By tracking the joints' angles and knowing precisely the camera's fixed position on the hand, a simultaneous localization and mapping approach (SLAM) enables the system to produce a high-quality map even if the camera is moving very fast or if some of the sensor data is missing or misleading, reported the researchers.
This helped them better determine the pose of the camera, not only to improving the accuracy of the 3D map but also locating more precisely the robot's hand within the newly modelled space.
Eventually, this approach could help robots reach into a tight space or pick up a delicate object, without prior knowledge of the landscape and without relying on too computational intensive sensor data crunching.
Using a small depth camera attached to a lightweight manipulator arm, the Kinova Mico, the researchers reconstructed a 3-D model of a bookshelf, producing reconstructions equivalent or better to other mapping techniques.
Visit the Carnegie Mellon University Robotics Institute at https://www.ri.cmu.edu