Hướng dẫn bình sai GPS Huace X20 - chương VI

6 Chapter VI Network Adjustment
Normally, you want to do further checking on the obtained baseline results of the
processing and translate the baseline vectors to wanted-state coordinates and local
coordinates. That is what the network adjustment is going to do.
6.1 Network Adjustment: Functions and Steps
Briefly mentioned fundamental theory as well as approaches of free network
adjustment, 3-dimensional adjustment, 2-dimensional adjustment, height fitting can be
fulfilled in our software.
The figure below shows steps of network adjustment in our software.
There are actually three steps, as the figure shows:
•Preparation done by users themselves. This is to say, you need to enter the coordinate
settings, as well as know points, horizontal coordinates and height.
•Network adjustment processed by software.
•Quality analysis and control, made by users.
102

Coordi nate Sys&Adjustment Setup
Input Known Coordinates
Build vector network using baseline-vectors
Check the Network
3D Min-constraint Adjustment
3D
2D Height Fitti ng
Quality Control
87. Network Adjustment Steps
What the software does is barely computation part. Works that are more important are
left to the users themselves to make most of the proper decision. It is normally a
through and through process.
103


6.2 Preparations before Hand
6.2.1 Coordinate System Settings
You should check your coordinate system settings before setting up the network
adjustment settings. Normally, China-domestic users should select Beijing-54 as used
coordinate system ellipsoid
File
>
Properties
Coordinate settings are available in the dialog-box below:
88. Coordinate Systom Managment
In fact, parameters of Beijing-54 have been input into application when it is installed. In
addition, parameters of the coordinate system are entered when a new project is built.
The Coordinate settings here are only for re-checking these parameters.
For Details about the coordinate system settings, refer to relevant materials.
104


6.2.2 Network Adjustment Settings
6.2.2.1 Network Adjustment Settings
Select ‘Adjustment’ from Menu Bar > ‘Settings’. A dialog-box will appear, as shown
below.
There are three panels on this dialog-box: Network Adjustment Settings, 2-D
Adjustment Settings and Height Fitting Method.
89. Network Adjustment Settings
Here, you can select wanted network adjustment method, such as 3D network
adjustment, 2D network adjustment and height fitting. Free network adjustment is not
available here because the software has done the free network adjustment before
doing the above adjustment.
105


You can also reset the central meridian here. Normally, Domestic users only need one
set of ellipsoid parameters-Beijing 54 with different values of central meridian in
different areas.
6.2.2.2 2D Adjustment Settings
Only geodetic coordinates in WGS-84 comes out from free adjustment, there is a need
of state/local projection coordinates. To obtain these coordinates, you need to run a
combined measurement with some known points in the static baseline network. Then,
consequently, other points in the network can be translated into required coordinates.
The 2D combined adjustment is the most often used method.
When the free adjustment is done, four parameters are needed to translate the
obtained geodetic coordinates in WGS-84 into projection coordinates precisely. There
are a couple of ‘moving’ parameters, a ‘rotate’ parameter and a scale parameter. 2D
adjustment settings dialog-box is shown in the figure below:
90. 2D Ajustment Settings
106

You can have different options of for these parameters. The Default choice is ‘Moving,
Rotation, and Scales’. It is highly recommended in common situation that the default
option should be selected.
107


6.2.2.3 Height Fitting
Some network adjustment setting methods have been introduced in the previous
chapters. You can select a method in ‘Height Fitting Methods’ panel. As shown in the
figure below, the default selection is ‘Surface Fitting’
91. Height Fitting Methods
108


6.2.3 Input know Point
You should input the known points’ coordinates after setting up the network adjustment
parameters. Right click and select ‘Properties’ on list window in the Observation
Stations section, a dialog-box will appear, as shown below.
Enter the known points’ ‘fix method’ and ‘fixed coordinates’.
Please do not forget to tick constrained!
92. Set Up Known Points
Notes:
1. Every known point of a project should be in the same system
2. Make know points reasonably distributed
3. It must be determined, which of the following heights is used: geodetic height
(ellipsoidal height), trigonometric height and leveled height - 3D adjustment requires
geodetic height while height fitting requires ellipsoidal height. If there is a need for of
entering both geodetic height
and leveled height in the same project, you should set
another project to deal with them one after another.
109

