ABERDEEN PROVING GROUND, Md. (Jan. 31, 2013) -- Natural disasters like earthquakes, hurricanes and tsunamis can unveil points of weakness in man-made infrastructure, and now robots are being called in to help.
Scientists at the U.S. Army Research Laboratory, or ARL, and Edgewood Chemical Biological Center, or ECBC, here, are developing suction cups that could be placed on robots designed to perform tasks in unstructured and contaminated environments.
The self-sealing suction cup is a collaborative project between the two Army laboratories and the University of Maryland, where Chad Kessens, a robotic manipulation researcher for ARL, is pursuing his doctorate in mechanical engineering
As part of the program, Kessens tested the limits of robotic grasping by developing a new suction technology to expand the range of graspable object shapes and sizes. An expanded grasping capability could improve how emergency responders observe areas of devastation by increasing the effectiveness of robotic operations while reducing human risk.
The collaborative effort between ARL and ECBC demonstrates a desire to improve technology, share resources and utilize the expertise of personnel working in laboratories across the U.S. Army Research, Development and Engineering Command.
"Manipulation of unknown objects is a very difficult task for a robot. In traditional applications, the robot would have a model for the object it wants to pick up and would then know how to pick it up. The self-sealing suction cup design could enhance grasping technology, making grasping of unknown objects easier," Kessens said.
On Dec. 7, 2012, a 7.3-magnitude earthquake off the coast of Japan, shaking buildings in Tokyo and caused a small tsunami to revisit an area that was destroyed by the Fukushima-Daiichi disaster in 2011.
"When something like Fukushima happens, it would be very useful if the robots that are sent in could perform some sort of manipulation activity like closing a valve, recovering an object or operating a tool in a contaminated area," Kessens said. "Even opening a door or a hatch could allow the robot to better observe what's going on inside the reactor while eliminating the risk of exposing people to radiation."
Inspired by the octopus, Kessens' design features a self-sealing component that imitates the sea creature's ability to individually actuate suction cups based on the object it wants to pick up -- from large and small fish to rocks and even a jar of peanut butter. Though suction technology has been applied to the robotics field since the 1960s, it has been limited in its scope and practical only for objects with a specific size and shape.
According to Kessens, a traditional suction grasper uses one vacuum pump as a central suction source, which limits the effectiveness of the technology for grasping if some cups on the grasper do not attach to a given object, creating leak points where air enters at the point of engagement.
Instead, Kessens is modifying the technology so a robot could grasp a large range of items by maximizing the strength of the suction. The self-sealing suction cup features a plug that sits nominally in the suction inlet. When the source pump is turned on, the plug of any cup not in contact with an object gets sucked in, sealing itself. This increases the pressure differential and strengthens the suction capability of the cups that are engaged on an object. The design also uses passive reaction forces that cause the cup to activate and open when the lip contacts an object, breaking the seal to initiate suction.
While Kessens has demonstrated remarkable success in air, he believes his design might work even better underwater.
"There are several advantages. Objects are typically not porous and there are generally smoother surface features underwater. There are also higher pressure differentials," Kessens said. "When you are operating in the atmosphere using air, you're limited to the atmospheric pressure for how much force you can generate from the suction cup. But when you go underwater, you have all of the extra pressure from the depths of the sea so that gives you more force to utilize for the effectiveness of the cups."
The joint project is in the middle of its lifecycle and comprehensive prototype testing still needs to be done Kessens said. While the ARL scientist provided the concept and design, ECBC generated the prototypes through its expertise in rapid prototype manufacturing. According to Brad Ruprecht, engineering technician and senior model maker in the Advanced Design and Manufacturing Division of ECBC's Engineering Directorate, the biggest challenge was determining how small the cups could be while still making them functional. Part of the process was ECBC's design capability, including experienced engineering personnel and advanced equipment, to craft a prototype using a multi-material 3D printer.
"What I loved about the project is Chad came to ECBC first and foremost because we had the multi-material machine, and he leveraged that to get a working model right off of the 3D printer," Ruprecht said. "It has levers and springs and everything else needed to be a working prototype, and it's worked very well for him. He's received a lot of good data from it and is definitely moving forward with his designs."
Ruprecht used the 3D printer to create prototypes composed of elastomeric materials such as a liquid photo polymer that solidifies into plastic once exposed to ultraviolet light, and more rigid materials like nylon. In about 20 minutes, the ECBC engineer could produce 20 prototypes of different shapes and sizes.
One of the challenges for Ruprecht was handling the small parts of the suction cup like the central plug crucial to the design. The 3D printer fills the space, or clearance, between parts with support material that stabilizes the cups during printing. This material, however, needs to be removed upon final production, forcing Ruprecht to be creative when removing the support material without destroying the prototype itself.
"When the suction cup shrunk in size, there was a huge challenge in getting the support material out of the clearances and overhangs without destroying it because it was very delicate at that point," Ruprecht said. "Eventually we bought a Waterpik, and it was a nice, fine-pointed stream of water that could spray out the support material. Especially on the first iteration of the prototype, there were a lot of delicate parts."
Now on its fourth iteration of the design, the self-sealing suction cup ranges anywhere in size from the palm of a hand to the point of a fingertip. Four fingertip cups can pick up a bottle of wine. The next step is developing a substrate such as a hand or tentacle, where the cups would be located on a robot. Until then, there are several prototypes to finalize the design and conduct testing.
"With 3D printing, you're getting a working ensemble of suction cups right off of the machine with the elastomeric and rigid materials together," Ruprecht said. "But if you were to go underwater with it, you probably wouldn't use the same materials. They tend to absorb moisture and degrade faster. You'd want something that is going to hold up to salt water like a thermal plastic."
Though the 3D printer is limited to the materials it was designed to print, Ruprecht said the technology serves the purpose of giving a researcher adequate time to gather large amounts of data from the prototypes. Mass manufacturing for commercial or industry purposes, on the other hand, would more likely use injection molding that melts down any thermal plastic into a mold, allowing the user to select from a variety of materials. According to Ruprecht, injection molding would also be cost effective and quicker to produce on a large scale, a secondary area of expertise for ECBC.