An innovative localization system of transformer substation automation robots based on pattern recognition and ICP algorithm

This paper proposes a novel, laser-based localization method for the autonomous mobile robot in transformer substations. Iterative closet point (ICP) algorithm has been widely used in current laser localization system for locating the mobile robot. However, after initialization, current localization system cannot finish positioning without human assistance. There is a growing interest in finding an effective way to improve the stability and correctness of laser system. We describe a method that combines the pattern recognition with laser localization system. Robot positioning in feature map is to search the relative distance and orientation to landmarks.