Positioning and navigation of mobile robot with asynchronous fusion of binocular vision system and inertial navigation system

Abstract

Binocular stereovision–based positioning and inertial measurement–based positioning have their respective limitations. Asynchronous fusion of a binocular vision system and an inertial navigation system (INS) is therefore introduced to global positioning system–denied environments with a fuzzy map. It aims to provide a sequential and progressive update with regard to mobile robot positioning and navigation. The system consists of two off-the-shelf cameras and a low-cost inertial measurement unit (IMU). The localization procedure fuses the inertial data from the IMU and the absolute position data from the binocular vision system based on corners. The main contribution of this article is a novel fusion method adaptive to different data rates at which the two modules operate. Utilization of an asynchronous Kalman filter is proposed to fuse the results from the two modules, which can achieve intermittent correction for INS localization. Experiments were carried out in an indoor laboratory environment where dynamic tests validated the reliability and effectiveness of the proposed asynchronous fusion algorithm.