At IROS last year, we met a curious looking fleshy-appendaged explosive jumping robot from the Harvard Microrobotics Laboratory. When we asked the researchers about their plans for the future, they talked about “an entirely different design, and capable of either self-righting or reliably landing upright, enabling multiple successive jumps.” Now the Harvard team, in collaboration with UCSD researchers, has completed that redesign, creating a robot that can jump and land upright, which is totally cool. What’s equally cool is how they did it: with a multimaterial 3D printer that lets them fabricate a robot with the optimal combination of soft and rigid structures.
First, let’s take a look at how this robot jumps. “Explosive” doesn’t just refer to the dynamic nature of the jump; it’s an actual butane-oxygen explosion that gets this thing off the ground. Listen for the pop:
On top of the robot, there’s a core module with a custom circuit board, a high-voltage power source, a battery, a miniature air compressor, a butane fuel cell, six solenoid valves, an oxygen cartridge and pressure regulator, and ducts to move the gas and stuff around as necessary. To jump, the robot first aims itself by selectively inflating one or more of its pneumatic legs to point its body in the direction that it wants to do. Then, it fills its body with a mixture of oxygen and butane and ignites itself, which rapidly expands the flexible bottom of the robot to launch it into the air. The robot can jump more than 20 times in a row, reaching 0.75 meters in height (six times its own height) and 0.15 meres laterally.
What makes the robot so capable and resilient is the multimaterial 3D printing process used to create it. Different parts of the robot grade over three orders of magnitude from stiff like plastic to squishy like rubber, through the use of nine different layers of 3D printed materials. For the main hemispherical chamber in the body of the robot (the bit that pops), for example, the top is stiffer to allow the core module to attach and to help transfer the energy of the explosion to the bottom of the hemisphere, which is the bit that expands downward to launch the robot. Going too rigid would cause the robot to smash into tiny bits on impact, and too squishy would reduce the efficiency of the jumps.
It’s certainly possible to achieve a similar effect if you print out rigid and flexible parts separately and then combined them together somehow, but this causes problems in two ways. First, it’s annoying to have to do it, because it adds steps (and cost) to your fabrication process. And second, everywhere you attach something to something else, you’ve introduced a weakness into your robot through a joint or seam.
This kind of flexibility gradient is something that animals have been doing since there were animals, but it’s new to robots, and has enormous potential for manufacturing beyond robotics, as well:
The fabrication of soft robots using multimaterial 3D printing has numerous advantages over traditional molding techniques. This strategy promotes high-throughput prototyping by enabling rapid design iteration with no additional cost for increased morphological complexity. By allowing designers greater freedom, 3D printing also facilitates the implementation of good robotic design principles, such as modularity and the separation of power and control actuators. Beyond soft robotics specifically, the ability to print a single structure composed of multiple materials enables investigation into mechanically complex designs, without the drawbacks of complicated assembly or inconsistent manufacturing repeatability.
“A 3D-Printed, Functionally Graded Soft Robot Powered by Combustion,” by Nicholas W. Bartlett, Michael T. Tolley, Johannes T. B. Overvelde, James C. Weaver, Bobak Mosadegh, Katia Bertoldi, George M. Whitesides, and Robert J. Wood, from Harvard University and UCSD, will be published in tomorrow’s issue of the journal Science.
[ UCSD ]