Intelligent vehicles operate in open and uncontrolled environments. They may assist the driver or have control of some driving tasks.
For autonomous navigation, localization is still an open problem. Global Navigation Satellite Systems (GNSS) provide global localization and need to be assisted by exteroceptive and proprioceptive sensors and map information to increase the performance particularly in terms of accuracy and integrity. I study methods and algorithms able to merge these sources of information and able to provide reliable confidence indicators.
Perception systems that use on-board exteroceptive sensors (e.g. cameras or lidars) and information shared by cooperative agents need also deep improvements to enable drivable space characterization and scene understanding. I consider frameworks and methodologies able to handle uncertainties and partial knowledge in highly dynamic open conditions.
I am also interested in the definition and elaboration of enhanced digital maps able to support these different processes, being convinced that they will play a key role in this field.