9) Autonomous map building and localization of a service robot


Figure 14: Localization at the dynamic indoor environment

Three maps are employed for applications. 1) Information map used as a user interface 2) Topological map for global path planning. 3) a local grid map for local path planning and localization by map-matching scheme. Accurate environmental maps can be obtained by developed scan registration and probabilistic accumulation even in unknown environment. For localization, we derive a motion model from experiments and map-matching algorithm is carried out with the probabilisticMonte Carlolocalization framework. Range sensor based navigation scheme was developed, then it was successfully tested in various real environments. An integrated navigation strategy including active behavior control guarantees self-recoverable autonomous navigation.

Posted by ISR