Qualitative homing

A homing problem is addressed in which a robot is given the bearings of the landmarks at the robot´s current location, the bearings of the landmarks at the target location, and the correspondences between landmarks across the current and target locations. It is required to make an admissible movement, one that takes it closer to the target location. It is shown that an admissible homing movement can be determined in a qualitative way from angular constraints derived from inexact landmark bearings. The algorithm can be implemented with simple integer counting operations, without using costlier operations such as multiplication or matrix inversion. The algorithm (in conjunction with the consistency filtering algorithm) can handle uncertainty from two sources: incorrectly matched landmarks and inexact landmark bearings. This efficiency and robustness make it suitable for real robots running in real time