Implementation of UAV localization methods for a mobile post-earthquake monitoring system

The goal of this paper is to develop an integrated mobile sensing system that enhances Unmanned Aerial Vehicles (UAV) to provide an autonomous means of locating and mapping damage characteristics of a structural system. The UAV brings mobility to the integrated sensor payload so that the sensors can access remote sites and collect data of structures and field features following an earthquake or a natural disaster. A light detection and ranging (LiDAR) sensor and high resolution camera are installed on a UAV to obtain highly detailed and accurate topological maps of physical surfaces from which structural cracks, surface chances and minor faulting can be detected. One of the challenges in automating the operation of the UAV in challenging operational environments such as GPS-denied environments (e.g., under bridges, indoor environments) is accurate navigation. In a GPS-denied environment, localization relies on on-board sensors, which has limited accuracy when positioning the UAV in a fixed coordinate system. This paper explores solutions to this problem with vision- and marker-based localization methods. The experimental results are based on a LiDAR and camera payload integrated with an octo-rotor UAV platform that has an onboard inertial measurement unit. The experimental results show that three dimensional navigation and geometric mapping in unknown operational environments are both feasible and accurate.