Elsevier

Advances in Space Research

Volume 68, Issue 8, 15 October 2021, Pages 3225-3232
Advances in Space Research

Scaling uncertainties on asteroid characteristics to prepare datasets for machine learning

https://doi.org/10.1016/j.asr.2021.06.007Get rights and content

Abstract

Context

Physical and dynamical characterisations of asteroids are used in different fields, such as Solar System formation modelling, Planetary Defence and Resources Prospecting. The vast majority of asteroids are not known in detail - have at best their orbit well defined - and the knowledge on the composition or internal structure is derived by models of reflectivity curves, with limited certainty. Machine learning methods have begun to be used on asteroid datasets, but the major uncertainties about their characteristics are slowing down the applicability.

Aims

This paper reviews some stakes and challenges of asteroid exploration, and why the introduction of common characterisation factors would be beneficial for the asteroid science community, especially with the application of machine learning methodologies. A preliminary scale to quantify the characterisation of asteroids is proposed, and finally discusses its interests and limitations for machine learning applications.

Method

The investigated characteristics of asteroids are: size/shape, orbital dynamics, mass/density, spin, internal structure and composition. This paper reviews the current methods used to determine these parameters, and provides a preliminary scale based on the certainty associated with the different measurements.

Results

Characterisation factors are useful to build datasets that will be used in machine learning algorithms applied to asteroid science. The ratio of currently known asteroids in each defined bin of characterisation factor is estimated. Moreover, a total characterisation factor that yield a preliminary quantification of our knowledge about a specific asteroid (i.e. the sum of all the certainties about its different characteristics) is defined. Finally, characterisation factors for specific applications can be introduced using an adapted weighting system.

Next steps

This preliminary work provides a baseline for scaling uncertainties of asteroids properties. The next step is to create viable datasets using application specific characterisation factors that would allow the use of advanced machine learning algorithms already available.

Introduction

The exploration of asteroids is currently driven by three fields of interest: the knowledge of the Solar System Formation (SSF), Planetary Defence (PD) and Resources Prospecting (RP) to prepare its utilisation. Machine learning (ML) methodologies appear very promising in that context, considering the expected large datasets (soon measured in Exabytes in multiple astronomy experiments). While databases currently present large heterogeneities preventing the full implementation of ML, it is only a matter of time before algorithmic datasets play a dominant role in the area of asteroid sciences

This paper focuses on a preliminary normalisation of the quantification of asteroids characteristics to prepare the datasets for ML. Section 2 presents the motivations for asteroid characterisation for a high number of targets. Section 3 details the currents and foreseen uses of ML in asteroids science. Section 4 presents the state-of-the-art methodologies used to determine different characteristics of asteroids (size/shape, orbital dynamics, mass/density, spin, internal structure and composition) and introduces proposals for characterisation factors. Finally, Section 5 concludes on how the characterisation factors can be used to create datasets for ML, to verify the efficiency of its predictions and states on the value and limitations of the proposed methodology, as well as the next identified steps.

Section snippets

Knowledge of the formation and evolution of the Solar System

There are many unanswered questions about SSF, as identified in (Morbidelli and Raymond, 2016), such as “1) the structure and evolution of proto-planetary disks; 2) the growth of the first planetesimals; 3) orbital migration driven by interactions between proto-planets and gaseous disk; 4) the origin of the Solar System’s orbital architecture; and 5) the relationship between observed super-Earths and our own terrestrial planets”. Small bodies are remnants of the proto-planetary disk that the

ML in asteroid science

ML or Artificial Intelligence (AI) is mainly used for classification, regression, clustering, forecast generation and reconstruction, discovery and insight. (Fluke and Jacobs, 2019) describes the state of the art of ML used with modern astronomical datasets. ML is usually classified as supervised or unsupervised:

  • Supervised ML, relying on a training dataset (typically 70% to 90% of the overall data), a validation dataset and a testing dataset. Supervised ML can be used to perform classification

Asteroids characterisation: What do we know, and how?

There is a gap between the rate of asteroid discoveries (around one hundred each months) and characterisation (Galache et al., 2015), that is done using remote observation campaigns, fly-bys, or through dedicated missions.

Asteroid characterisation is made using photometric, spectroscopic, ephemeris and more rarely, imaging datasets. The following sub sections are a review of the different characteristics of asteroids, the methods that are currently being used to determine each characteristic

Conclusions

Preliminary elements for characterisation scaling factors are defined in this article, in order to define the level of knowledge available about a specific characteristic of an asteroid. Moreover, application specific characterisation factors have been proposed in Section 4.9. While the maximum TCF of 36 is currently not reached for any asteroid, it is reasonable to think that it could be obtained for a limited number of bodies in the coming decades, notably in the frame of RP and mining. A

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors would like to thank B. Segret for his support, R. Tsai for proofreading and the two anonymous reviewers for their constructive and helpful comments.

References (38)

  • J.C. Sercel et al.

    Practical applications of asteroidal ISRU in support of human exploration

  • F. Vilas et al.

    Iron alteration minerals in the visible and near-infrared spectra of low-albedo asteroids

    Icarus

    (1994)
  • B.D. Warner et al.

    The asteroid lightcurve database

    Icarus

    (2009)
  • S. Bandyonadhyay et al.

    Silhouette-Based 3D Shape Reconstruction of a Small Body from a Spacecraft

  • P. Bartczak et al.

    Shaping asteroid models using genetic evolution (SAGE)

    MNRAS

    (2018)
  • W.F. Bottke et al.

    The Yarkovsky and YORP effects: Implications for asteroid dynamics

    Annu. Rev. Earth Planet. Sci.

    (2006)
  • P. Brown et al.

    The flux of small near-Earth objects colliding with the Earth

    Nature

    (2002)
  • V. Carruba et al.

    Machine-learning identification of asteroid groups

    MNRAS

    (2019)
  • V. Carruba et al.

    Machine learning classification of new asteroid families members

    MNRAS

    (2020)
  • Cited by (2)

    View full text