With each passing year, it seems robots evolve faster than humans do. Last week, a group of students at the University of Freiburg's humanoid robots lab in Germany detailed how they gave robots the ability to maneuver extremely difficult obstacles, such as stairs and ramps, without assistance.
To achieve this relatively new level of robotic maneuverability, the researchers implemented a "2D laser scanner, a monocular camera, an inertial measurement unit, and joint encoders" into a Nao robot, according to a research document (PDF).
For a robot to climb stairs, it first must get a good grasp of its surroundings. The Nao robot creates a 3D model of a staircase using laser range data, which serves as a digital map. Before the climbing process begins, the Nao robot uses a camera to detect the edges and contours of the stairs. With the information collected, the robot estimates its location on the stairs and positions itself to take a step.
Descending ramps appear to be much easier for the enhanced Nao robot, as the researchers only needed a camera and inertial data to achieve success. The robot "locates the beginning of the ramp using visual observations, walks down with regular corrections based on the inertial data, and finally determines the end of the ramp by detecting the ending edge before exiting the ramp," the humanoid robots lab explained in a statement.
University of Freiburg students Stefan Obwald, Armin Hornung, and Maren Bennewitz worked on this robotic research with Jens-Steffen Gutmann, a former Sony employee who now works with Evolution Robotics.