A Kansas State University computer science professor has been awarded a five-year, $450,000 National Science Foundation CAREER Award to research how future teams of robots can better work together.
"At that point (9/11 search crews) had single robots that were basically controlled by humans," Scott DeLoach, an assistant professor of computing and information science, said. "There are many applications for cooperative robotic teams, including search and rescue, waste site cleanup, extra-terrestrial exploration, military operations, transportation and factory automation, uninhabited air, ground, and undersea vehicles for intelligence gathering, as well as a variety of industrial applications."
DeLoach said the future of cooperative robotics lies in autonomous cooperative robotic teams, which can operate without direct human intervention. Unfortunately, cooperative robotic teams are, by nature of the problems they are employed to solve, susceptible to loss of individual robots or functionality, which can significantly reduce the team's ability to accomplish its goals.
According to DeLoach, most cooperative robotic teams are designed to work with a limited set of configurations. Even when these teams possess the capabilities to accomplish their goals, they are limited by their own rigid organizational structure.
"However, we will not necessarily send robot teams out to do jobs such as exploring Mars totally on their own," DeLoach said. "For most applications, some level of human control will always be necessary. Right now, for each Mars rover, there is literally a team of people controlling each robot. To control uninhabited air vehicles, the Air Force has a team of four or five people controlling one unpiloted plane.
"We would like to invert that relationship. We would like to have one person control a team of robots."
With a team of robots, DeLoach said scientists could make less expensive robots that do one or two kinds of things, but more importantly, can work together. If one is lost it is not catastrophic.
"The idea is that you need to give the robots certain capabilities as well as the knowledge of how to work together to carry out their mission," DeLoach said. "When one of them fails or for some reason they cannot achieve their team goals, then they can get together and they have enough intelligence to say 'I can't perform this particular role anymore.' Then the team can choose another robot to take its place."
DeLoach hopes to give the robots enough information so that the team can organize and reorganize in order to continue to perform their mission. He one day envisions having one or two humans controlling a whole team of robots, instead of the other way around. DeLoach hopes to establish organizational reasoning as a key component in a new approach to building highly robust robotic teams.
"In order to reach this goal, organizational knowledge and reasoning techniques must be incorporated into practical methods and tools or designing and developing real world cooperative robotic systems," DeLoach said. "We want to put theory into practice; we want to determine how to capture and use organizational knowledge and reasoning techniques in designing and developing realistic cooperative robotic systems."
DeLoach admits that meeting this goal will require advances across several fronts such as technology, tools and methods and education. DeLoach said that the award will also allow him and his students to participate in a series of activities designed to impact pre-college students from elementary to high school. These activities include participation in K-State summer programs such as the Engineering Summer Studies Institute, Exploring Science Technology and Engineering, and Girls Reaching our World workshops to motivate pre-college students to select science and technology majors and to reach to specific underrepresented groups. In addition, DeLoach plans to expand current outreach efforts in the form of the 'Robot Road Show,' which takes robots to elementary students to motivate interest in science and engineering related fields.
"The actual research results will also have a broad impact to society as a whole," DeLoach said.