10 February 2009—Attempts by the U.S. Defense Advanced Research Projects Agency (DARPA) to create cybernetic insects (hybrids of biological and electronic bugs) have yielded ultralow-power radios to control the bugs’ flight and a method of powering those circuits by harvesting energy, according to research that will be reported this week at the IEEE International Solid-State Circuits Conference (ISSCC)

Two papers being presented at ISSCC reveal the latest initiatives in the DARPA-sponsored Hybrid Insect Micro-Electro-Mechanical Systems (HI-MEMS) project, which is currently in its third year. The program’s goal is the creation of moths or other insects that have electronic controls implanted inside them, allowing them to be controlled by a remote operator. The animal-machine hybrid will transmit data from mounted sensors, which might include low-grade video and microphones for surveillance or gas sensors for natural-disaster reconnaissance. To get to that end point, HI-MEMS is following three separate tracks: growing MEMS-insect hybrids, developing steering electronics for the insects, and finding ways to harvest energy from the them to power the cybernetics.

Researchers at the Boyce Thompson Institute for Plant Research, in Ithaca, N.Y.—which is one of the contractors on the HI-MEMS project—presented progress on the first goal at the IEEE MEMS 2009 conference in Italy two weeks ago, describing silicon neural interfaces for gas sensors that were inserted into insects during the pupal phase. At ISSCC, the HI-MEMS projects focused on new chip technology for the second two goals: Researchers led by DARPA contractor MIT will present a low-power ultrawide-band radio, a digital baseband processor, and a piezoelectric energy-harvesting system that scavenges power from vibrations.

The HI-MEMS project was conceived in 2005 by program manager Amit Lal, an electrical engineering professor on leave from Cornell University while he coordinates the four-year DARPA effort. MIT is one of three major contractors, including the University of Michigan and Boyce Thompson. The research also draws on the work of entomologists, electrical engineers, and mechanical engineers at the University of California, Berkeley, the University of Arizona, and Washington University in St. Louis, Mo. To be considered successful, the final HI-MEMS cybernetic bug must fly 100 meters from a starting point and then be steered into a controlled landing within 5 meters of a specified end point. On landing, the insect must stay in place.

The electronic and MEMS components of the system must consume little power and be absolutely featherweight. After all, an average hawk moth weighs 2.5 grams; with too much extra weight it would be unable to fly.

Anantha Chandrakasan, an electrical engineering professor at MIT, is a coauthor on each of the ISSCC papers. The first is an ultrawide-band receiver system on chip, a radio that works at extremely low power over a broad swath of spectrum. (Earlier research had created the transmitter.) The device was specifically built for the HI-MEMS project in order to steer the moth. To control the moth’s flight direction, Chandrakasan and MIT graduate student Denis Daly designed a small, lightweight, low-power radio connected to a tungsten 4-electrode neurostimulator. When this radio picks up the right commands, the device stimulates the nervous tissue in the moth’s abdominal nerve cord. The stimulation makes the moth’s abdomen move in a way that alters the direction of its flight. The radio and stimulator are powered by a hearing-aid battery.

The second chip is a low-power digital baseband processor that can very quickly synchronize with wireless signals. That solves a particular problem with wireless communication. “When you send a piece of data through a wireless link, the receiver takes some time to lock to the transmitter,” Chandrakasan says. “Our new algorithms can very quickly synchronize, which means that you can turn on the radio, take the piece of data, and then turn the radio back off very quickly. That saves a lot of power.

A third chip being presented at ISSCC, which Chandrakasan says is unrelated to the radio chips and not funded under HI-MEMS, could nevertheless be used to meet the DARPA project’s goal of finding ways to efficiently harvest energy from the moth. While a cyborg insect would be fairly autonomous and self-fueling, there would be no way to recharge its equipment payload on missions. Batteries are heavy. So the researchers are seeking a method by which the insect’s flight itself generates the electrical energy the payload electronics require. Harvesting ambient vibration energy through piezoelectric means—in which energy is converted between mechanical and electrical forms—could supply between 10 and several hundred microwatts of power.

The research presented at ISSCC addresses a common problem with energy-harvesting circuits: The power consumed by the harvesters’ control circuits reduces the amount of usable electrical power. The solution, a circuit called a bias-flip rectifier, improves the power-extraction capability by “more than four times,” according to the paper by Chandrakasan and graduate student Yogesh K. Ramdass.

PHOTO: ALPER BOZKURT, BOYCE THOMPSON INSTITUTE CONTROLLED FLIGHT: The moth can be made flap its wings under computer control.

The HI-MEMS project is not the first attempt at creating cyborg animals. The list is long, including pigeons, beetles, cats, and bees. Perhaps the most famous example is the cyborg rat. In 2004, John Chapin, a professor at the State University of New York Health Science Center, in Brooklyn, demonstrated Rescue Rats. These were lab rats with neural implants that encouraged them to steer through rubble piles with a camera and GPS locator to find people. Using a radio remote control, Chapin stimulated a part of the rats’ brains that mimicked the sensation of being touched on the whiskers. In response, the rats turned in the direction of the sensation. When they turned, Chapin rewarded them with a quick jolt of electricity in the pleasure center of their brains.

Jelle Atema, a biologist at Boston University and at the Woods Hole Oceanographic Institute, was also funded by DARPA in 2005 to research steering sharks with similar neural implants. Atema says that while he applauds the HI-MEMS project for its technical ambition and engineering virtuosity, he is concerned about its ultimate biological feasibility: Electronic control would compete with natural brain processes. He cites some limitations for insects, including a tendency for moths to approach light sources (the proverbial flames) and a powerful sex pheromone response that could override attempts at remote electronic control. “Pheromones are incredibly powerful,” he says.

In addition, modifying just one moth would be prohibitively time-consuming and expensive, especially in light of the life span of the animal, says Atema.

Even if HI-MEMS never produces a working cyborg moth, Chandrakasan says that the usefulness of these devices is not limited to the specific DARPA project. You can repurpose the chips for assistive technologies and implantable devices. In particular, he says, the energy-harvesting system would be a promising technology for prosthetic arms, which have a similar problem with weight and battery life.