Orbit-centered atmospheric density prediction using artificial neural networks

Pérez, David; Wohlberg, Brendt; Lovell, T. Alan; Shoemaker, Michael A.; Bevilacqua, Riccardo

doi:10.1016/j.actaastro.2014.01.007

Cited by 36 publications

(17 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The estimation of maps of total electron content (TEC) has been tackled in Acharya et al (), Habarulema et al (), Habarulema et al (), Hernandez‐Pajares et al (), Leandro and Santos (), Watthanasangmechai et al (), Wintoft and Cander (), and Tulunay et al (). The foF2 parameter (which is the highest frequency that reflects from the ionospheric F 2‐layer) has been studied in Oyeyemi et al (), Poole and McKinnell (), and Wang et al (), and thermosphere density in Choury et al () and Pérez et al ().…”

Section: Review Of Machine Learning In Space Weathermentioning

confidence: 99%

The Challenge of Machine Learning in Space Weather: Nowcasting and Forecasting

2019

View full text Add to dashboard Cite

The numerous recent breakthroughs in machine learning make imperative to carefully ponder how the scientific community can benefit from a technology that, although not necessarily new, is today living its golden age. This Grand Challenge review paper is focused on the present and future role of machine learning in Space Weather. The purpose is twofold. On one hand, we will discuss previous works that use machine learning for Space Weather forecasting, focusing in particular on the few areas that have seen most activity: the forecasting of geomagnetic indices, of relativistic electrons at geosynchronous orbits, of solar flares occurrence, of coronal mass ejection propagation time, and of solar wind speed. On the other hand, this paper serves as a gentle introduction to the field of machine learning tailored to the Space Weather community and as a pointer to a number of open challenges that we believe the community should undertake in the next decade. The recurring themes throughout the review are the need to shift our forecasting paradigm to a probabilistic approach focused on the reliable assessment of uncertainties, and the combination of physics-based and machine learning approaches, known as gray box. Plain Language SummaryIn the last decade, machine learning has achieved unforeseen results in industrial applications. In particular, the combination of massive data sets and computing with specialized processors (graphics processing units, or GPUs) can perform as well or better than humans in tasks like image classification and game playing. Space weather is a discipline that lives between academia and industry, given the relevant physical effects on satellites and power grids in a variety of applications, and the field therefore stands to benefit from the advances made in industrial applications. Today, machine learning poses both a challenge and an opportunity for the space weather community. The challenge is that the current data science revolution has not been fully embraced, possibly because space physicists remain skeptical of the gains achievable with machine learning. If the community can master the relevant technical skills, they should be able to appreciate what is possible within a few years time and what is possible within a decade. The clearest opportunity lies in creating space weather forecasting models that can respond in real time and that are built on both physics predictions and on observed data.

show abstract

Section: Review Of Machine Learning In Space Weathermentioning

confidence: 99%

The Challenge of Machine Learning in Space Weather: Nowcasting and Forecasting

2019

View full text Add to dashboard Cite

show abstract

“…These prediction techniques are formulated based on non-linear Artificial Neural Network (see other applications in, e.g. Choury et al 2013;Pérez et al 2014), which include Non-linear AutoRegressive (NAR), Non-linear Autoregressive with Exogenous (External) Input (NARX), and Non-Linear Neural Network Regression (NLNNR). All these techniques are discussed in Section 2.3.…”

Section: Introductionmentioning

confidence: 99%

Estimating and predicting corrections for empirical thermospheric models

Forootan

Farzaneh

Lück

et al. 2019

Geophysical Journal International

View full text Add to dashboard Cite

Quantifying spatial and temporal changes in thermospheric neutral density is important for various applications such as precise orbit determination, estimating mission lifetime and reentry prediction of Earth orbiting objects. It is also crucial for analysis of possible collisions between active satellite missions and space debris. Empirical models are frequently applied to estimate neutral densities at the position of satellites. But their accuracy is severely constrained by model simplifications and the sampling limitation of solar and geomagnetic indices used as inputs. In this study, we first estimate thermospheric neutral density by processing the high-accuracy accelerometer measurements on-board of the twin-satellite mission Gravity Recovery And Climate Experiment (GRACE). Daily density corrections (in terms of scales) are then computed for the commonly used NRLMSISE-00 empirical model. The importance of these daily scales is examined within an orbit determination practice. Finally, three data-driven prediction techniques based on Artificial Neural Network (ANN) are applied to forecast the daily density corrections for few days to months. Our numerical results indicate that GRACE derived scales are correlated with solar and geomagnetic indices and can improve the timing (from few hours to days) and magnitude of model simulations (up to 10-100 times) during high solar or geomagnetic activity when they usually perform poorly. We found that the Non-linear Autoregressive with Exogenous (External) Input (NARX) ANN technique performs well in predicting the corrections with an average fit of 0.8 or more in terms of squared correlation coefficients for timescales of 7-90 days.

show abstract

“…As was shown in [18], out of all the elements in Eq. 3, the density is the one that varies the most.…”

Section: A Nonlinear Relative Dynamicsmentioning

confidence: 81%

“…To create the reference trajectory using predicted density, the methodology presented in [18] is exploited. In particular, an STK scenario, in which NRLMSISE-00 is used for modeling the density, was propagated using STK's high-precision orbit propagator (HPOP) for a period of four days before and two days after the maneuver's initial time.…”

Section: Reference Trajectory For a Rendezvous Maneuvermentioning

confidence: 99%

“…This is not desirable from the point of view of creating reference trajectories, as this assumption does not account for a realistic behavior of the control force. In [18], Artificial Neural Network (ANN) predictors capable of forecasting the density along the future orbit of a spacecraft were introduced. In the present work, these predictors are used for designing rendezvous reference trajectories created specifically for differential drag-based maneuvering, provided that density estimates or measurements of the present value for the density are available.…”

mentioning

confidence: 99%

See 1 more Smart Citation