Perception and Localization of Unmanned Systems in Complex Environments

Perception and navigation are essential capabilities for intelligent unmanned systems which have a lot of applications in intelligent transportation, surveillance, structure inspection, logistics, and service. GPS has been widely used for positioning and navigation. However, in indoor and many outdoor environments such as urban canon, forest, tunnel, GPS may not be available or unreliable. Hence, there has been a lot of studies on perception and localization based on various sensors such as WiFi, UWB, camera, LiDAR and radar. Each type of sensor has its pros and cons. Different application scenario will require different sensors and there is no one-size-fits-all solution. To achieve accurate and reliable localization in complex and dynamic environments, heterogeneous sensor fusion has been a recent trend and attracted significant interest. In this talk, we shall present some recent development leveraging on heterogenous sensor fusion and deep learning algorithms for perception and localization, and discuss possible future directions of the area. The applications of unmanned systems in structure inspection and logistics will be demonstrated.