Desarrollo de un sistema completo de navegación autónoma basado en GNSS y LiDAR para robots terrestres que operan en entornos dinámicos y no estructurados

Résumé

In this thesis, we describe the development of a complete autonomous navigation system for car-like robots operating in dynamic and unstructured environments. These works cover the main areas of ground robotics, localization, detection and tracking of moving objects, terrain analysis, path planning, nonholonomic control, and even the creation of robotic research platforms. We take special care of the system's modularity and scalability while pushing towards the integration of GNSS and LiDAR sensors as two very convenient information sources for robust navigation, taking advantage of the LiDAR precision for local navigation and the bounded errors of GNSS on global positioning. Using this approach, our system is able to navigate autonomously in absence of any prior knowledge (no maps, nor waypoints) in large outdoor environments populated with lots of obstacles. This thesis is a description of the journey that led us from the abstract conception of the system to the actual implementation of a robot able to navigate in a university campus, and the contributions we made along the way.