Engineers building robots usually have a choice between making a fast machine or a flexible one. But now engineers at Tufts University are changing the game with "soft robots" who can worm their way into hard-to-reach places very quickly.
How did they do it? By imitating nature. These researchers created robots that behave like caterpillars escaping from a threat, by turning their bodies into wheels and rolling away. This seemingly simple engineering feat is actually very hard - and it's about to transform the way our machines can move.
Mighty Morphin' Caterpillars
The video on the left shows two examples of animals reconfiguring their body structures (a process called morphing) to form a wheel and roll out of harm's way. There is a key difference, however, between the salamanders in the video and the caterpillars. The salamanders rely on an external force, namely gravity, to drive the motion of their wheel. While this particular ability has evolved countless times in nature, the caterpillars exhibit a much rarer adaptation. The creatures themselves are the ones generating the rolling momentum by curling quickly into a wheel. Because caterpillars change the shape of their bodies in less than a tenth of a second, they generate a linear velocity that can exceed 0.2 meters per second, making it one of the fastest self-propelled wheeling behaviors in nature.
In search and rescue situations, robots that are good at accessing difficult-to-reach places tend to sacrifice quickness and efficiency for versatility, while faster-moving robots give up adaptability in favor of speed. By morphing their bodies, animals like those seen above can have the best of both worlds by exploiting the advantage of limbs in one situation and wheels in another. It was this observation that led researcher Huai-Ti Lin, and fellow researchers Gary G. Leisk and Barry Trimmer, to investigate how one might design a robot that mimics this behavior, and develop a better understanding of the underlying mechanics of the caterpillar's ballistic escape method.
Enter GoQBot, a soft robot that takes its name from the progression in conformational changes that it undergoes as it pushes off into a free-wheeling state. GoQBot is 10 centimeters long, giving it dimensions roughly five times that of its biological inspiration, and while it takes about four times longer to accelerate the robot (it weighs in at over 80 times heavier than its caterpillar counterpart), that doesn't keep it from cranking its locomotion speed to over half a meter per second. That's over 20 times faster than its crawling speed.
To build the robot, Lin and his colleagues had to reduce the caterpillar body into its functional components. To simulate the caterpillar's flexible frame, GoQBot's bendy body is made primarily of a high performance, castable silicone rubber called — wait for it — Dragonskin.
A hammer head and two tail skids function like the terminal prolegs and thoracic legs of the caterpillar, respectively, and serve to anchor and stabilize the robot in the momentum-building stage of movement.
To mimic muscle contraction along the length of the caterpillar body, the researchers turned to a shape memory alloy (SMA) called nitonol. SMA's are lightweight, solid-state alternatives to conventional actuators like hydraulic, pneumatic, and motor-based systems, and have the ability to deform in response to heat. The researchers explain:
To mimic muscle, our robots use SMA wire in the form of coils. In SMA coils the relatively small temperature-induced changes in linear wire strain are converted into large displacements similar to that of caterpillar muscles. The coils are activated by resistive heating using pulses of the current that simulate muscle tetanus.
Once the components have been assembled, GoQBot requires a power source. Although radio-controlled, untethered versions of GoQBot have been developed, the low instantaneous power output of existing batteries tend to limit robot-rolling performance. Consequently, the new GoQBot designed for this study is powered by an external power supply through thin wire tethers.
So What's Next?
So what's next for GoQBot and soft robotics? In addition to engineering GoQbot, Lin and his colleagues studied the robot's kinematic and dynamic properties in hopes of better understanding the control issues of ballistic rolling. Their results indicate that there is still much work to be done before we'll be seeing robotics like this in practical applications.
As of today, ballistic rolling is only effective on smooth surfaces, demands a large amount of power, and often ends unpredictably. The future's soft robots will need to be capable of sensing their environments, operating remotely, and self-righting, all of which are goals that Lin and colleagues, who propose the ballistic roll as a viable solution to fast locomotion in robotic devices, believe are well within reach.
Top image is "Curl Up" by M.C. Escher
Stills of GoQBot via Huai-Ti Lin's Research Blog