ABSTRACT

The simultaneous localization and mapping (SLAM) is a problem of robot localization and also of map building or upgrading the already available map simultaneously or in a boot-strap, or hand-in-hand manner. The SLAM is also very useful for unmanned aerial vehicles, micro- and mini-air vehicles and many other types of autonomous vehicles: from indoor to outdoor robots, underwater vehicles, unmanned ground vehicles and airborne systems, for the latter it is called aerial SLAM. The mathematical models can be considered to be linear for the sake of the simplicity of the exposition and to clearly understand the problem of estimation in SLAM. A mobile robot can traverse through an unknown environment by taking into consideration, the relative observations of the landmarks provided. For robust target tracking in cluttered environments, the multi-hypothesis data association is essential, because it resolves association ambiguities by generating a separate track estimate for each association hypothesis and thereby creating, over time, an ever-branching tree of tracks.