Location and robotics (Q3147669)

In this chapter, a very interesting but marginal problem from the point of the classical location theory is dealt with. Whereas in a classical location problem a location for a facility is to be determined, so that some benefit function is minimal, in the considered case a position of a mobile robot should be recognized. The robot is located somewhere in a region of the two-dimensional space and it is equipped with a general map of the region. Making use of sensors, the robot is able to scan vertices of a visible polygon of the region and based on this information, its program should determine the point in the map, which fits the current position of the robot. The authors introduce the geometric theory, which concerns this problem of visibility and recognition. They point out, how to create so-called visibility skeletons, which can be recognized having information of their vertices. The authors pay great attention to the complexity of the associated computational process and to the selection of the most suitable distance for the process of recognizing. They discuss the Hausdorff distance and the polar coordinate metric and they select the second one as the most convenient. Finally, the resulting robot location program is described and the results of experimental tests of the associated implementation are reported. For the entire collection see [Zbl 0988.00044].