Robot localization and mapping problem with bounded noise uncertainties

摘要

This paper deals with H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> Filter based SLAM which is also known as minimax filter to estimate robot and landmarks location whose able to stand for non-gaussian noise characteristics. Based on our findings, by selecting appropriate γ and initial state covariance matrix in H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> Filter, the estimation results can show better performance in comparison to the Kalman Filter approach. From the analysis of convergence properties of H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> Filter, it is found that the filter is capable to provide a reliable estimation. Besides, from the simulation results, H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> Filter produces better outcome than the Kalman Filter in the nonlinear case estimation. These condition subsequently provides alternative estimation techniques with the capability to ensure and improve estimation in the robotic mapping problem especially in SLAM.