Method and apparatus for orbit determination

• US 6,708,116 B2
• Filed: 11/12/2002
• Issued: 03/16/2004
• Est. Priority Date: 11/13/2001
• Status: Expired due to Term

• Current Assignee: Ansys Government Initiatives Incorporated (Ansys Incorporated)
• Original Assignee: Analytical Graphics
• Inventors: Wright, James
• Primary Examiner(s): NGUYEN, TAN QUANG
• Application Number:
  US10/292,726
• Publication Number:
  US 20030096609A1
• Time in Patent Office: 490 Days
• Field of Search: 701/226, 701/13, 701/213, 342/357.02, 342/357.01, 342/357.12, 342/357.03, 455/12.1, 455/13.2, 455/427
• US Class Current: 701/226
• CPC Class Codes: B64G 1/242 Orbits and trajectories

Abstract

A method and apparatus for estimating the orbits of spacecraft or other objects relative to primary celestial bodies, given applicable measurements. The present invention uses the variation of parameters in universal variables for spacecraft trajectory propagation during operation of its optimal sequential orbit estimation. This method provides a universal capability for all three conics: elliptic, parabolic, and hyperbolic, forward-time and backward-time numerical integration for filtering and smoothing, respectively. A method for determining local and global estimates of atmospheric density is also provided. A method for realistic orbit simulation provides a means for validating an orbit estimate.

Claims

1. A method of estimating in real-time an orbit of a satellite using a recursive filter and wherein the recursive filter first receives an a priori state estimate and an a priori state error covariance matrix for the satellite, and wherein the method comprises:

inputting tracking measurements for the satellite received in real-time from tracking stations to the recursive filter at filter time update “

T”

;

inputting filter tracking station platform data to the recursive at filter time update “

T”

;

computing recursively the current state estimate, the current error covariance and the current filter measurement update of the satellite at filter time update “

T”

;

moving the state estimate and error covariance from filter time update “

T”

to filter time update “

T+1”

;

incorporating into the filter measurement update the information provided by the tracking data measurement at filter time update T; and

determining a current estimate of the orbit of the satellite.

2. The method of claim 1 further comprising determining an estimate of density {circumflex over (ρ

)}(H) in real-time for use in the computation of the acceleration of the satellite due to atmospheric drag during satellite trajectory propagation, wherein the satellite has a current height “

H”

, and wherein the determination of the estimate of density {circumflex over (ρ

)}(H) comprises;

computing from an atmospheric density model (ADM) atmospheric density {overscore (ρ

)} at height H ({overscore (ρ

)}(H)), wherein the ADM derives {overscore (ρ

)}(H) in part from current observatory measured solar and geomagnetic indices;

propagating through time an estimate of the relative density correction at the perigee height {circumflex over (D)}(H_P) of the satellite, wherein H_P is the satellite height at perigee and $\hat{D}<br><br>\left(H_P\right)=\frac{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\widehat{\mathrm{\rho <br><br>}}<br><br>\left(H_P\right)}{\stackrel{\_}{\mathrm{\rho <br><br>}}<br><br>\left(H_P\right)};<br><br>$

deweighting the variance estimate of {circumflex over (D)}(H_P);

computing an estimate of the correction to atmospheric density at a perigee height, Δ

{circumflex over (ρ

)}(H_P);

computing from Δ

{circumflex over (ρ

)}(H_P) and the current height of the satellite “

H”

the value of a mapping function, S<sub>Δ

ρ</sub>

;

computing an estimate of the density correction at the current height, Δ

{circumflex over (ρ

)}(H), by multiplying S<sub>Δ

ρ</sub>

by {circumflex over (D)}(H_P); and

adding the density correction estimate Δ

{circumflex over (ρ

)}(H) to {overscore (ρ

)}(H) to determine {circumflex over (ρ

)}(H).

3. The method of claim 2, wherein propagating through time an estimate of the relative density correction at the perigee height {circumflex over (D)}(H_P) of the satellite, wherein H_P is the satellite height at perigee and $\hat{D}$

(HP)=Δ



 



ρ

^

(HP)ρ

_

(HP)

comprises applying a Gauss-Markov transition function Φ

(t_k+1, t_k) to propagate {circumflex over (D)}(H_P) from time t_k to time t_k+1 where;

4. The method of claim 2, wherein deweighting the variance estimate of {circumflex over (D)}(H_P) comprises applying a Gauss-Markov deweighting function for variance on the filter time update where:

{overscore (σ

)}_k+1²({circumflex over (D)}_k(H_P))=└

1−

Φ

²(t_k+1, t_k)┘

{overscore (σ

)}<sub>Δ

ρ</sub>

²(H_P)

where {overscore (σ

)}<sub>Δ

ρ</sub>

²(H_P) is a positive time constant.

5. The method of claim 2, wherein the current observatory measured solar and geomagnetic indices comprises the ten centimeter solar flux measurement F₁₀, its mean across several months {overscore (F)}₁₀, and a geomagnetic measurement.

6. The method of claim 5 wherein geomagnetic measurement is selected from the group consisting of the three-hourly geomagnetic measurement K_P and a_P.

7. The method of claim 6 wherein the recursive filter gain is determined, in part, by the value of K_P.

8. The method of claim 6 wherein the recursive filter gain is determined, in part, by the value of a_P.

9. The method of claim 1 further comprising validating an orbit determination using simulated data wherein the simulated data is generated by;

obtaining a priori state of the satellite;

computing an initial simulated orbit estimate and a simulated truth orbit;

determining the initial orbit error by computing the difference between simulated truth orbit and the simulated orbit estimate; and

adjusting the initial orbit error using error estimates reflected in the state error covariance matrix.

10. The method of claim 9 wherein the error estimates reflected in the state error covariance matrix are selected from the group consisting of serially correlated transponder biases for two-way measurements, serially correlated measurement biases for two-way measurements, serially correlated clock biases for one-way measurements;

and serially correlated force model error biases for all orbits.

11. The method of claim 1 wherein computing recursively at filter time update “

T”

the current state estimate, the current error covariance and the current filter measurement update of the satellite, moving the state estimate and error covariance from filter time update “

T”

to filter time update “

T+1”

, incorporating into the filter measurement update the information provided by the tracking data measurement at filter time update T, and arriving at a current estimate of the orbit of the satellite are performed in an object-oriented computer program.