Visual-Inertial Fusion for Indoor Autonomous Navigation of a Quadrotor Using ORB-SLAM

Abstract

This brief addresses the accuracy of the pose estimation of a Low-Cost Quadrotor using visual SLAM and sensor data fusion for making an autonomous indoor navigation in a known GPS-denied environment. To perform this approach, two kinds of modules are employed. For the former, a Simultaneous Localization and Mapping (SLAM) system is used. This SLAM system uses Oriented Fast and rotated BRIEF (ORB) features, also known as ORB_SLAM. The second module is Extended Kalman Filter (EKF) that is applied for combining the obtained pose from ORB_SLAM and Inertial Measurement Unit (IMU). Since in this paper the SLAM process is performed by the frontal monocular camera of Quadrotor, so developing scale factor estimation in order to calculate the scale of the map with minimum error is necessary. In addition, a Proportional Integral Derivative (PID) is utilized to conduct the maneuver of the robot. All of these processes are done by sending data via Wi-Fi to a ground station in order to perform the data processing. To show the performance of the proposed algorithm, two experiments are carried out on an AR-Drone 2.0 quadrotor. The first one is position holding that shows the robustness of the estimated position of the system in hovering situation after take-off. The second one is an indoor trajectory tracking that illustrates the difference between the downward camera output position and the estimated position.