Coordinate System Transformations (GPS)

Any Cartesian coordinate system can be transformed to another Cartesian coordinate system through three succeeded rotations if their origins are the same and if they are both right-handed or left-handed coordinate systems. These three rotational matrices are:

tmp2A88_thumb

where a is the rotating angle, which has a positive sign for a counter-clockwise rotation as viewed from the positive axis to the origin.tmp2A89_thumbare called the rotating matrix around the x, y, and z-axis, respectively. For any rotational matrix R, there aretmp2A90_thumbthat is, the rotational matrix is an orthogonal one, where R-1 and RT are the inverse and transpose of the matrix R.


For two Cartesian coordinate systems with different origins and different length units, the general transformation can be given in vector (matrix) form as

tmp2A93_thumb

wheretmp2A94_thumbis the scale factor (or the ratio of the two length units), and R is a transformation matrix that can be formed by three suitably succeeded rotations.tmp2A95_thumbdenote the new and old coordinates, respectively;tmp2A96_thumbdenotes the translation vector and is the coordinate vector of the origin of the old coordinate system in the new one.

If rotational angle a is very small, then one has sin a~ a and cos a~ 0. In such a case, the rotational matrix can be simplified.

tmp2A100_thumb

where a1, a2, a3 are small rotating angles around the x, y and z-axis, respectively. Using the simplified R, the transformation 2.8 is called the Helmert transformation. As an example, the transformation from WGS-84 to ITRF-90 is given by (McCarthy 1996):

tmp2A101_thumb

wheretmp2A102_thumbthe translation vector has the unit of meter.

The transformations between the coordinate systems of GPS, GLONASS and Galileo can be generally represented by Eq. 2.8 with the scale factor ,tmp2A103_thumb(i.e., the length units used in the three systems are the same). A formula of velocity transformations between different coordinate systems can be obtained by differentiating the Eq. 2.8 with respect to the time.

Next post:

Previous post: