Basic program operation
track has three main components in its analysis of kinematic data:
Rinex file reading and bias parameter assignment;
Resolution of the biases to the correct integer values;
Computation of the kinematic trajectory using the resolved values of the biases.
The most time consuming step is (2) and the time taken for this part depends on the quality of the input data and the separations of the moving and fixed stations. Some of the user inputs in the command file will also effect the speed of this task.
In step (1), all the data to be processed in read into memory and biases are assigned to all data. A new bias is introduced whenever:
A satellite rises above the minimum elevation angle assigned by the user (default is 15 degrees);
There is a gap in the data of more than max_gap (default is 1), i.e. any missing data causes a new bias to be introduced;
There is a jump in the computed ionospheric phase of more than
max_ion_jmp. Default is 1.0 cycles. (Due to the ratio of the L2 to L1 frequencies, a 1-1 slip at L1 and L2 will change the ionospheric phase by 0.28 cycles.).
During step (1) reports will be issued about bias parameters being added, and why blocks of data are being removed from the analysis (usual reason is too small an interval of good data between bias flags). These reports look like:
Marking EX-WL slip at epoch 553 Site 2 PRN 03 Size -24.18 cycles Mask 26
Marking EX-WL slip at epoch 554 Site 2 PRN 03 Size -23.93 cycles Mask 26
Marking EX-WL slip at epoch 555 Site 2 PRN 03 Size -24.13 cycles Mask 26
If the number of cycles is small (0.2–0.4 cycles) then the bias flag might not be needed. The first entry of the ION_STATS command controls the size of a jump flagged.
Some kinematic data have gaps because of recording problems rather than cycles slips.
If large numbers of bias flags are being added (typical behavior should be one per satellite per site), the BF_SET command can be used to increase the gap size that is automatically flagged.
Once all the biases have been assigned, approximate estimates are found from the mean Melbourne-Wubbena (MW) wide lane (this is combination of range and phase that gives the difference between the L1 and L2 biases and it independent of the ionospheric delay and the geometry), and the mean of \(LG = L1 - \left(\tfrac{f L1}{f L2}\right)L2\) phase values (called the ionospheric phase or extra-wide lane). The latter quantity is independent of geometry but will only equal zero when the ionospheric delay is zero. These mean values are computed using double differences. Some of the one-way bias values can be arbitarily set to zero when double differences are formed.
The bias parameter table is output by the program and has the following format:
BIAS FLAG REPORT: Type INITIAL
. # Site PRN Start Stop Fixd L1 cycles L2 cycles DD Bias Refs Mean MW-WL +- Mean Ion +- INITIAL
1 1 PRN 05 1 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
2 1 PRN 06 1 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
3 1 PRN 17 1 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
4 1 PRN 10 1 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
5 1 PRN 24 1 6422 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
6 1 PRN 30 1 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
7 1 PRN 15 3360 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
8 1 PRN 23 4103 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
9 1 PRN 26 4379 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
10 1 PRN 18 5232 6892 1 0.0 0.0 0 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
11 2 PRN 05 1 6892 1 0.0 0.0 -1 0 0 WL Res +- 0.000 0.000 0.000 0.000 INITIAL
12 2 PRN 06 1 6892 1 -8200515.0 -3079519.0 11 2 1 WL Res +- -0.257 0.150 0.007 0.011 INITIAL
13 2 PRN 17 1 6892 1 -2305580.0 -1862667.0 12 3 2 WL Res +- -0.427 0.171 -0.137 0.014 INITIAL
14 2 PRN 10 1 6892 1 -8768807.0 -6842458.0 11 4 1 WL Res +- -0.200 0.130 -0.003 0.007 INITIAL
15 2 PRN 24 1 2453 1 5173625.0 4017810.0 11 5 1 WL Res +- 0.037 0.196 0.029 0.007 INITIAL
16 2 PRN 24 2493 5739 1 5173625.0 -3947627.0 11 5 1 WL Res +- 0.091 0.336 -0.006 0.027 INITIAL
17 2 PRN 24 5982 6421 1 5173625.0 -7378939.0 12 5 2 WL Res +- 0.133 1.531 -0.048 0.070 INITIAL
18 2 PRN 30 1 6892 1 -7074955.0 -5537433.0 11 6 1 WL Res +- -0.176 0.137 0.008 0.009 INITIAL
19 2 PRN 15 3360 6892 1 -6106087.0 -4838049.0 12 7 2 WL Res +- 0.371 0.384 0.004 0.028 INITIAL
20 2 PRN 23 4103 6892 1 -7242531.0 -5594566.0 12 8 2 WL Res +- -0.480 0.346 -0.111 0.021 INITIAL
21 2 PRN 26 4379 6892 1 -7464580.0 -5773140.0 12 9 2 WL Res +- 0.342 0.414 0.016 0.023 INITIAL
22 2 PRN 18 5232 6892 1 -8632784.0 -6768836.0 12 10 2 WL Res +- -0.400 0.688 -0.124 0.023 INITIAL
The table gives a list of all bias parameters needed for the data being analyzed. The bias flag table is printed several times in the output. The different types of output of the table are:
- Type INITIAL
This indicates that these are initial approximate estimates.
- Type UNIFIED
This indicates that nearby bias parameters which have similar values have been assigned the value of the best approximate estimates. In the case above, the additional bias parameters are introduced due to data gaps (which are probably due to recording errors rather then loss of lock on the satellite since most of the values are nearly equal).
- Type INPUT
User input values when ambiguity file is passed through the command line or in the command file (see
AMBIN_FILEbelow)- Type FLOATXX
where XX is a number. This is the output from the FLOAT resolution of the bias parameters.
- Type PASSXX
where XX is a number. This is the output of the iterative fixing of the bias parameters to integer values. With each pass more bias parameters are fixed using the previously fixed ones. (The output of this part of the program is explained below.)
- Type FINAL
Final estimates of the bias parameters used in the kinematic trajectory determination.
The columns in the Table are:
- #
Bias parameter number. Runs sequentially from 1 to the number of bias parameters needed.
- Site
Site number (1 to number of sites)
- PRN
PRN number of the satellite to which the bias parameter refers.
Start
- Stop
Start and stop epoch numbers for the region where the bias parameter is needed. The epoch numbers are computed from the start time and increment by one for every epoch of data.
- Fixd
This column indicates the status of the bias flag. If the bias is fully resolved then the value will be 2 or 3 in the PASSXX and FINAL outputs of the bias flag table. A value of 1 means an approximate result fixed relative other defined biases, 0 means an approximate value that is estimate relative an another uncertain bias value (in double differences).
- L1 cycles
Estimates of the number of cycles needed at L1
- L2 cycles
Estimates of the number of cycles needed at L2
- DD Bias Refs
Indicate the other bias parameters used in forming the double differences for this bias parameter.
0 0or-1 0indidates that this bias has been set to zero. All biases at station 1 are set this way, and at least one entry at the remaining stations can be arbitarily set this way. (The longest contiguous block of data is used as the criteria).
Mean MW-WL +- Mean Ion +- INITIAL
- 12-15 (Numerical column numbers). These are the estimates of residual errors in the Melbourne-Wubbena (MW)
widelane and the Extra-wide lane and their expected sigmas based on a correlated noise model for the data (10-minute correlation time). The order is <MW WL> <Sigma> <EX WL> <Sigma>. The values and sigmas of these estimates plus the likely contribution of the ionospheric delay are used to compute the range of L1/L2 cycles to search over.
Final column is the label for the type of report.
Once the bias parameters have been established, track then, in a iterative loop, tries to fix the biases to the correct integer values.