Carnegie Mellon's Solar Powered Robot Demonstrates Concept That Could Lead To Long-Term Exploration Of Planets And Moons

Late last month, the robot, named Hyperion, successfully completed field experiments on Devon Island in the Canadian Arctic. It tested the concept of Sun-Synchronous Navigation, a technique that involves tracking the sun while exploring terrain. Hyperion performed navigation experiments during a period of 24-hour sunlight, exploring the terrain it encountered while simultaneously monitoring its solar panels to ensure that they collected enough energy to complete each segment of its planned traverse. During the initial experiment on July 19, Hyperion traveled 6.1 kilometers and made history by circumnavigating the Von Braun Planitia, an area near Devon Island's Haughton Crater. The crater is the subject of ongoing study by the NASA Haughton-Mars Project because of its apparent similarity to the terrain on Mars. According to project manager David Wettergreen, a Robotics Institute research scientist, about 95 percent of the circuit was completed autonomously, with the remainder under remote supervision. The experiment began and ended with the robot's batteries fully charged and ready to continue operation. As it encountered the unknown terrain, Hyperion sometimes fell behind its timetable for operation, but each time it caught up when it emerged into a more easily navigable region. "The ability of the robot's perception and navigation system to find routes was very impressive," said Wettergreen. "Analysis of telemetry recorded from Hyperion will reveal the thousands of obstacles it detected and evaluated, the tens of thousands of steering corrections and the statistics of planned versus actual distance and power. Qualitatively, Hyperion wiggled through some pretty tight spots."

In the extended experiment, which ended July 22, Hyperion covered more distance-9.1 kilometers, traveled through rougher terrain, including scree slopes and mud flats, and was challenged with a mission plan that at times put desired goal locations in conflict with the position of the sun.

Wettergreen said that the extended experiment intentionally pushed the limits of Hyperion's capability to find where further research is necessary. In this experiment, Hyperion had greater difficulty due to communication drop-outs, areas of extremely ruggedterrain, and dazzling of its stereo cameras by the sun. In the end, however, it arrived at its destination on schedule with charged batteries.

In only one instance was manual intervention required to correct a steering problem. "There is more work to do to move new technologies from research into development, but the fundamental ideas have been proven," Wettergreen says.

Wettergreen and six colleagues, including graduate students and Robotics Institute engineers, left for Devon Island on July 3. They began to work with Hyperion building up toward the main experiments that were scheduled for July 10. But the time they had was decreased by bad weather and more snow than they had expected during this time of year. They had a narrow window of time to complete their work- -after the snow has melted in mid-July and when the sun begins to drop below the hills toward the end of the month.

"The technology is a grand leap for planetary exploration," said William "Red" Whittaker, the Fredkin research professor in the Robotics Institute and principal investigator of the Hyperion project.

Wettergreen emphasized that Hyperion is a concept vehicle designed to operate only on earth. The objective is to develop technologies, like reasoning about terrain, sunlight and power, that are broadly applicable to robotic explorers but specifically tailored robots that could operate at the poles of the moon or in the polar regions of Mars.

For more detailed information about Hyperion, and the sun-synchronous navigation experiments successfully performed in the Canadian Arctic, see http://www.frc.ri.cmu.edu/sunsync. For more information on the NASA Haughton-Mars project, see http://www.arctic-mars.org