3D Mapping and Navigation for Autonomous Quadrotor Aircraft

Autonomous navigation is a challenging problem in GPS-denied environments. For small flying robots and quadrotors, inherent limitations such as limited sensor payload and small system time-constant add another layer of challenge to the problem. This paper presents a solution for autonomous navigation of a quadrotor by performing autonomous behaviors, such as exploration, returning home, or following waypoints. The solution relies on accurate localization, mapping, and navigation between waypoints. A 3D mapping algorithm is proposed that enables the fast moving quadrotor to make a reliable 3D model. One of the contributions of the paper is the 3D mapping algorithm which is built upon a 2D mapping algorithm and IMU information. Compared with state-of-the-art algorithms, the proposed 3D mapping allows the quadrotor to generate more accurate models with low-bandwidth channels. Additionally, the algorithm allows the quadrotor to have aggressive changes in rotation and translation that occur frequently due to the fast dynamics of the quadrotor. Also 2D and 3D maps are displayed on multiple remote devices, such as tablets. Multiple tests were performed in simulated and real-world environments to show the effectiveness of the proposed solution.