A Multi-functional Robot for Civil Infrastructure Inspection

Abstract

This thesis focuses on designing and building a robotic system for civil infrastructure inspections. The robot is equipped with two non-destructive-evaluation (NDE) types of sensors: a ground penetrating (GPR) sensor and two electrical resistivity (ER) sensors. For localization, the robot employs an extended Kalman filter (EKF)-based approach using global positioning system (GPS) signal. Differ from its predecessor, the currently built robot is capable of localizing and navigating in GPS-denied environments by using a visual-inertial odometry method with a stereo camera. In addition, the robot can build an elevation map of its surrounding areas using a PrimeSence camera for obstacle avoidance.