SENSOR BASED ROBOT LOCALISATION AND NAVIGATION: USING INTERVAL ANALYSIS AND NONLINEAR KALMAN FILTERS.

Abstract

Multiple sensor fusion for robot localisation and navigation has attracted a lot of interest in recent years. This paper describes a sensor based navigation and localisation approach for an autonomous mobile robot using an interval analysis (IA) based adaptive mechanism for the non-linear Kalman filter namely the Extended Kalman filter (EKF). The map used for this study is two-dimensional and assumed to be known a priori. The robot is equipped with inertial sensors (INS), encoders and ultrasonic sensors. A non-linear Kalman filter is used to estimate the robots position using the inertial sensors and encoders. The ultrasonic sensors use an Interval Analysis (IA) algorithm for guaranteed robot localisation. Since the Kalman Filter estimates may be affected by bias, drift etc. we propose an adaptive mechanism using IA to correct these defects in estimates. In the presence of landmarks the complementary interval robot position information from the IA algorithm with uniform distribution using ultrasonic sensors is used to estimate and bound the errors in the non-linear Kalman filter robot position estimate with a Gaussian distribution.