Quaternion-Based Unscented Kalman Filter for 6-DoF Vision-Based Inertial Navigation in GPS-Denied Regions

Abstract

This article investigates the orientation, position, and linear velocity estimation problem of a rigid-body moving in 3-D space with six degrees-of-freedom (6-DoF). The highly nonlinear navigation kinematics are formulated to ensure global representation of the navigation problem. A computationally efficient quaternion-based navigation unscented Kalman filter (QNUKF) is proposed to imitate the true nonlinear navigation kinematics and utilize onboard visual-inertial navigation (VIN) units to achieve successful Global Positioning System (GPS)-denied navigation. The proposed QNUKF is designed in the discrete form to operate based on the data fusion of photographs garnered by a vision unit (stereo or monocular camera) and information collected by a low-cost inertial measurement unit (IMU). The photographs are processed to extract feature points in 3-D space, while the six-axis IMU supplies angular velocity and accelerometer measurements expressed with respect to the body frame. Robustness and effectiveness of the proposed QNUKF have been confirmed through experiments on a real-world dataset collected by a drone navigating in 3-D and consisting of stereo images and six-axis IMU measurements. Also, the proposed approach is validated against state-of-the-art filtering techniques.

For a demo of the quaternion-based navigation unscented Kalman filter (QNUKF) and the experimental setup, watch the embedded YouTube video on this page. Use the links below to access the full IEEE paper and the scripts in the QNUKF repository.