This card refers to a concept from a branch of mobile robotics known as simultaneous localization and mapping. As a robot navigates through an environment, it is important to understand where it is, and, depending on the application, to generate an accurate map of the environment. A robot can take measurements of features external to it, using camera images or a laser range finder, for example. It may also keep track of where it is by considering its own history of motion. If it relies solely on the latter, the localization error will inevitably accumulate, in much the same way that you may drift out of a straight lane while driving if you close your eyes for a few seconds. The net result is that the robot will think it is in a very different location from where it actually is.
In the technical context, drift is something we generally wish to avoid. A robot will periodically correct where it thinks it is based on its internal sensors with external observations. However if one’s purpose is to move to previously unexplored territory, and get a bit lost, it can be useful to reduce one’s reliance on external references, and follow one’s internal compass. Disorientation breaks old habits and moves you towards new emotional terrain.
Watch a demonstration of the use of simultaneous localization and mapping on a mobile robot.
Return home.