Gray-dynamic EKF for mobile robot SLAM in indoor environment

Abstract

The Gray-dynamic EKF (GEKF) algorithm is proposed to estimate the states of a mobile robot in an indoor environment. First, the gray prediction theory is adopted to predict the states of a mobile robot and the feature positions in the environment; next, based on the predictions, a mobile robot system model is built dynamically; then, the GEKF is used to estimate the mobile robot states and the feature positions. Experimental results show that the GEKF can achieve almost the same estimation accuracy with EKF, while without the need of a fixed system model. To improve the head direction estimation accuracy of the mobile robot, a head direction match algorithm is proposed, and relatively better results are shown by experiments.