chip is 2.5mm long, 0.4mm thick, and weighs just 5.4mg – while a honeybee can carry up to 12mg of water, pollen or nectar. A system collects data from special RFID readers placed at check points, such as hive entrances or feeders that the bees visit. The data is then analysed if a hive collapses.
The company uses its low-power Intel Edison compute module, in the hope that the system can be used in remote areas such as the Amazon. The modules can be customised to measure different parameters, such as temperature and honey production.
First results are already trickling in, allowing the researchers to “better understand bee behaviour and their responses to stress factors and weather conditions,” says de Souza.
The scientists can now better estimate changes in the life expectancy of bees, once they have been exposed to stressors such as pesticides, pathogens, parasites and poor diet, he adds. “We were also able to determine when bees move from one hive to another, which is an important factor for disease spread.” Researchers in Kenya, the United Arab Emirates, New Zealand, Argentina, Brazil and Mexico are already using the technology – and Europe and North America are next on the list, he says.
Another company engaged in monitoring is Bee Smart Technologies, which allows beekeepers to survey their hives remotely and respond early if they suspect a problem. The company, based in Bulgaria and California, provides beekeepers with a tiny box of sensors that attaches to the frames inside a hive and monitors not only temperature and humidity, but how active the worker bees are and even whether the queen is mating.
Then there is Plan Bee, a firm in Scotland that also offers a monitoring service for hives – but, on top of that, provides data analysis with the help of a small Raspberry Pi computer.
Meanwhile, the Open Source Beehives project has a different approach. It aims to support bee colonies and track their health and behaviour with the help of open-source sensory kit, the Smart Citizen Kit, which transmits information from a hive to an open data platform, Smartcitizen.me. The designs of these sensor-enhanced hives are freely available online, and the gathered data is published together with geolocations, to provide further analysis of the hives and their buzzing inhabitants.
Monitoring is important, but what’s to be done about the deadly Varroa mites that contaminate hives, latch on to bees and suck the life out of them? Pesticides have traditionally been the preferred option, but many mites are now resistant – and, once a pesticide is being applied, the entire hive has to be shut down for a while.
Czech scientist Roman Linhart believes he has created a solution with his Thermosolar Hive. It uses solar power to heat up the hive, killing the mites without affecting the bees.
Linhart had the idea on a hot summer’s day when he was in pursuit of snacks, he says. Going to a convenience store, he noticed a beehive under its tin roof. The store owner told him it had been there for 17 years. Linhart began to inspect the hive closely, and managed to collect a swarm that he placed in a hive of his own. He guessed – correctly – that heat was responsible for the mite-free prospering colony, and began developing his own technology.
After concluding that other heat sources proved to be time- and energy-consuming, Linhart settled on solar energy for his thermal treatment. His Thermosolar Hive sports a wooden outer layer with a thermosolar ceiling underneath, which is exposed to sunlight during treatment. The increase in temperature is measured by sensors. Once the temperature inside reaches 47°C, the outer layer is placed back on. The mites are killed – while the bees and honeycombs remain unharmed. Linhart started selling his hives in January.
And how about learning from humans? Entomologists in the US have recently started the world’s first bee sperm bank, collecting semen samples from male bees, and freezing them in liquid nitrogen, where they can remain for decades. The idea is to boost the genetic diversity of bees, and allow beekeepers to selectively breed for traits to better fight against colony collapse disorder.
Still, the collapse of the world’s honeybee population has not yet been contained. And, while Intel’s, Linhart’s and some other technologies are already being deployed on a commercial scale, Miyako’s drones are yet to fly out of the laboratory. “I wanted to use the drones outside in the campus of our institute, but it was impossible,” he says. “The Japanese government has certain laws for controlling drones outdoors and I haven’t yet received permission.”
But he hopes that his robotic ‘bees’ will eventually take off, in about a decade or so – helping farmers, alongside honeybees, to get food on our dinner tables.
Hairy bees
Hairs matter – and, it turns out, bee hairs especially so.
Researchers at Georgia Institute of Technology in Atlanta, US have an idea how to refine the design of future robotic bees – by improving their ‘fur’.
The average bee is covered in three million tiny hairs, even on its eyes. The hairs are strategically placed to ensure the insect picks up as much pollen as possible. The amount of hairs on the bee’s legs are five times denser than the ones on its eyes, say the scientists.
The scientists found that the gap between each hair is the width of a dandelion pollen – about 45 micrometres. The hairs help bees sweep up nearly 15,000 particles of pollen from one flower to the next – effectively making the insect carry up to 30% of its body weight in pollen.
So the bottom line is, for effective cross-pollination, future drone bees have to have hairs, say the scientists. They even designed a smooth robotic leg and covered it in wax – but it picked up four times less pollen than a real bee’s leg.