NASA robots

NASA robots are robotic devices used to aid, augment, or substitute for astronauts in order to do difficult or rote tasks such as repairs in dangerous environments (such as those with radiation or micrometeorite risks), routine procedures (video capture), etc.

Robonaut

Robonaut resting on a Segway HT.

Main article: Robonaut

Robonaut is a joint DARPA–NASA project designed to create a humanoid robot which can function as an equivalent to humans during extra-vehicular activity (space walks). There are multiple versions of this robot using various locomotion methods. Some versions of the robot use the Segway HT for locomotion[1]. Robonaut uses telepresence and various levels of robotic autonomy. While not all human range of motion and sensitivity has been duplicated, the robot's hand has fourteen degrees of freedom and uses touch sensors at the tips of its fingers.

Mobile Robot Programming Toolkit

The Mobile Robot Programming Toolkit (MRPT) is a cross-platform and open source C++ library aimed to help robotics researchers to design and implement algorithms related to Simultaneous Localization and Mapping (SLAM), computer vision and motion planning (obstacle avoidance). Results obtained from this project have appeared in some of the major robotics journals and conferences^[1].

MRPT is open source and distributed under the GPL.

Some features included in the project as user-applications:

• Visualization and manipulation of large datasets.
• Off-line SLAM algorithms, using ICP, Extended Kalman filtering, or a Rao-Blackwellized particle filter.
• Grabbing datasets from robotic sensors.

Upon a selection of the individual libraries provided by MRPT, users can develop new applications. This is a summary of the topics covered by these libraries:

• mrpt-core: Linear algebra, statistics, SLAM algorithms, 3D geometry, GUI classes for 2D and 3D visualization, serialization, etc.
• mrpt-hwdrivers: Interfaces to robotic sensors: ARIA-compatible robots, GPS, Hokuyo laser scanners, etc.
• mrpt-reactivenav: An implementation of obstacle avoidance.
• mrpt-hmtslam: An implementation of Hybrid Metric-Topological SLAM.

Representation of metric maps

Through polymorphism in the design of metric-map classes, observations (such as a laser scans) can be inserted into a grid-map or a map of points , or both simultaneously, transparently to the user.

The following representations of metric maps are implemented:

• Occupancy grid maps.
• Point maps.
• Landmark maps: discrete elements are 3D points sensed through range and bearing. For example, visual landmarks.
• Beacon maps: elements are also 3D points, but sensed by means of range-only devices.
• Coloured point maps.
• Gas concentration maps.
• A "multi map", collections of any of the other maps, behaving as a single map.