Leveraging Robotic Insects to Support Mechanical Pollination

Flowering plants, including crops, rely on pollination, or the transfer of pollen, to support fertilization and seed production. Insects, including bumblebees, are the main perpetrators of this vital task, flying from flower to flower, picking up pollen from the anther and then later releasing it with a gentle vibration on the plant’s stigma. 

For decades, researchers have looked for ways to artificially pollinate farms and gardens, with limited success. Now, a group of roboticists at the Massachusetts Institute of Technology (MIT) have developed a paperclip-sized aerial robot that can hover for one thousand seconds (nearly 17 minutes)—bringing us closer to the creation of viable mechanical pollination applications.

Flying pollinators

Kevin Chen, head of the Soft and Micro Robotics Laboratory within MIT’s renowned Research Laboratory of Electronics (RLE), said the idea of developing flying robotic insects is not a new one, however, trying to manage the physics of such miniscule robots has long been a challenge.

“For an application like pollination, you need a robot that can easily interact with a very soft and compliant object like a flower or leaf,” he explained. “Those things are difficult, if not impossible, for large, heavier robots.”

Chen and his team have spent the last decade trying to create a unique structure that behaves like an insect—right down to its tiny tendons and muscles. He said his group had a breakthrough when they were able to develop a new type of artificial muscle that can drive the robot’s flying capabilities. They leveraged that advance to create a 4x4x1 centimeter robot that has four actuators and four sets of wings.

“The actuators behave like artificial muscles. If you think of them as a cylinder, when you give that cylinder a high voltage, it elongates. When you let the voltage go, it retracts,” he said. “If you send voltage at around 400 hertz, the actuator oscillates very quickly, and the translational motion converts into the rotational motion of the wings.”

Those wings not only flap but rotate relative to the pitch axis, Chen said. Resembling the flapping wing kinematics of insects, this motion allows for an average speed of 35 centimeters per second. The design also allows the robot to follow a precise trajectory, as well as move with more speed and agility than previous insect robot designs. In addition, this design also minimizes mechanical stress on the artificial wing flexures, enabling increased endurance—and, as an extra benefit, provides a bit of cargo space that could potentially house batteries or sensors.

Robotic insect

While Chen’s robot shows remarkable improvement over past robotic insect designs, he and his team are already hard at work to make it even better. While this robot boasts flight time 100 times longer than other designs, he hopes to refine the design to go even longer. Furthermore, he hopes to tailor that design to support new capabilities.

“Over the next few years, we plan to pursue autonomy, meaning that when the robot flies, it does not require an external motion capture system to provide motion feedback,” he said. “In addition, we can gain more precision. Being able to fly at a particular point in space isn’t that difficult. But what is difficult is supporting higher level computation and power autonomy.”

He said the group is working on incorporating a tiny camera on board and using it to, first, recognize that an object in front of it is a flower, localize itself to understand how far it is from the flower, and then conduct proper path planning so it can successfully land on its intended target.

Then there is the matter of powering the insects. Chen envisions that one day a swarm of insect robots could work in tandem to pollinate an area. Today, his robotic insects are tethered to get their power. But a terrestrial robot that runs alongside the insects could provide the answer.

“When you consider the state of battery technology right now, robots would only be able to fly for a minute or so before they’d need to recharge. That amount of time is too short for it to do anything practical,” he said. “But if we have a hub robot that rolls along with the insect robots through the vegetation, the robots could potentially stay charged and do their thing.”

While there is still plenty of work ahead before artificial pollination applications are feasible, Chen said that this work shows that roboticists can create softer robots that can combine agility and robustness, much like the animals or insects that inspire them.

“Most of the commercialized robots we see today are still driven by rigid systems like motors. That’s because soft actuators are intrinsically slower and less precise. That means the functionality of soft robots can be quite limited,” he said. “But future soft robots can work alongside humans and interact with delicate objects yet still be forceful and precise. There’s a lot of interesting foundational science happening, ranging from flapping aerodynamics to the physics of flexors, which can help us get to that point.”

Kayt Sukel is a technology writer and author in Houston, Tex.