Do bees, swarms of bees, make you nervous? Maybe not. Maybe they remind you of honey, flowers and warm summer days. You stay out of their way and they stay out of yours. What if, however, the bees weren't bees at all but hundreds (or thousands) of autonomous microbots, facsimiles of the real thing, buzzing around in the real world?
It won't be easy, according to Rob Wood, the project's principal investigator.
"The challenges that you get when you scale these things down mean that you have to reinvent everything, everything has to come from scratch, every one of the technologies," Wood said in an interview last week. "There is nothing off the shelf."
His team has recently taken a significant step forward with the demonstration of a manufacturing process that they believe will pave the way to industrial-scale production of microbots, spitting out thousands of them with ease.
The prototype creature built with this process has been dubbed the "Mobee," short for Monolithic Bee. With its carbon-fiber airframe and titanium wings, it doesn't really look like a bee. It's more like an intricate box frame with insect-like wings. The creatures are so lightweight that Harvard says the mass of 63 Mobees is equal to the mass of one United States quarter.
Roughly speaking, the Mobee starts as more than 18 plain sheets of material. Each sheet of carbon fiber, titanium, plastic, etc. is cut to a certain pattern by a laser. The sheets are then sandwiched together into a cohesive whole. The final step is where the magic happens.
Jacks beneath the flat sandwich push upward, allowing hinges in the design to transform the package into a three-dimensional microbot. Largely speaking, it's finished at that point. Harvard describes the process as something akin to what happens in the pages of a pop-up book or in the creation of complex origami figures.
Graduate students Pratheev Sreetharan and Peter Whitney helped develop the manufacturing process. Sreetharan, who also designed the Mobee prototype, says the hard part was hand drawing the CAD files for each of the 18 layers the laser-milling device had to shape.
"For a device like Mobee, I would say I designed it in 2 1/2 weeks. That's actual design work," Sreetharan said. "Two-and-a-half months after that was literally drawing lines [for the CAD files]."
Before the development of this pop-up manufacturing concept, the Robobees team had built each individual creature painstakingly by hand. Only about 15 percent of the bees built this way worked properly. With the new technique, where robots are essentially building robots, based on a human design, the yield is in the range of 100 percent; perfection.
The beauty of the technique, besides precision and efficiency, is that it can integrate future elements of the project directly into the designs. So, when the sensor, control and power elements of the Robobee project are ready, they can be added as integral component layers of the bee structure, rather than as bolt-on additions to the design.
Before the fully functioning Robobee becomes a reality, however, Harvard hopes to take this manufacturing process into other arenas. They believe, among other things, that it could used today in the manufacture of articulated, electro-mechanical surgical instruments. The team is working with the Wyss Institute for Biologically Inspired Engineering to apply what they've learned with Mobee to the manufacture of instruments for laparoscopy and endoscopy.
Graduate student Whitney says their technique is a fit for these applications because if falls in a zone between the very small processes applied to computer chips and the larger processes found in traditional machine manufacturing.
"When you want to make a fully integrated device that's on the scale of 1 or 2 cm, it's not economical to scale up the computer-chip, integrated-circuit manufacturing technology; it's just too expensive," Whitney said. "On the other hand, trying to scale down traditional manufacturing, you're sort of in a Swiss-watch situation where you can make tiny components, but then you have to assemble them all together, and that's a tedious process; so we sort of live in between those two worlds."
The real question hanging in the air, so to say, is how the bees themselves might be used once they've been endowed with the power, sensor and control mechanisms needed to fly and operate on their own.
The obvious answer is surveillance of all types, whether it's for the military in combat or scientists tracking changes in the environment or spooks keeping tabs on their targets. Oh, and they might also be able to pollinate crops of vegetables and flowers, too.
We have a few years, at least, to figure out how how swarms of robotic insects might fit into our lives. The Robobees team is working along three tracks: body, brain and colony. Each one presents its own challenges. Integrating the three strands, which are being worked on in parallel, is a whole other set of hurdles.
Mobee is just one hurdle cleared. We'll have to wait and see two years from now if the swarm is really on the horizon, and a little longer to decide if we need to run from it.