Photo: University of Washington
For robots of all sizes, power is a fundamental problem. Any robot that moves is constrained in one way or another by power supply, whether it’s relying on carrying around heavy batteries, combustion engines, fuel cells, or anything else. It’s particularly tricky to manage power as your robot gets smaller, since it’s much more straightforward to scale these things up rather than down—and for really tiny robots (with masses in the hundreds of milligrams range), especially those that demand a lot of power, there really isn’t a good solution. In practice, this means that on the scale of small insects robots often depend on tethers for power, which isn’t ideal for making them practical in the long term.
At the IEEE International Conference on Robotics and Automation in Brisbane, Australia, next week, roboticists from the University of Washington, in Seattle, will present RoboFly, a laser-powered insect-sized flapping wing robot that performs the first (very brief) untethered flight of a robot at such a small scale.
There are a bunch of different ways to potentially power a robot remotely, but most of them suffer from low range, or low efficiency, or don’t deliver a lot of power. For some small robots, this is just fine, but a flying robot demands a lot of power transmitted over a relatively long distance in order to be useful. There’s one obviously best way of doing this, and it’s obviously best because it’s the solution for pretty much every problem in science: lasers.
RoboFly—a design based on the RoboBee flapping-wing microrobot from Harvard’s MicroRobotics Lab—is about the size of a bumblebee, and weighs just 190 milligrams (a bit more than a toothpick). It’s powered by an infrared laser aimed at that tiny little photovoltaic cell, which can harvest the 250 mW required to get the robot airborne. In the video, the laser doesn’t track the robot, so as soon as the solar cell moves out of the beam, it loses power and the robot stops flying.
Ultimately, a RoboFly could be controlled by a ceiling-mounted laser that tracks it wherever it goes, or even lasers mounted on moving vehicles (or other robots) that can follow the RoboFly around and provide power to it indefinitely
While delivering the power is a significant part of the challenge, developing the electronics necessary to turn that power into flight is the initial focus of the University of Washington group, which included Johannes James, Vikram Iyer, Yogesh Chukewad, Shyamnath Gollakota, and Sawyer B. Fuller. The wings are driven by two piezoelectric actuators, which require around 250 volts to maximize power density. The photovoltaic cell outputs just 7 volts, so the researchers had to custom design a boost converter to drive the wings, and add a microcontroller to control them. They managed to cram both of these things into an electronics package that weighs under 100 mg, enabling the robot to function completely untethered.
It’s tempting to watch the video and be like, “Uh, that’s it?” But note that the primary thing that needs to happen for the robot to achieve longer flights is for the laser to track the photovoltaic cell so that it can provide power continuously, and this is a problem that other researchers have worked on before with some success. The UW team has been able to power the robot at ranges of up to 1.23 meters indoors, and with a higher output laser (or one that can focus a bit better), they expect that a range of some tens of meters shouldn’t be much of a problem. Ultimately, a RoboFly could be controlled by a ceiling-mounted laser that tracks it wherever it goes, or even lasers mounted on moving vehicles (or other robots) that can follow the RoboFly around and provide power to it indefinitely.
Photo: University of Washington
For more details, including what RoboFly could potentially be used for, we spoke with UW professors Sawyer B. Fuller, who directs the Autonomous Insect Robotics Laboratory, and Shyam Gollakota, who leads the Networks and Mobile Systems Lab, via email.
IEEE Spectrum: What are some of the challenges of developing electronics that can function efficiently at this scale, and how have you approached them?
Sawyer B. Fuller: The electronics package weighs less than a toothpick, and generates a sinusoid signal. It also incorporates the first microcontroller ever to fly on a honeybee-sized robot like this one. The challenge was to come up with a fast way to build and iterate ultra-light circuits, and integrate this with an efficient power delivery system, and do so at very high (>200) voltages.
Do you think that flight at this scale powered by a remote laser source is viable outside of a controlled environment? What applications can you imagine?
Shyam Gollakota: A number of applications like in farms and for finding leaks in oil pipes, we most likely will still have a line of sight. And hence laser based powering of the insect is useful. Another question you might have is about safety and we have indeed designed, in the last few months, approaches to achieve a safe transfer of high power (2 W) for consumer electronics like phones.
Fuller: By cutting the wire and incorporating an on-board boost converter, we’re also building a foundation for powering these robots from other sources, like solar, or energy harvesting Wi-Fi or cellular signals.
Photo: University of Washington
Do you think that wireless power transfer is necessary for systems like these, or do you think that energy storage is likely to improve to the point where these robots will be able to power themselves?
Gollakota: The issue with current battery technology is that they weigh a ton (relatively) and hence having sustained flight without recharging will be hard using just batteries. We envision our solar cell to be something that can also harvest from the sun directly and create an insect-scale robot that can hop around. So the cell we did introduce in this design will in principle enable robots that will be able to harvest power. One could also add supercapacitors to store some energy.
Fuller: Eventually, we see them as using a combination of stored energy and external energy sources. That said, we’d like to incorporate a battery, but that requires a special battery that is both very small and high-drain, which is not commercially available.
There have been some recent commercial demos from companies like Energous and Ossia on wireless power transfer. Any thoughts on how that might apply to insect-scale flying robots?
Gollakota: Energous and Ossia use RF signals (we believe in the 2.4 GHz, 5 GHz, and lower) to power devices. The challenge with using RF for power delivery in contrast to lasers is efficiency. Since RF signals are broadcast in nature, they attenuate significantly with distance (1/r^2 in ideal scenarios and 1/r^3 in practical scenarios). Thus to deliver 100 mW at a distance of 10 meters, the transmit power is multiple orders of magnitude higher. In contrast, since laser signals are focused they do not have a similar attenuation. Further, while RF signals can be made directional by using multiple antennas, to get the beam-like behavior at 2.4 GHz requires hundreds of antennas, which is quite difficult, large in size, and expensive to achieve in practice.
Fuller: I think it would be really great to explore using their technology to potentially power insect-scale robots—but it would have to be over very short range as Shyam suggested. Our robots use very little, opening up many potential sources of power that might be too minute—or too dangerous in density—for larger drones.
What are you working on next, and what will it take to extend the flight time/altitude and make a robot like this controllable?
Fuller: Our next step is to steer the laser at the robot as it flies, so that it can hover indefinitely or fly longer distances. The laser beam has been demonstrated at up to a meter range, but we think it can easily go much farther.