Robust particle filter localization by sampling from non-corrupted window with incomplete map

We present robust localization algorithm using particle filter for mobile robots. The blueprint of building that cannot completely represent the real environment or a simple map of environment can be used in this work. With this incomplete information, it will not be easy to estimate the pose of robots because un-modeled obstacles will make the sensor observation inconsistent. Proposed algorithm provides robustness in such a dynamic environment because the effect of corrupted sensor observations can be suppressed by a selective update or a sampling from non-corrupted window. Results of experiments and simulations show the robustness of proposed algorithm.