Robot mapping with range data

Explores methods by which range-sensor data can be used to generate maps for robot navigation. The authors assume an idealized 2-D environment consisting of flat walls, a point robot, and a perfect range measuring system. The purpose of this idealization is to understand how much effort is to be expended to accomplish the mapping of a region. Several theorems are developed that place limits on the number of robot movements required to completely explore obstacles and boundaries. A mapping procedure is described that works within the limits indicated by these theorems. Results of applying this procedure in a simulated environment are also shown.<>