A Robust Localization Algorithm for Mobile Robots with Laser Range Finders

Abstract

We present a robust localization algorithm for mobile robots with laser range finders, which takes feature based approach for reliable matching process as well as point based approach for accurate and stable pose calculation. An efficient sequential segmentation algorithm is suggested, which also performs least square fitting processes simultaneously. Hence robust line segment features can be obtained with much less computation time. A novel cost function for pose calculation is suggested and formulated, which makes it possible to compute the location and orientation accurately and reliably. We examined the validity of the proposed algorithm with various simulations and experiments, and revealed its robustness and accuracy compared to other typical localization algorithms. The results show that the localization error of the proposed algorithm is 40% to 80% less than conventional localization algorithms.