6.3 Running Network Adjustment
To run the network adjustment, select ‘
Network Adjustment’>’Run Adjustment
’ from the
Pop-up Menu;
OR
click ‘Network Adjustment’ button on the Tool Bar.
The software will do the adjustment based on known baseline computation result,
network adjustment settings and observation station coordinates.
6.3.1 Derive Baseline Vectors Network
The first step of network adjustment is deriving the baseline vectors network. The
Following principles are applied to forma baseline vectors network:
1. Given that the baseline exists in the project and is not deleted.
2. Given that the baseline has both starting point name and computing point name.
3. Given that the baseline has been computed and is shown ‘qualified’ on the baseline
vectors list.
4. Given that the baseline has is not been set as ‘not involved in the baseline
computation and network adjustment’
Any baseline that meets all four requirements above will be automatically loaded to
form the baseline vectors network.
6.3.2 Connectivity Check of Baseline Vectors Network
If you run the adjustment without a positive connectivity, it may lead to a negative
co nstring ency. Therefore, before running the adjustment, the software will self-check
the connectivity. If negative connectivity is found, a warning information will be shown:
You should check the baseline vectors and observation names that form the network.
The Connectivity can also be checked in the properties of the network.
110



6.3.3 Free Adjustment
After connectivity checking, the software carries out free adjustment, of in which the
brief theory has been given in the previous chapter.
The Results of the free adjustment will be shown in the
baseline list window
,
Observation Station Window
and
Detailed Result Window
.
The
Baseline list window
shows the residual and the adjusted value of every baseline
involved in free adjustment, as shown in the figure below:
93. Baseline Adjustment
(Problematic Display)
The
Observation Station Window
shows the free adjustment coordinates of the
observation stations, as shown in the figure below:
94. Free Adjustment Coordinates
(Blank fields due to the lack of 3D,2D,Height data
in testing file)
111


Right-click and select
Detailed Result
shows the residual and the adjusted value of
every baseline involved in the free adjustment, as shown in the figure below:
95. Detailed Result
You can run free adjustment before setting up the adjustment parameters. Sometimes,
free adjustment is also used to check the quality of the baseline computation.
112


6.3.4 3-Dimensional Constrained Adjustment
If ‘3-Dimensional Constrained’ is chosen in the
Network Adjustment Settings
, and BL
(longitude & latitude) or BLH ((longitude, latitude & height) is constrained on at least
one observation point in the network, 3-Dimensional constrained adjustment would be
carried out here.
96. Adjustment Settings
The 3-Dimensional constrained adjustment determines the number of unknowns
according to the constraint condition of the observation points. For example, if 3 or
more than 3 known points with fixed BLH are selected, 7 parameters will be derived
from two bases. Under an insufficient constraint conditions, transformed parameters
will be selected according to given conditions. Three constraint conditions, for example,
will only be derived from three ‘move’ parameters.
113



The Result of the 3-Dimensional constrained adjustment is shown in the
Observation
Station Window
and
Detailed Result Window
. The figure below shows the adjustment
information in the Observation Station Window
.
97. 3D constrained adjustment coordinates
(Blank fields due to the lack of
3D,2D,Height data in testing file)
The Coordinates and their errors derived from the 3-Dimensional constrained
adjustment are shown in the tables in the
Detailed Result Window
.
98. 3D constrained adjustment result
6.4 2-Dimensional Constrained Adjustments
If the ‘2-Dimensional Constrained is chosen in the
Network Adjustment Settings
, and xy
(north, east) is constrained on at least one observation point in the network, the
2-Dimensional constrained adjustment would be carried out.
The result of the 2-Dimensional constrained adjustment is shown in the
Observation
Station Window
and
Detailed Result Window
. The figure below shows the adjustment
information in the
Observation Station Window.
114



99. 2D constrained adjustment
(Blank fields due to the lack of 3D,2D,Height data
in testing file)
The table below shows the 2-Dimensional constrained adjustment information in
detailed result. The following information is shown: 4 transform parameters derived
from adjustment, plane distance between each pair of points and its error and plane
coordinates/errors of each station.
100.
2D constrained adjustment result
(didn’t deliver given report)
6.5 Height Fitting
If the ‘Height Fitting’ is chosen in the
Network Adjustment Settings
, and the height of
one of the BLH, xyH or H is constrained on at least one observation point in the network,
the 2D constrainted adjustment would be carried out.
The result of height fitting is shown in the
Observation Station Window
and
Detailed
Result Window
. The figure below shows the adjustment information in the
Observation
Station Window.
115



101. Height Fitting
(Blank fields due to the lack of 3D,2D,Height data in testing file)
The table below shows the height fitting information in the detailed result. The following
information is shown: height fitting method and derived transform parameters as well as
.
fitted height and it errors. Refer to figure below
102. Height Fitting and Its Error
(didn’t deliver given report)
6.6 Result Checking
It is necessary to check the network adjustment results. The adjustment results are
mainly estimated by residual, mean square error and accordingly statistic checking
results, etc.
For details about statistic checking, see the next section.
You can seek the reason for poor adjustment result by using the following hints:
1. Did you set a correct coordinate system?
2. Were the known points correct and were they in the same system?
3. Was every baseline in the network correct? You can delete unqualified static
baselines to exclude them from the adjustment. If you cannot delete an unqualified
baseline, or if it is important in your network, you have to run the computation again,
116

