基于方向加速法的RLV再入预测制导研究

针对可重复使用运载器(RLV)再入制导问题,为克服算法复杂度高、时效性差、以及环境扰动引起制导精度降低的问题,提出了一种基于修正方向加速法的再入预测制导方法,通过实时预测落点偏差及性能指标来在线搜索最优控制量增量。该方法以无约束极值理论为基础,将多约束的非线性最优化问题转化为无约束最优化问题。采用修正的方向加速法将多维搜索转化为一维搜索;同时进行最佳搜索方向替换,在性能指标收敛最快的方向上直接搜索可行解,降低了算法复杂度并提高了RLV抗扰动能力。仿真结果表明,与单纯形替换法相比,该方法可有效提高制导系统的时效性和精度,具有一定的工程应用价值。

Study on Prediction Guidance Method Based on Powell Algorithm for Reentry Vehicles

A new method based on evolved Powell algorithm was proposed to increase the timeliness and precision of predictive guidance affected by the jamming deriving from environment for reusable launch vehicles,searching for the optimal control value increment through predicting the land point error and performance index.Based on no-constraint extremum theory,this nonlinear optimal problem with multi constraints was simplified into that with no constraint.The search process in multi dimensions was simplified into that in one dimension by the evolved Powell algorithm.Through replacing the direction vector by the optimal one,optimal control value increment was searched in the direction which was the fastest convergence direction of the performance index function.The complexity of algorithm has been decreased and the ability of self-adapt has been increased.Simulation results showed this method has highly increased the timeliness and precision of guidance system when compared with simplex algorithm,great practical value in engineering of this method was also showed.