高精度火星大气制动轨迹智能高效优化方法

Fast optimization method for Mars high-fidelity aerobraking trajectory using a neural network

大气制动技术是减少深空探测任务燃耗的重要手段.本文面向火星大气制动轨迹优化问题,提出了一种高精度火星大气制动轨迹快速优化方法.首先,构建了全维度非球形引力摄动动力学方程,并基于此完成了火星大气制动轨迹参数的敏感性分析,结果表明近火点高度在高精度模型下存在波动现象.然后,应用深度神经网络发展了大气制动轨迹参数快速估算方法,提出了受控火星大气制动轨迹优化模型,并采用遗传算法进行优化.数值仿真结果表明,本文所提方法较传统简化动力学方法在计算效率和精度方面优势明显.

Aerobraking technology is a significant way to save fuel for interplanetary missions,and has been employed several times for Mars missions.A novel intelligent optimization method is proposed in this study to achieve high-fidelity aerobraking trajectory for Mars missions.First,a full-dimensional,nonspherical gravitational perturbation dynamic model is derived;this model is employed to analyze the sensitivity of the Mars aerobraking trajectory.The results indicate that the periapsis altitude of the aerobraking trajectory is volatile rather than being a constant value under the high-fidelity dynamic model.Then,two deep neural networks are applied to replace the complex dynamic equations for efficiently approximating the parameters of the aerobraking trajectory.Finally,a controlled aerobraking trajectory optimization approach is introduced for Mars missions,and the genetic algorithm is used to optimize the control variables.The numerical simulation results verify the effectiveness of the proposed method and show its outstanding advantages with respect to efficiency and accuracy.