A cloud-based framework for risk-aware intelligent navigation in urban environments

Abstract

Remotely Piloted Aircraft Systems (RPAS) are being widely studied and developed due to their mission flexibility, reconfigurable architecture and low cost. In this paper, a novel Cloud-based framework for intelligent navigation of RPAS in urban environments is proposed, toward achieving fully-autonomous missions without compromising safety. The advantages of the proposed approach reside in the flexibility in designing and implementing complex systems and algorithms with partial independence of the specific flying and control hardware. Thanks to the real-time use of Cloud-based algorithms, advanced control ability and risk-aware navigation and planning can be implemented, without increasing the flying pay-load. The proposed framework is structured in stacked logical layers, distributed between the Cloud and the RPAS, implementing the tasks of autonomous flying, processing information and decision-making. The architecture comprises five layers, including map and risk-aware path planning generation, and control layers. Important novelty elements are: (i) the definition of a dynamical risk-map, and (ii) an On-board Control System able to perform emergency maneuvers, if communication with the Cloud is poor or missing.