Localization of mobile robot based on fuzzy-adapted Kalman filtering

Abstract

In order to resolve the problem of mobile robot localization with unknown noise characteristics,this paper proposed a mobile robot localization method based on fuzzy-adapted extended Kalman filtering.Combined fuzzy logic and covariance-matching technique together to adjust the measurement noise covariance R and on-line improve the performance of the localization algorithm.Moreover,it used a sensor fault diagnostic and recovery algorithm to monitor the sensors' states and improved the algorithm's robustness.Then applied the algorithm to mobile robot localization with unknown measurement noise characteristics.Experimental results show that this method can effectively reduce the effect of incomplete a prior knowledge of R,and improve the localization accuracy.