Mobile robot pose tracking by correlation of laser range finder scans in Hough domain

Abstract

This paper proposes a correlation scan matching method based on matching consecutive readings of laser range finder. The method uses the Hough transform and the Hough spectrum for determining the relative rotation. By calculating also translation in the Hough domain, the method can also be successfully used in non-perpendicular environments. This paper also looks at validity of utilizing the probabilistic Hough transform with regard to the complexity and the accuracy of the algorithm. Proposed scan matching methods are used for mobile robot pose tracking. Developed algorithms are validated by simulation and experimental tests.