UAV Cargo Operations: Path Planning and Decision Making for Emergency Landing

Abstract

[EN] This thesis develops a Path Planning and Decision Making algorithm for Emergency Landing after engine failure under Wind Effects. The main contri bution of this work is the design of a novel wind-aware Dubins and trochoidals based Rapidly-exploring Random Tree (RRT) path planning algorithm, along with a high-level reasoning decision making logic which can accommodate un predicted changes in the scenario while en descent. The path planning algorithm has been optimised to consider wind effects, together with time and UAV manoeuvrability constraints in an emergency land ing. The RRT algorithm has been formulated to produce directly flyable 3D paths which minimise cross-wind length while avoiding airspace restricted ar eas and reaching one of the designated landing sites. A diversion route search algorithm has also been developed. The decision making algorithm provides the system with high-level reason ing during forced landing. The developed logic can accommodate common dy namic scenario changes during descent: landing site denial, airspace restrictions reassessment and cross-track error rise due to wind uncertainty in path plan ning. Redundancy in path scheduling is achieved by combination of main and diversion landing routes. Extensive simulation of the system is provided in a realistic environment, testing performance for all considered scenario alterations during descent. Re sults are discussed, showing performance and reliability of the system