Mobile robot localization in a structured environment cluttered with obstacles

The authors describe a method for mobile robot localization in an a priori known structured environment cluttered with unknown obstacles. The method is nearly independent of the number of obstacles encountered. An environment model based on ultrasonic range readings was built and compared with the reference model using clustering techniques. The method is very robust and works well if only a few parts of the environment are known. Results from an implementation on the mobile robot RAMSIS are presented.<
>