An enhanced pose tracking method using progressive scan matching

The purpose of this paper is to propose an enhanced pose tracking method using progressive scan matching, focusing on accuracy, time efficiency and robustness.,The general purpose of localization algorithms is to dynamically track a robot instead of globally locating one. In this paper, progressive scan matching is used to promote the performance of pose tracking. Rotational and translational samples are separately generated to accelerate the calculation and to increase the accuracy. Progressive iteration of sample generation can ensure localization to achieve a specific precision. The direction of localization uncertainty is taken into consideration to increase robustness. Nonlinear optimization is adopted to achieve a more precise result.,The proposed method was implemented on a self-made mobile robot. Two experiments were conducted to test the accuracy and time efficiency of the method. The comparison with the basic Monte Carlo localization shows the advantages of the method. Another two experiments were conducted to test the robustness of the method. The result shows that the method can relocate a robot from an inaccurate place if the offset is moderate.,An enhanced pose tracking method is proposed to promote the performance by separately processing rotational and translational samples, progressively iterating the sample generation, taking the direction of localization uncertainty into consideration and adopting nonlinear optimization. The proposed method enables a robot to accurately and quickly locate itself in the environment with robustness.