Neighboring Optimal Guidance for Rocket Powered Descent Based on Costate Estimation
Aiming at the requirement of accurate soft landing for rocket powered descent,a neighboring optimal guidance method is proposed based on costate estimation.According to the first order optimal necessary condition,a closed-loop linear feedback guidance method is obtained by solving the optimal guidance problem near the reference trajectory.The reference trajectory is designed using the sequential convex optimization method.Due to the limitation of computational resources and the introduction of trust region constraint,the costate obtained from the reference trajectory under fewer iterations does not satisfy the first order necessary condition,resulting in a decrease in the accuracy of the neighboring optimal guidance method.Therefore,a costate estimation method is proposed.At first,the costate differential equations are discretized by Gaussian pseudo-spectrum method.Then,based on the constant of Hamiltonian function,a new and well-adapted performance index is designed.Finally,a costate estimation algorithm is proposed based on minimum principle.The above method is applied to the simulation example of rocket powed descent,and the Monte Carlo simulation results show that the proposed method has good guidance accuracy and robustness under drag coefficient deviation,thrust acceleration deviation and atmospheric density deviation.
Powered descentNeighboring optimal guidanceReference trajectoryCostate estimationMonte Carlo
NETL
NSTL
万方数据
