Research in robots that can enter hazardous locales will make these machines the “right-hand man” of humans.
Chun-Sing George Lee, a professor in the College of Engineering, is working on a U.S. Defense Advanced Research Projects Agency-funded project to create a robot called Atlas. Researchers will improve the Hubo II robot created by the Korean company Rainbow and use these improvements on Atlas. One goal is for Atlas to be able to enter disaster sites and perform complex challenges such as climbing a ladder or moving over rough terrain.
“The research has been moving toward this kind of robot,” Lee said.
Lee hopes to use the research on the Hubo II to make an effective mechanical clone of a human. Many types of research can be done on the robot such as analyzing its interactions with humans to its mechanical motions.
“We want to improve its capabilities to function as a human,” Lee said.
According to Lee, the “integrated system for research” has many systems to be improved. Roy Chan, Manas Paldhe and Rami Alazrai, all graduate students in Lee’s lab, are working on different aspects. Paldhe is working on changing the software on the robot from Windows to Linux. Chan is working on how to detect the positioning of the robot and Alazrai is programming the robot to interpret human emotions.
“We want (the robot) to know to predict what is going on and how two humans interact with each other,” Alazrai said.
Andy Park, a graduate student in Lee’s lab, said he works with the movement of robot legs and simulated human-like movement. Despite the 50 degrees of freedom the robot has, it is not close to the 150 degrees of freedom of a human. This makes the goal of climbing a ladder difficult.
“There is a balancing problem we have to resolve,” Park said.
Yung-Hsiang Lu, an associate professor in the College of Engineering, is working to improve the battery life of the Hubo II, which is currently 30-45 minutes. Lu wants to increase the efficiency of the motor system and replace the battery.
“It’s a lot like if you are in a 100-meter sprint or in a marathon,” Lu said. “You use a different strategy.”
Lu is also working on another limitation of the robot: Its processing speed. He said he plans to have some of the processes of the robot run off of cloud software because the system would run faster. This should smooth out movement.
“Mechanical systems have a real time constraint,” Lu said.
The project is more than just a robotics challenge. The developments in improving the Hubo II robot can be translated to other areas in technology as well.
“We don’t want to develop something specific to this but can also be generalized to other uses,” Lee said.