Learn more about the Mars rover landing.
Get more information about Raj Dasgupta's robotics research.
Find out about UNO's summer robotics workshops for junior high and high school students.
* * *
As NASA prepares for the Sunday landing of its most sophisticated Mars rover yet, University of Nebraska researchers are working on the next generation of robots that could go even deeper into space.
Someday some distant asteroid or faraway moon could be explored by self-reconfiguring robots — think Transformers — such as those being developed by a University of Nebraska at Omaha computer science professor, a University of Nebraska-Lincoln engineering professor and their two teams of researchers.
“What we're doing is changing the paradigm to adapt to environments we don't know as much about,” said Carl Nelson, the UNL engineering professor. “We're not designing to replace the robots NASA's already building. This is a different system, with different capabilities that could expand upon the vision for space exploration.”
The Nebraska research comes in an area of study that has the potential to transform space exploration in coming decades, said Dimitri Zarzhitsky, a robotics software specialist at the Jet Propulsion Laboratory in California, where the Mars rover was designed and built.
“We're all very excited about this,” he said.
To be sure, the $2.5 billion Mars rover, named Curiosity, is one awe-inspiring piece of equipment. NASA informational materials describe it as carrying “the most advanced payload of scientific gear ever used on Mars' surface. ... Its assignment: Investigate whether conditions have been favorable for microbial life and for preserving clues in the rocks about possible past life.”
“It is a mind-blowing machine,” said Michael Watkins, a mission project engineer who also is at the Jet Propulsion Laboratory.
The 2,000-pound rover was launched from Cape Canaveral, Fla., in November and is scheduled to land Sunday evening on Mars.
In comparison, the experimental robots of UNO computer science professor Raj Dasgupta seem almost rudimentary. Some are no bigger than hockey pucks; others are about the size of a milk crate. They look more like Roomba robotic vacuum cleaners than C-3PO or R2-D2 of “Star Wars” fame.
In fact, Dasgupta buys robots off the shelf.
“We buy robots, we don't build robots,” Dasgupta said while giving a tour of his laboratory at UNO's Peter Kiewit Institute. “We're interested in computer science — in putting the brains on the robots.”
Artificial intelligence is Dasgupta's area of specialty. On the UNO faculty since 2001, he became intrigued by robots in 2004. He obtained a $100,000 grant from the Department of Defense's Naval Air Systems Command to explore the possibilities of using multiple robots, moving in coordination, to detect land mines in war zones.
Dasgupta reasons that a swarming team of robots might do a better job of finding mines than a single robot, despite the sophistication of its equipment.
His work on COMRADES, the mine-detecting robots, continues at UNO under a $1.3 million grant awarded by the Office of Naval Research.
Last year he and Nelson landed a NASA grant for their $1.1 million ModRED project. To put it briefly, Dasgupta's team is working on the robot's brain and Nelson's team is developing its physique.
Both Curiosity and the ModRED robots are machines with autonomy — that is, they “think” for themselves as conditions dictate. Distance makes a remote control device unworkable.
The rover Curiosity is fed an upload of instructions each Martian day, taking into account factors such as safety, power reserves and communication availability. A ModRED robot's programming would allow it to respond in a more adaptive way to unforeseen situations.
Plutonium-powered Curiosity may travel more than 12 miles as it makes its way to the middle of Mars' Gale Crater and climbs Mount Sharp, the informal name for a three-mile-high mountain at the crater's center.
It will sample layers of rock during the course of a Martian year — 687 Earth days — and analyze them in a self-contained laboratory. It will transmit video back to an Earth-based team of 350 international scientists.
Zarzhitsky said the Nebraska research and that of other scientists exploring similar concepts could be a game changer.
Instead of a single car-size machine equipped with an array of tools, a handful of small modular robots could be dispersed over an extraterrestrial body.
The data gathered by each would be combined and shared. The robots could assemble and reassemble themselves to form tools and solve problems. They might join into a self-made bridge to cross a ridge or a ravine; they might form an inchworm configuration to move over rough terrain; they could link into a wheel to roll across the landscape. They might even stack themselves into a crane to receive deliveries from space.
The specialties of the 10 members on Dasgupta's research team, who range from undergraduate students to post-doctoral researchers, give a glimpse of the complexity of programming robots to make decisions.
Undergraduate Brad Woosley, for example, works on the algorithms for robot motion planning. Postdoctoral researcher Jihyun Yoon works on systems using global positioning satellites to allow the robots to know their location. Doctoral student Janyl Jumadinova is writing her dissertation on “fusing,” developing a method to combine data from multiple sensors on different machines into a cohesive whole.
Jose Baca, who earned his doctoral degree at a university in Spain and moved to Omaha to join Dasgupta's team, helps program the various motions of the Transformer-like robots.
Dasgupta says it likely will be 2018 before the ongoing research is completed, and there's a decision from NASA and the Department of Defense on the next steps.
Nelson said to stay tuned for more Nebraska robots.
“We are going to grow this project,” he said. “Modular robots and the space exploration robotics will continue as a theme in our lab.”
This report includes material from the Los Angeles Times.
Contact the writer: