A New Fast Histogram Matching Algorithm for Laser Scan Data

摘要

This paper presents a new fast histogram matching algorithm for tracking the position and orientation of robots without the help of any odometers. Histogram avoids extracting geometrical primitives from the sensor data, acts independently from odometers, and is easy to be implemented. Here, the authors employ it to analyze laser scan data. A new concept named "hierarchical histogram structure" is introduced for constructing and matching histograms. This structure makes full use of the information in a laser scan and speeds up the histogram matching to a satisfying level. Computational complexity analysis and experiments show the feasibility of the new algorithm both mathematically and factually.