ECEF frame is commonly used as a reference frame and resolving axes for application such as satellite navigation solutions or airbone photogrammetry.
ECEF frame navigation equations block diagram is shown in Figure 1. The suffixes (-) and (+) are used to denote a priori (at time \(t\)) and a posteriori values (at time \(t + \tau_i\)).
Attitude Update
To track the attitude change, we must track the rotation matrix through time. Consider the attitude change over time interval \(t\) to \(t + \tau_i\). The time derivative of the trasnformation matrix is:
where \(\boldsymbol{\Omega}^b_{ib}\) is the skew-symmetric matrix of the IMU's angular-rate measurement and \(\boldsymbol{\Omega}^e_{ie}\) is the
skew-symmetric matrix of the Earth-rotation vector defined in section >Rotation of Earth. Hence, the rotation of
the Earth must be accounded for in updating the attitude.
Since this is a first-order matrix ordinary differential equation, assuming \(\boldsymbol{\Omega}^e_{ie}\) is constant:
Since the rotation of the Earth is much slower compared to the angular rate measurements from the IMU, this small angle approximation is always valid for the
Earth rate term of the attitude update equation.
Precision Attitude Update
We can replace the first order approximation in eq (\(\ref{4.3.3}\)) with the attitude update matrix:
Since the specific-force measurement is an average over time \(t\) to \(t + \tau_i\), the transformation matrix should be similarly averaged. Assuming a constant angular rate:
However, the mean of two transformation matrices does not precisely produce the mean of the two attitudes. The less the attitude varies over the time interval, the smaller the errors introduced by this approximation.
Precision Specific-Force Frame Transformation
Retaining the first-order approximation for the Earth-rate term, the prcecise transformation of the specific force to ECEF-frame axes is:
The equation shows that the rate of change of velocity resolved about the ECEF axes incorporates a centrifugal and a Coriolis term due
to the rotation of the resolving axes. From the section on gravity models, we know that the applied acceleration \(\mathbf{a}^e_{ib}\) is the sum of the measured specific force, \(\mathbf{f}^e_{ib}\), and the acceleration due to the gravitation force, \(\boldsymbol{\gamma}^e_{ib}\).
Furthermore, the acceleration due to gravity, \(\mathbf{g}^e_b\), is the sum of the gravitational and centrifugal accelerations. Hence:
Since the Coriolis term is much smaller than the specific-force and gravity terms (except for space applications), it is reasonable approximation
to neglegt the variation of the Coriolis term over the integration interval. Hence:
In ECEF frame implementation of the navigation equations, the time derivative of the Cartesian position is simply velocity as the reference frame and resolving
axes are the same:
Exact evaluation of the Coriolis and transport-rate terms require knowledge of the velocity at the end of the update interval, requiring recursive solution. A good but processor-intensive solution is
a two-step recursive method.