Abstract
A new position identification method for mobile robot is proposed. The method is to find out own position and orientation of the robot in wide unknown environment by means of horizontal cross section information measured by the range finder and compare it with pre-registered map information. The map information is expressed by five geometrical characteristics of the pair of segments which simplifies the contour of the local environments and the map is registered in the form of binary tree. On-line searching of the position and orientation is then executed on the binary tree. Position identification experiment is performed in a corridor of 50m in length resulted with 57% and 79% correct answer rates both in the condition of single point measurement and double points measurement, respectively. The required times are proved to be 10 second for range measurement by the range finder and 3 second for position identification calculation.
