Evaluation and Comparison of the Positional Accuracy in Various Localization Systems for Mobile Robots in GPS-Denied Environments
Navigation has guided the development of humans throughout history. Nowadays, navigation is a vital part of autonomous robotics. A major component to robotic navigation is the robot's ability to localize, or keep track of where it is within its environment. As robots become a more ubiquitous part of our world, it will be of vital importance for robots to localize in GPS-denied environments, such as heavily wooded areas, cities with many tall buildings, or even other planetary bodies. This will ensure they properly operate and assist in human development. This thesis explores the current state of the art in LiDAR-based localization systems in GPS-denied environments. A total of nine localization systems are tested in three environments. These systems were tested in urban environments to establish a baseline before testing in an off-road environment to determine how they might extrapolate to harsher environments. Each system is tested with both high and low end LiDARs to help determine the most cost effective method of localization. It was found that LiDAR-based inertial localization solutions produce the most effective localization estimate. Furthermore, fusion of these systems with wheel encoders and inertial measurement units can improve their localization performance.