Off-road environment Localization of unmanned vehicles based on Visual-Inertial

Abstract

Nowadays, simultaneous localization and mapping(SLAM), as one of the basic problems of autonomous driving technology, is used in more and more scenarios and fields, such as robot exploration, driverless cars, VR (virtual reality) ,and AR (augmented reality). However, most of the simultaneous localization and mapping studies in the past were conducted in a structured environment, and there was a lot of room for improvement in the performance in the complex environment in the field. In this paper, a visualization monocular visual-inertial odometry consisting of a monocular camera and an inertial measurement unit is introduced, and the front end of the system can operate robustly in the environment of large changes in field light, low texture and bumpy roads by fusing visual information and inertial measurement unit information at the front end, and using nonlinear optimization at the back end. The method in the paper can accurately measure six degrees of freedom (DOF) by combining only a monocular camera and an inertial measurement unit. We verified the performance of the algorithm on the public datasets ROOAD and EuRoC. The results are compared with other state-of-the-art algorithms to show that our method can indeed operate robustly in the field environment and surpass the best performing algorithms in EuRoC.