Three-dimensional dubins path generation and following for a UAS glider

Abstract

Small unmanned aircraft systems (UAS) typically do not offer the redundancy typical in commercial aviation due to budget and weight limitations. UAS, therefore, have a nontrivial probability of encountering a scenario in-flight that requires an emergency landing. For fixed-wing UAS, one of the most common in-flight emergencies is a total loss of engine or motor thrust, necessitating that the UAS glide to a safe landing site. This paper builds on the extensive previous work in Dubins path planning based on a point-mass model to provide a complete solution to the emergency glider landing problem applicable to a high fidelity fixed-wing aircraft simulator. Path planning and guidance algorithms are proposed to feed a standard autopilot low-level controller. While most previous work focused on maximizing the range or minimizing flight time, this work focus on providing a “safe solution” in the sense that the aircraft is maintained well inside its flight envelope. This provides margin to the controllers to deal with the different sources of environmental and vehicle performance uncertainties. The proposed solution can also be applied to an unmanned fixed-wing platform for which a detailed dynamics model is not available.