Morphing Robot Under Development

At only 16 inches wide and 27 inches long, iRobot Corporation’s PackBot Explorer can enter tunnels, sewers and collapsed structures. However, the Defense Advanced Research Projects Agency is now looking for a shape-shifting robot that can enter a structure through an even smaller opening, then regain its shape and perform search and rescue or reconnaissance duties.

Military seeks pliable options for traversing small openings.

The Defense Advanced Research Projects Agency has embarked on a quest to develop a soda-can-size robot that can shape shift enough to fit through a hole the diameter of a quarter. Working with industry and academia, the agency’s Chemical Robots program seeks to create a new class of soft, flexible, meso-scale mobile device that can navigate through arbitrarily shaped openings. As envisioned, the robot would then perform tasks related to search and rescue or reconnaissance, depending on the payload.

Work began earlier this year on the first phase of the potentially three-phase project to develop the morphing robot. The Defense Advanced Research Projects Agency (DARPA) Defense Sciences Office awarded contracts to Tufts University; the University of Chicago; Boston Dynamics, working with the Massachusetts Institute of Technology (MIT); and a team led by iRobot Corporation, Bedford, Massachusetts, which includes experts from Harvard University and MIT. These creative teams have begun their exploration into the world of designing a robot that resembles a furry mouse more than R2-D2 of Star Wars movie fame.

Phase one of the effort is dedicated to research. Although the agency has allowed up to two years for this stage, it would prefer to see results sooner.

Milestones for this first stage include designing a chemical robot (ChemBot) with a circumference of less than 12 inches, diameter of 4 inches and volume of slightly less than 30 cubic inches. The ChemBot then must be able to achieve a tenfold reduction of its largest dimension. The ChemBot must be able to travel a distance of nearly 16.5 feet at a speed of approximately 10 inches a minute. Finally, it must be able to traverse through an opening of less than one-half inch of arbitrary geometry, return to its original size and shape in 15 seconds, and perform a function or task using an embedded payload. The agency hopes to be able to demonstrate these capabilities by the end of this phase of the project.

At this point in the program, DARPA has not specified the need for robustness in terms of indestructibility; however, it must be able to survive typical military operational environments, which include exposure to humidity, rain and extreme temperatures. A per-unit cost ceiling has not been specified.

Critical features of the final ChemBot design include a three-dimensional morphing capability and architectures that can sense and change shape in response to the size and shape of openings. The latter could include using a built-in tactile sensing capability. In addition, it will need flexible backbone structures or architectures that can morph or dissolve, then reconstitute. Potential ChemBot payloads must remain viable after traveling through an opening, which means they must be smaller than the largest part of an opening, but novel soft payloads can be larger. ChemBot power requirements must be modest, and the device can be self-powered, self-consuming or an energy scavenger. The final device can be either semi-autonomous or user-controlled.

Dr. Mitchell Zakin, program manager, ChemBots, DARPA, explains that as a chemist and inventor, his job at the agency is to solve important problems for warfighters. In the case of ChemBots, success would mean providing them with the ability to collect intelligence and reconnaissance information in areas where the access point is less than an inch wide.

“Chemistry, employed in concert with other sciences and disciplines, offers tremendous potential for game-changing technological surprises for our armed forces. Our armed forces often use robots to gain access to or intelligence about an area controlled by our adversaries. On today’s battlefield, they frequently confront situations in tight, closed-in areas to which access seems impossible. What if the only way to get into an area or building is through a small hole in a wall or window or through a crack under a doorway? Mechanical and electronic robots can’t do that. Could there be such a thing as a ‘soft’ robot?” Zakin relates.

Although the concept of a soft robot may be new, it is not an entirely new field for Zakin. Before joining DARPA as a program manager in 2005, he developed the Life Sciences business area at Physical Sciences Incorporated, Andover, Massachusetts, where he applied innovative chemistry to medical and defense problems.

He points out that interesting examples of the ability to squeeze through tight spaces exist in nature. “Many soft creatures, including mice, octopi and insects, readily traverse openings barely larger than their largest ‘hard’ component, through a variety of reversible mechanisms. These mechanisms include using elastic materials to twist, crumple and bend with many degrees of freedom, utilizing the flexibility of the musculoskeletal structure to squeeze through openings, and exploiting reversible changes in modulus to achieve dimensional reductions. These soft invertebrates typically move by crawling—earthworms and caterpillars, for example—pedal waves, such as snails and slugs, or cilial motions, and utilize means such as gripping, hooking and suction to ensure sufficient traction with the terrain,” Zakin relates.

“Initially, we’ll focus on the development of novel materials, material systems and/or robot architectures that can both move and morph under the influence of an appropriate driving force, for example electromagnetic, acoustic or chemical, as well as the demonstration of a rudimentary meso-scale ChemBot,” he adds.

iRobot’s PackBot with the ICx Fido explosives detector features a manipulator arm with flexible elbow joints and an 87-inch reach.

Dr. Chris Jones, research program manager, iRobot, notes that DARPA is going back to the drawing board on the mobility of robots with this project. He agrees with Zakin that work must begin with basic material science and physics to design and develop new components for robotic systems. “The vision is to build a robot that is roughly the size of a soda can, but that is completely squishy to the point that it could squish through a hole the diameter of a quarter. So you can’t have large, chunky pieces of metal in it. It really needs to be completely soft, flexible. Imagine a water balloon or something like that. You really can’t meet that vision by using traditional technologies,” Jones says.

And it is the creation of these building blocks of the ChemBot that likely will be the most challenging to the experts working on this project. Jones relates that, for example, it will require advances in material science and physics that can be brought to bear on creating a practicable actuator.

Another challenge will be figuring out how to fabricate many of the needed components. Because the project likely will involve working with new materials that feature very small electronics and embedded sensors, it will require new and novel approaches, Jones notes. “It’s not bolting a couple of parts together. If you have all sorts of weird materials, and you need sensors that you’ve never put together before, how do you fabricate some subsystem?” he says.

Jones believes the third challenge will be systems integration. Once the novel actuators, sensors, communications schemes, processing capabilities and power systems are developed, the quandary will be how to put them in a functional robot. He describes this as an “exciting challenge” for his company.

Tackling these problems on the iRobot team will be technical experts from the company as well as from MIT and Harvard, Jones relates. The diverse group of approximately eight people will include specialists in the areas of chemistry, material science, mathematical topology, computer science and robotics.

Although the company has not created a robot quite like the ChemBot that DARPA seeks, iRobot will be able to draw on some of the inspiration and knowledge it has learned from its design work on other tactical robots, Jones says. The firm’s relationship with DARPA began in the late 1990s when it started developing manpackable unmanned ground vehicles for the military. To date, iRobot has delivered to the military more than 1,600 of these types of platforms that have been part of current operations and have been credited with saving lives.

“We’ve done work before on robots that have a very high degree of freedom. …When you start talking about a very soft robot, you’re essentially talking about an almost infinite number of degrees of freedom or ways in which it could change its shape. …Our experience in building very highly integrated electromechanical robots has forced us to gain an expertise in how to do that very tight integration between the electrical components, mechanical components, sensors, processing and communications,” Jones maintains.

DARPA officials say the results of phase one of the project will determine the specifics of phases two and three.

Web Resources
Chemical Robots Program, DARPA: www.darpa.mil/dso/thrusts/materials/multfunmat/chembots/index.htm
iRobot Corporation: www.irobot.com