|
SUPPORT SITE |
|
|
|
SUPPORT SITE |
|
Path: SpacecraftEstimation/OrbitEstimation
% Continuous discrete iterated Kalman Filter.
The state and covariance matrices are numerically integrated.
The model is of the form:
dx/dt = f(x,t,u)
y = h(x)*x + y0
F(x,t,u) = df/dx
fName is a RHS function for the state and covariance of the form
xDot = fName( x, t, d.fData, q )
hName is an observation partials function of the form
[z, dhdx] = hName( x, t, d.hData )
See OrbitKalmanFilter for a demo. There are m measurements with a total
of j elements (the measurements do not need to be the same type or size).
--------------------------------------------------------------------------
Form:
d = CDKF45( d )
--------------------------------------------------------------------------
------
Inputs
------
d (:) Data structure
.time Current Julian date (days)
.dT Integration time step (sec)
.secFromStart Seconds from start (sec)
.nIterations Number of iterations
.propagator Choice of propagator, RK4 or RK45
.hLast Last step for RK45
.tol Tol for RK45
.x (n,1) State
.p (n,n) Covariance
.k (n,j) Gain matrix (calculated)
.q (n,n) Plant noise matrix
.r {m} Measurement noise matrix
.meas Measurements
.z(j,1) New measurement(s) in column vector
.fName fRHS function name
.fData fRHS data
.dFName Derivative function
.nStates Number of elements in the state vector
... (user defined)
.hName h function name
.hData(m) h data for measurements
... (user defined)
.xMeas (n,1) State from measurements (calculated)
.pMeas (n,n) Covariance from measurements (calculated)
-------
Outputs
-------
d (:) Data structure
--------------------------------------------------------------------------
References: Gelb, A. Ed., Applied Optimal Estimation, MIT Press. p.188.
Table 6.1-1. Also, pp. 190-191.
--------------------------------------------------------------------------
Math: Integration/RK4 Math: Integration/RK45 SpacecraftEstimation: OrbitEstimation/ReshapeCovariance
Back to the SpacecraftEstimation Module page