According to Dr. Jim Overholt, director of RSJPO’s Joint Center for Robotics, the primary task to look at how warfighters are using what they already have, understand what they want next and then determine how to meet those requirements. But what is being demanded for combat applications also is driving the future structure of the overall robotics industry.
“Plug-and-play, from an industry standpoint, will require the establishment of metrics and standards, so if we do finalize on a series of robotic platforms, the payloads going onto those will not require a lot of different tools,” he said. “There are many problems to solve, which is not necessarily a bad thing. It engages the community, with industry and academia around the world tackling the many technology issues involved in the future of robotics.”
As a result, Braden added, industry, academia and government have begun moving in the same direction for all types of robots.
“We’re all looking at standards we can share or are similar, especially when it comes to platform interoperability,” he said. “If I’m using a small UAV and a small ground robot, I’d like to have only one soldier dealing with both and have some quick situational awareness off a single controller unit.”
A central concern is determining the best method of locomotion – wheels, tank-like treads or legs. Dr. Robert Mandelbaum, program manager on numerous robotics efforts at the Defense Advanced Research Projects Agency (DARPA), is focused on walking robots.
“When building robots that are meant to take the place of humans in some activities, such as going into a dangerous locale, you can’t redesign a preexisting space intended for humans; you need robots that can operate in that space, with arms and legs on the same scale as humans,” he said. “Honda, Toyota and others have humanoids now, but most only address one of four critical criteria – mobility, manipulation, cognition and power. The vision of creating humanoids in the future requires first conquering each of those four individually.
“LS3 (Legged Squad Support System, a program Mandelbaum recently launched) is specifically targeted at not having direct human control. Full autonomy is not quite ready yet, so that vehicle will have to follow a leader, but that leader won’t need to control it with a joystick.”
Other major challenges include creating a self-sustained system that can carry its own fuel source and power system and giving robots the ability to manipulate their environment.
“We’re working on a study in autonomous manipulation,” Mandelbaum said. “The idea is to have a vehicle control its own arms and hands to pick up objects and operate tools, such as a screwdriver or shovel, that are made for humans. A lot of work has been done in this field over many decades and we are looking to galvanize all that into a very focused program.”
Practical robotic manipulation, especially autonomously, requires major advances in artificial perception – the ability not only to see but to understand the environment around it.
“Essentially, the metric of success is how often does a human have to intervene to get a robot out of trouble,” Mandelbaum said.
Much remains to be done before ground robots are ready to move and work safely in urban settings, such as a busy marketplace filled with animals, other vehicles, a maze of stalls and what Braden terms “probably the most difficult object on the face of the Earth: The moving human in the equation; not being able to solve how to deal with a 6-foot object – some friendly, some neutral, some enemy – that moves around in an unpredictable manner.”
As to what comes next and how to get there: “People talk about robotics and look for applications, sort of a technology push,” Mandelbaum concluded. “My approach is to look at a user pull – what do the services really need.”
See more on unmanned ground vehicles in J.R. Wilson’s “Unmanned on Land: UGVs Today” in the upcoming Year in Defense Land Forces Edition.