A surrogate-based optimization method is proposed in this paper to realize the optimization of the entire trajectory from launch to reentry. Different from traditional strategy of multiphase trajectory optimization that directly utilizes the original dynamical model, a surrogate model is substituted for the original time-consuming model to improve the iterative efficiency and decrease the computational cost. The procedures of modeling, solving, and optimizing in each substage are addressed, including air-launch ascent trajectory optimization, orbital maneuver planning, deorbit braking, and atmospheric reentry. Instead of optimizing every substage separately and setting interfaces by experience, in this study, substages are patched as an entire system in advance and then the surrogate model is utilized to replace it to perform a systematic optimization. Additionally, the Fourier series is applied to fit the air launch stage during the iteration. Numerical simulations show that the proposed method is acceptable for the pre-design of a full space mission and the results are optimal.

本文提出了一种基于代理的优化方法来实现从发射到再入的整个轨迹的优化。与传统的多阶段轨迹优化策略直接利用原始动力学模型不同，替代模型代替原始耗时模型，以提高迭代效率并降低计算成本。讨论了每个子阶段的建模、求解和优化程序，包括空中发射上升轨迹优化、轨道机动规划、脱轨制动和大气再入。在本研究中，不是单独优化每个子阶段并根据经验设置接口，预先将子阶段作为一个完整的系统进行修补，然后使用代理模型替换它以执行系统优化。此外，傅里叶级数适用于迭代期间的空中发射阶段。数值模拟表明，所提出的方法可以用于全空间任务的预设计，并且结果是最优的。