Core Journal of Peking University

Excellent Sci-Tech Journal of Chinese Universities

Journal of Committee of Deep Space Exploration Technology, Chinese Society of Astronautics(CDSET-CSA)

Advanced Search
Turn off MathJax
Article Contents

NI Yang, PAN Binfeng. Deep Neural Network Approximation of the Asteroid Polyhedron Model[J]. Journal of Deep Space Exploration. doi: 10.15982/j.issn.2096-9287.2022.20200074
Citation: NI Yang, PAN Binfeng. Deep Neural Network Approximation of the Asteroid Polyhedron Model[J]. Journal of Deep Space Exploration. doi: 10.15982/j.issn.2096-9287.2022. 20200074

Deep Neural Network Approximation of the Asteroid Polyhedron Model

doi: 10.15982/j.issn.2096-9287.2022. 20200074
  • Received Date: 2020-11-19
  • Rev Recd Date: 2022-07-19
  • Available Online: 2022-08-31
  • In this paper, based on the polyhedral model of irregular asteroids, the gravity calculation and position judgment criteria of irregular asteroids are studied and approximated by using Deep neural network(DNN), and a fast calculation method for the unpowered descent mission of asteroids is proposed. The polyhedron method is used to generate training data, and two DNN models are trained to calculate the gravitational field around the asteroid and judge whether it reaches the asteroid boundary. This method can not only ensure the accuracy of gravity calculation near asteroids, but also save calculation time. Taking Eros 433 asteroid as an example, the trajectory and impact point position of the unpowered descent process are simulated. The simulation results show that the calculation accuracy of the DNN model can meet the mission requirements, the deviation of the impact point position is within a reasonable range, and the calculation efficiency is high, which can be used in large-scale simulation experiments.
  • ● Based on the data generated by the polyhedron method,Deep Neural Network(DNN) is used to learn and approximate the gravitational acceleration and boundary determination criteria near the asteroid. ● The influence of different neural network hyperparameters and training datasets for training DNN are analyzed in detail. ● 3. A fast calculation method the unpowered descent mission of asteroids based on DNN is proposed and applied to the 433 Eros asteroid.

  • [1]
    李俊峰, 宝音贺西, 蒋方华. 深空探测动力学与控制[M]. 北京: 清华大学出版社, 2014.

    LI J F, BAOYIN H X, JIANG F H. Dynamics and control of interplanetary flight[M]. Beijing: Tsinghua University Press, 2014.
    [2]
    WERNER R A,SCHEERES D J. Exterior gravitation of a polyhedron derived and compared with harmonic and mascon gravitation representations of asteroid 4769 Castalia[J]. Celestial Mechanics and Dynamical Astronomy,1996,65(3):313-344.
    [3]
    GARMIER R,BARRIOT J P,KONOPLIV A S,et al. Modeling of the Eros gravity field as an ellipsoidal harmonic expansion from the NEAR Doppler tracking data[J]. Geophysical research letters,2002,29(8):1-3.
    [4]
    崔祜涛,张振江,余萌. 多面体模型的Eros433引力场计算与分析[J]. 哈尔滨工业大学学报,2012,44(3):17-22. doi:  10.11918/j.issn.0367-6234.2012.03.004

    CUI H T,ZHANG Z J,YU M. Computing and analysis of gravity field of Eros433 using polyhedron model[J]. Journal of Harbin Institute of Technology,2012,44(3):17-22. doi:  10.11918/j.issn.0367-6234.2012.03.004
    [5]
    束磊,刘睿,张迎春. 不规则形状小行星引力场组合建模方案[J]. 飞行力学,2012,30(6):560-564. doi:  10.13645/j.cnki.f.d.2012.06.019

    SHU L,LIU R,ZHANG Y C. Combined modeling for gravitation field of asteroid with irregular shape[J]. Flight Dynamics,2012,30(6):560-564. doi:  10.13645/j.cnki.f.d.2012.06.019
    [6]
    肖尧,阮晓钢,魏若岩. 基于433 Eros的多面体引力模型精度与运行时间研究[J]. 深空探测学报(中英文),2016(1):41-46.

    XIAO Y,RUAN X G,WEI R Y. Research on the accuracy and operation time of polyhedron gravity model base on 433 Eros[J]. Journal of Deep Space Exploration,2016(1):41-46.
    [7]
    GAO A,LIAO W. Efficient gravity field modeling method for small bodies based on Gaussian process regression[J]. Acta Astronautica,2019,157:73-91. doi:  10.1016/j.actaastro.2018.12.020
    [8]
    YANG H,LI S,SUN J. A fast Chebyshev polynomial method for calculating asteroid gravitational fields using space partitioning and cosine sampling[J]. Advances in Space Research,2020,65(4):1105-1124. doi:  10.1016/j.asr.2019.11.001
    [9]
    FURFARO R,BAROCCO R,LINARES R,et al. Modeling irregular small bodies gravity field via extreme learning machines and Bayesian optimization[J]. Advances in Space Research,2021,67(1):617-638. doi:  10.1016/j.asr.2020.06.021
    [10]
    CHENG L,WANG Z,SONG Y,et al. Real-time optimal control for irregular asteroid landings using deep neural networks[J]. Acta Astronautica,2020,170:66-79. doi:  10.1016/j.actaastro.2019.11.039
    [11]
    KINGMA D P, BA J. Adam: A method for stochastic optimization[J]. arXiv preprint arXiv: 1412.6980, 2014.
    [12]
    李晓宇. 小行星软着陆动力下降段制导与控制方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2015

    LI X Y. Guidance and control law for asteroid soft landing in power decent phase[D]. Harbin: Harbin Institute of Technology, 2015.
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Article Metrics

Article views(20) PDF downloads(0) Cited by()

Proportional views
Related

Deep Neural Network Approximation of the Asteroid Polyhedron Model

doi: 10.15982/j.issn.2096-9287.2022. 20200074

Abstract: In this paper, based on the polyhedral model of irregular asteroids, the gravity calculation and position judgment criteria of irregular asteroids are studied and approximated by using Deep neural network(DNN), and a fast calculation method for the unpowered descent mission of asteroids is proposed. The polyhedron method is used to generate training data, and two DNN models are trained to calculate the gravitational field around the asteroid and judge whether it reaches the asteroid boundary. This method can not only ensure the accuracy of gravity calculation near asteroids, but also save calculation time. Taking Eros 433 asteroid as an example, the trajectory and impact point position of the unpowered descent process are simulated. The simulation results show that the calculation accuracy of the DNN model can meet the mission requirements, the deviation of the impact point position is within a reasonable range, and the calculation efficiency is high, which can be used in large-scale simulation experiments.

NI Yang, PAN Binfeng. Deep Neural Network Approximation of the Asteroid Polyhedron Model[J]. Journal of Deep Space Exploration. doi: 10.15982/j.issn.2096-9287.2022.20200074
Citation: NI Yang, PAN Binfeng. Deep Neural Network Approximation of the Asteroid Polyhedron Model[J]. Journal of Deep Space Exploration. doi: 10.15982/j.issn.2096-9287.2022. 20200074
WeChat followshare

Top

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return