Notice: This is an archived and unmaintained page. For current information, please browse
  1. Squaring Off With a Robotic Serpent

    Text Based on a NASA Ames Research Center Press Release

    NASA engineers are developing an intelligent robot snake that may help explore other worlds and perform construction tasks in space.

    The robot serpent, able to independently dig in loose extraterrestrial soil, smart enough to slither into cracks in a planet’s surface and capable of planning routes over or around obstacles, could be ready for space travel in five years, NASA engineers predict.

    “The snake will provide us with flexibility and robustness in space,” said Gary Haith, lead “snakebot” engineer at NASA’s Ames Research Center located in California’s Silicon Valley. “A snakebot could navigate over rough, steep terrain where a wheeled robotic rover would likely get stuck or topple.”

    “One of our first steps was to make a simple mechanical test snake, and we constructed it in less than a day thanks to previous work at other labs,” said Haith. “It is a direct model of a ‘polybot’ developed by Mark Yim of Xerox Palo Alto Research Center, Palo Alto, CA, with whom we are cooperating. We have slightly different electronics in our version.”

    “The test snake has a wire that carries communications and power to and from the computer brain,” Haith explained. “All of the identical hinge-like modules are easy to make, and we attached the snake segments together in a chain. It has off-the-shelf hobby motors in its hinged segments that cause it to move. Each of the many motors takes a signal from the snake’s main computer brain,” he said.

    “Our first test robot does what we tell it to do, no matter what the results are. If it comes to an obstacle, the robot will continue to try to go over it, even if the task is impossible,” he said. “We made the first, simple robot because we wanted a working snakebot in a day or two, a robot that would help us to think about how a snakebot could and should move.”

    Robotic serpents can “inchworm” ahead, can flip themselves backward over low obstacles, can coil and can side-wind, Haith said. “Future work will enable the snake to become a mast or a grasping arm. A rover would need to have a dedicated mast and arm that would cost extra weight, money and time.”

    “A snakebot is not as good at some jobs as other robots, but you get a lot more robot for the weight and the money,” he said. “The problem is it’s hard to tell the snakebot what to do. It is a complex robot that must operate independently, possibly far from Earth. Work on our second snakebot is aimed at making it capable of independent behavior.”

    “The key part of what we are striving for in the second snakebot version and beyond is sensor-based control in which the robot uses its sensors to decide what to do,’” Haith said. “We made two little microcontrollers, tiny computers, that we put in each hinged section that also includes a motor, electronics and gears to get the hinge to move to certain positions,” he explained.

    The snakebot will have a main computer that will tell its little computers in each segment what to do in a higher, planning sense. The tiny computers in the segments could provide “reflexes” that take care of simple, but important jobs.

    “In the next couple of months, we hope to simulate the snakebot in a computer program so we can automatically develop computer routines that can control the robot,” Haith said. Engineers have added strain sensors to the robot on metal ribs inside the snake. “They will tell the snake whether or not it is contacting anything, and where and how hard it is touching,” Haith explained.

    “We hope to write software that allows the snake to learn on its own by experience,” he said. “Some lessons we hope it will learn are how to crawl from soft to hard surfaces, and how to go over rough surfaces that have rocks. We even hope to show that it can climb scaffolds and go into cracks. These abilities would help the robot look for fossils or water on another planet,” he added.

    Adaptable, reconfigurable robots like the snakebot may prove useful in furthering the goals of NASA’s Astrobiology Institute (NAI), which include the search for signs of life on other worlds. On Mars, for example, a snakebot equipped with appropriate measuring instruments or sample-collection hardware could make it possible to extend the hunt for evidence of ancient life to areas not accessible by rovers.

    What Next?

    The snakebot can save spacecraft weight because the snake-like design enables the robot to do many tasks without much extra equipment, according to engineers. “One of the many advantages of the snake-based design is that the robot is field-repairable. We can include a bunch of identical spare modules with the snake on a space mission, and then we can fix the snakebot much easier than a regular robot that needs specific parts,” said Haith. “Other benefits are: the snakebot can crawl off a spacecraft lander and doesn’t need a ramp, the snake’s moving parts can be sealed inside artificial skin to avoid exposure to the outside environment and the robot can still function, even if one joint freezes.”