NAVIGATION AND AUTONOMOUS CONTROL OF MAVS IN GPS-DENIED ENVIRONMENTS

Abstract

This thesis studies issues related to the guidance, navigation and autonomous control of customized micro aerial vehicles in GPS-denied environments. A quadrotor platform with light weight, compact size, sufficient payload, long endurance and high computational power has been developed to as a testbed for our research. Subsequently, a nonlinear model is derived for autonomous control implementation to achieve attitude-loop stabilization and position-loop trajectory tracking. Further, an optical-flow sensor based on visual odometry is developed to provide translational velocities and height measurement. In order to obtain position measurements in GPS-denied environment, a monocular-vision module is investigated and integrated to the airborne system for video streaming. A vision-based simultaneous localization and mapping (SLAM) algorithm is realized on a ground control system (GCS) with robot operating system (ROS) enabled packages based on parallel tracking and mapping (PTAM) structure. Methodologies have been verified with real flight tests in indoor scenario and part of the philosophies has been further extended to nano aerial vehicles (NAV).