Outdoor Localization of an Assistive Autonomous Robot

This thesis regards the design, development and testing of the localization system for Autonomous Mobile Robots in outdoor environments, a project carried out in collaboration with InnoTech Systems L.L.C. and California State University of Los Angeles. The ability to localize itself on the map is at the base of autonomous navigation, an imperative skill for any autonomous vehicle. The proposed solution to the outdoor localization problem is based on the combination of the current state-of-the-art GNSS technology with data from inertial sensors, encoders and stereo camera, where the fusion of multiple information is performed by means of a Kalman filter algorithm with the purpose of reducing the estimation uncertainty. In addition to the system development and validation, this work aims at showing the potential of sensor fusion for localization purposes. After a general introduction on autonomous mobile robots and on the localization problem, this document covers the theory behind the sensors and techniques at the core of the system, to conclude with the project development and the validation results.