Robust Neighboring Optimal Guidance for Endoatmospheric Powered Descent under Uncertain Wind Fields

This paper considers the problem of guiding a reusable rocket in the endoatmospheric powered descent phase. As well known, the powered descent guidance problem for reusable rocket is quite challenging due to the dispersions caused by the unknown time-varying wind fields. To this end, this paper proposes a robustified optimal guidance framework that estimates the unknown wind field, and compensates it with a neighboring optimal guidance law. This paper firstly develops a neighboring optimal guidance law by calculating the sensitivity of the optimal trajectory to the wind field vector, that assumes the wind field is constant. Then, a wind-field-observer-based neighboring optimal guidance law is developed based on it. A robustified policy is designed for the wind field observation error to improve the robustness of the guidance algorithm using a disturbance observer based compensator. Consequently, the proposed endoatmospheric powered descent guidance could achieve both optimality and robustness. Simulation results are provided to show the guidance performance under different wind fields.