Merging maps via Hough transform

We present a recently developed algorithm for merging multiple occupancy grid maps computed by multiple robots independently exploring a shared indoor environment. The algorithm exploits the well known Hough transform in a novel way in order to produce a set of ranked roto-translations aimed to overlap the partial maps provided as input. In this paper, after having briefly summarized such method, we investigate the impact on the performance of different variations of the Hough transform. In particular, we are interested in determining the repercussions in terms of accuracy and computational time when only a subset of points is used to compute the transformation. Results are analyzed while merging maps produced by two robots exploring an indoor environment, and also using public available data sets. It turns out that the proposed method is robust and positively influenced by the use of more refined approaches to compute the Hough transform.