Laser scan matching using multiplex histograms with feature components

A new laser scan matching algorithm is proposed in this paper. By constructing multiplex histograms, the new algorithm offers a simple way for importing geometrical features into the histogram method, obtaining more geometrical information from the laser scan data, and carrying out more accurate matching results. Also, an optimal maximum selection mechanism is employed to filter out the incorrect matching results, which further guarantees the matching precision. With the satisfactory matching results, the new algorithm shows a promising application prospect to the self-localization of robots. Experiments are performed to prove the feasibility and accuracy of the new algorithm.