This paper aims to combine high-fidelity shooting method with Shallow Neural Network (SNN) to generate optimal flight trajectories in real time. The flight trajectory optimization is formulated as a nonlinear optimal control problem. Pontryagin’s maximum principle is used to devise necessary conditions for optimality, which not only gives rise to the structure of optimal control, but also allows to formulate a shooting function in terms of initial state, initial costate, and switching times. A typical homotopy method is employed to generate initial costate and switching times for each initial state. All the generated data are finally used to train the SNN. As a result, given any initial state, SNN can immediately produce the initial costate and switching times for the optimal flight trajectory. The developments are finally verified through numerical simulations.
Real-Time Generation of Optimal Flight Trajectories via Shallow Neural Network
Lect. Notes Electrical Eng.
International Conference on Autonomous Unmanned Systems ; 2021 ; Changsha, China September 24, 2021 - September 26, 2021
Proceedings of 2021 International Conference on Autonomous Unmanned Systems (ICAUS 2021) ; Kapitel : 164 ; 1668-1676
2022-03-18
9 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Long Flight-Time Range-Optimal Atmospheric Flight Vehicle Trajectories
British Library Conference Proceedings | 1994
|Long flight-time range-optimal aircraft trajectories
AIAA | 1996
|