or you even have to redo your fieldwork.
4. Were the observation station data and antenna height correct? Incorrect station data
and antenna height often lead to a very poor closing Error or free adjustment result.
6.7 Statistic Checking of GPS Network
The minimum constrained free adjustment that is the estimating tool of the accuracy on
the network itself, aims to the elimination of discrepancy caused by redundant
measurements errors. Therefore, what is the criterion of the accuracy? The answer lies
in two statistic-checking methods in the Compass: overall data focusing method,
2
check, of which the pass/failure status is shown in the adjustment-result-files, and
individual object aiming method, check, of which failed values (whose is larger than
1.00) are taken as containing gross errors and would be deleted. Both
and
are
2
shown in the results.
6.7.1
Check
2
2
This method checks the variance of the unit weight
, say, checking if the estimated
0
ˆ
2
2
value
after the adjustment is in line with
before the adjustment.
0
0
If:
V
PV
V
PV
'
'
<
<
2
0
2
2
/
2
1
/
2
ˆ
2
2
Then
is in line with
0
0
square checking
You can view the ‘
’ to check whether the adjustment passed
. If the
2
adjustment failed to pass, you need to find the reason and make it succeed by
re-processing the baselines or by deleting the poor values.
Notes: The adjustment often fails to pass 2 if you use the default value without a
proper weight setting. In this software, you can make the adjustment pass it by setting
up its
Covariance Proportionality Factor
and
Variance Factor
. By default, both two
factors are ‘1’.
117



Normally, by simply modifying the
Covariance Proportionality Factor
makes the
adjustment pass but has the least influence on coming up Check.
An example is shown in the figure below, of which the adjustment did pass
Check at
2
first. Under the assumption that every observed value was qualified, we found that the
reference factor was 3.57. Therefore, we modified the
Covariance Proportionality
Factor
to 3.5 and run the adjustment again. It passed this time.
103.
Check Failed
2
104.
Adjustment Settings
118


Pass
Check
2
105.
Passed
The data quality is the key point of
Check. And the proper setting of the
Covariance
2
Proportionality Factor
is also very important. When the adjustment passes
Check,
2
you need to apply the Check to each of the observed value.
6.7.2 Check
The gross errors can be detected according to the value of the baseline vector residual.
If:
<
ˆ
v
q
t
i
0
i
1
/
2
Where,
v
: Residual error of the value number i
i
ˆ
: Variance of the unit weight
0
q
: Factor of the value number i
i
t
t
: Distribution interval of
in significance level
1
/
2
The software offers a ‘ ’, which is the ratio of the estimated value, and actual . If this
ratio is smaller than 1.0, then the according value should not be removed. While a ratio
that exceeds 1.0 should be removed.
Besides listed Check value, the Detailed Result Window also offers a distribution
chart of , as shown in the figure below. Columns on background stand for the
theoretical distribution of and the histogram in blue is practical distribution. In this
119



figure, see some Check values exceed acceptable range.
Baselines
didn’t pass
the Check
106. Not Every Baseline Passed
On the list of baseline vectors and the list of Check, the baselines whiich did not pass
the Check is marked in red. An example is given in the figure below:
107.
A Baseline Doesn’t Pass Check,
According to this figure, there is still one baseline that cannot pass Check, despite the
positive result of every baseline in the previous
Check. We found an incorrect
2
antenna height exist in a file that is related to this baseline. We entered the correct
antenna height and re-processed it, then ran the adjustment again. Every baseline
120


involved in the adjustment passed the Check this time.
All baselines
are located in
the section of
Abs. 1
108.
Pass Check
It has no influence on the overall structure of the network adjustment if there some
baselines overlap on this baseline or it is deleted. Therefore, the result meets all
requirements when you delete this baseline and run the adjustment again.
Besides the two solutions mentioned above, you can also have the
Check passed by
2
modifying the
Variance Factor
in the
General
dialog-box to, bring the changed weight of
a single baseline into computation. Refer to the figure below:
121


109.
Modify Variance Factor to Change the Weight
Normally, under the assumption that the network meets the requirement and the
2
baseline computation results are all qualified, the network will pass both
Check and
check and accomplish the minimum constrained three-dimensional adjustment
smoothly.
•
Conclusions and Suggestions
1. In the network adjustment, these two checking methods concerns the whole network
and individual element of it respectively. You need to pay attention on the adjustment
result file. Only when the adjustment passes both two checks, can it be taken as a
‘passed’.
2. The quality of the data involved in the adjustment is crucial for the adjustment checks.
Therefore, it is necessary to edit and reprocess or even to delete and re-measure
some data.
122

3. The Covariance Proportionality Factor setting plays an important role. Properly
selected Covariance Proportionality Factor and Variance Factor are important to
having a satisfied adjustment result.
4. The requirements in manufacturing and practicing also determine the quality of the
adjustment. And the Statistic checking is only part of the adjustment checking
procedure; your attentions should not be focused on it.
123