Algorithms for Simultaneous Localization and Mapping

Abstract

Simultaneous Localization and Mapping (SLAM) is the problem in which a sensor-enabled mobile robot incrementally builds a map for an unknown environment, while localizing itself within this map. Efficient and accurate SLAM is fundamental for any mobile robot to perform robust navigation. It is also the cornerstone for higher-level tasks such as path planning and exploration. In this talk, I will survey the three major families of SLAM algorithms: parametric filter, particle filter and graph-based smoother. I will review the representative algorithms and the state-of-the-art in each family. I will also discuss issues including submapping, data association and loop closing.