摘要
针对复杂动力学环境下GEO空间碎片清除任务的高精度交会问题,提出了基于深度学习技术的GEO卫星轨迹制导算法。首先,以经典预测校正技术为基础,建立通用的GEO卫星转移轨迹制导算法架构,现有制导技术以及人工智能技术都适用于该架构。然后,基于深度学习技术对GEO卫星动力学模型参数进行拟合,利用一个深度神经网络同时输出高阶非球形引力加速度以及高精度日月星历,从而提高GEO卫星动力学模型参数的预测效率。最后,利用微分修正算法进行高精度轨迹修正,在此基础上提出结合深度神经网络以及高精度动力学模型的混合修正策略以同时保证转移轨迹修正精度和制导效率。仿真表明所提方法与传统基于高精度动力学模型的制导技术相比可有效提高转移轨迹制导效率。
A deep-learning-based geosynchronous satellite transfer trajectory guidance method is proposed to improve the rendezvous precision for space debris removal in complex dynamic environment.Firstly,based on the classical predictor-corrector guidance algorithm,a general framework for GEO satellite transfer trajectory guidance algorithm is established.This guidance framework is appropriate for existing guidance methods and intelligent methods.Then,with a deep neural network,the GEO satellite dynamic model parameters are fitted.The high-order non-spherical gravitational acceleration and luni-solar ephemerides are predicted to improve the calculation efficiency.Finally,a differential correction algorithm is employed for high-precision trajectory correction.To ensure the accuracy and efficiency of trajectory guidance,a hybrid correction strategy combining the deep neural network and high-precision dynamic model is developed.Simulation results show that the proposed method can effectively improve the guidance efficiency compared to traditional methods.
作者
黄旭辰
黄旭星
杨彬
李爽
HUANG Xuchen;HUANG Xuxing;YANG Bin;LI Shuang(College of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2023年第5期719-730,共12页
Journal of Astronautics
基金
国家自然科学基金(11672126)
空间智能控制国防重点实验室开放基金(2021-JCJQ-LB-010-04)。
关键词
空间碎片清除
高精度轨迹修正
预测校正制导
深度神经网络
微分修正算法
Space debris removal
High precision trajectory correction
Predictor corrector guidance
Deep neural network
Differential correction algorithm
