Visual Trajectory Based SLAM

SLAMstands for Simultaneous Localization AndMapping. It is a fundamental topic in Autonomous Systems and Robotics as it represents one of the most basic skills that any robot requires in order to be truly autonomous. This skill will allow a robot placed in an unknown environment at an unknown location to simultaneously build a consistent spatial representation (a map) and determine its location within this map. The classical approach to solve the SLAM problem was presented long ago [21] and many alternative solutions have been published since. The objective of any SLAM algorithm is to estimate as accurately as possible both the map and the robot location and their uncertainties based on the information conveyed by one or more sensors. The potential application of SLAM is incredibly wide, ranging from indoor and outdoor robots to underwater and aerial vehicles. Despite the fact that the SLAM problem is considered theoretically solved, many new developments are published every year that aim at the definition of a simple, practical and affordable solution, and try to overcome many of the practical issues that remain unsolved