In this project, we aim to address problems of estimation and control based on partial sensor information, with application to Optical Motion Capture Systems (OMCS) and Unmanned Aerial Vehicles (UAV). OMCS have become increasingly important over the last few years in a variety of applications and, in particular, in the field of aerial robotics. Their outstanding characteristics enabled the implementation of impressive trajectory tracking algorithms. However, their true potential is still overshadowed by some problems, namely, intermittent object reconstruction. In application scenarios where OMCS is not available, such as infrastructure inspection, environmental monitoring, or search and rescue operations, we can consider the dual problem of having sensors installed on-board the vehicles. In this context, similar problems of occlusions, difficult feature matching, allied to higher levels of sensor noise and lower sampling rates require extra care. We propose to explore the key idea of using sensor measurements directly to obtain nonlinear landmark-based or marker-based feedback laws. We also propose to combine hybrid systems theory and
geometric control defined on the natural configuration space of rigid bodies, i.e. the Special Euclidean Group SE(3) for position and attitude and the Special Orthogonal Group SO(3) for attitude only and thereby obtain a rigorous analysis for stability and robustness. We will investigate the possibility of complementing the marker measurements with inertial measurements, something which is overlooked by OMCS systems and can bring significant improvements.

In this project, we aim to address problems of estimation and control based on partial sensor information, with application to Optical Motion Capture Systems (OMCS) and Unmanned Aerial Vehicles (UAV). OMCS have become increasingly important over the last few years in a variety of applications and, in particular, in the field of aerial robotics. Their outstanding characteristics enabled the implementation of impressive trajectory tracking algorithms. However, their true potential is still overshadowed by some problems, namely, intermittent object reconstruction. In application scenarios where OMCS is not available, such as infrastructure inspection, environmental monitoring, or search and rescue operations, we can consider the dual problem of having sensors installed on-board the vehicles. In this context, similar problems of occlusions, difficult feature matching, allied to higher levels of sensor noise and lower sampling rates require extra care. We propose to explore the key idea of using sensor measurements directly to obtain nonlinear landmark-based or marker-based feedback laws. We also propose to combine hybrid systems theory and
geometric control defined on the natural configuration space of rigid bodies, i.e. the Special Euclidean Group SE(3) for position and attitude and the Special Orthogonal Group SO(3) for attitude only and thereby obtain a rigorous analysis for stability and robustness. We will investigate the possibility of complementing the marker measurements with inertial measurements, something which is overlooked by OMCS systems and can bring significant improvements.