GPS500 General Guide

7/25/2019 GPS500 General Guide

1/42
http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/http://gps500.pdf/


2/42

2

Congratulations on your purchase of a newSystem GPS500 from Leica Geosystems.
http://gps500.pdf/http://gps500.pdf/


3/42

3

Introduction

Overall planning for a GPS survey

Mission planning

Observation times and baseline lengths

Field observations

Importing the data to SKI-Pro

Deriving initial WGS 84 coordinates for one point

Data-processing parameters

Baseline selection - Strategy for computation

Interpreting the baseline results

Inspecting the logfile and comparing results

Storing the results

Adjustment, Transformation and output of results

Notes on single-frequency Static and Rapid Static measurements

6

7

13

15

17

22

23

24

30

32

35

37

39

40


4/42

4

Introduction ................................................... 6

Overall planning for a GPS survey .............. 7

Baseline length ............................................................ 7Temporary reference stations for Rapid Static surveys . 8Check the newly surveyed points ................................ 9Night versus day observations. Measuring long lines .. 10Observation schedule - .............................................. 10best times to observe ................................................ 10Consider the transformation to local coordinates ......... 11

Mission planning ......................................... 13GDOP - Geometric Dilution of Precision..................... 13Selecting good windows for successful GPS surveying13

Observation times and baseline lengths... 15

Field observations....................................... 17

Reference site ........................................................... 17Need for one known point in WGS 84............................. 18

Observing new points ................................................ 19Use the Stop and Go Indicator as a guide ...................... 19

Fill out a field sheet .................................................... 20

Importing the data to SKI-Pro .................... 22

Checking and editing during data transfer .................. 22Backing up raw data and projects .............................. 22

Deriving initial WGS 84 coordinates forone point ...................................................... 23

Data-processing parameters ...................... 24

Cut-off angle ............................................................. 24Ephemeris ................................................................ 25

Data used for processing ........................................... 25Fix ambiguities up to: ................................................. 26Rms threshold........................................................... 26Solution type ............................................................ 28Ionospheric model ..................................................... 28Use stochastic modelling ........................................... 29Frequency ................................................................. 29Tropospheric model................................................... 29

Baseline selection - Strategy forcomputation................................................. 30


5/42

5

Interpreting the baseline results ............... 32

Baselines up to the limitation value............................. 33Ambiguities resolved ...................................................... 33

Ambiguities not resolved................................................. 34Baselines above the limitation value ........................... 34

Inspecting the logfile and comparingresults .......................................................... 35

Baselines up to the limitation value............................. 35Baselines above the limitation value ........................... 36Compare the logfile against the field sheets ............... 36

Compare the results for double fixes .......................... 36

Storing the results....................................... 37

Adjustment, Transformation and output ofresults .......................................................... 39

Notes on single-frequency static and rapidstatic measurements................................... 40


6/42

6

Although this guide has been writtenspecifically for Leica GeosystemsGPS - System 500 and System 300,much of the information is of ageneral nature and applicable to allGPS surveying. Further informationmay be found in the variousguidelines contained in the System500 or System 300 documentationmaterial.

Surveying with GPS has becomepopular due to the advantages ofaccuracy, speed, versatility andeconomy. The techniques employedare completely different however,from those of classical surveying.

Provided that certain basic rules arefollowed GPS surveying is relativelystraightforward and will produce goodresults. From a practical point of viewit is probably more important to

understand the basic rules forplanning, observing and computingGPS surveys rather than to have adetailed theoretical knowledge of theGlobal Positioning System.

This guide outlines how to carry outStatic and Rapid Static GPS surveysand emphasizes those points to

which particular care has to be paid.


7/42

7

Rapid Static surveys feature short

observation times. It is particularlyimportant for Rapid Static thationospheric disturbances are more orless identical for both sites.

Thus, for all GPS surveying, and forRapid Static in particular, it is soundpractice to minimize baseline lengths.

A GPS receiver measures the

incoming phase of the satellitesignals to millimeter precision.However, as the satellite signalspropagate through space to earththey pass through and are affectedby the atmosphere. The atmosphereconsists of the ionosphere and thetroposphere. Disturbances in theatmosphere cause a degradation inthe accuracy of observations.

GPS surveying is a differentialmethod. A baseline is observed andcomputed between two receivers.When the two receivers observe thesame set of satellites simultaneously,most of the atmospheric effects

cancel out. The shorter the baselinethe truer this will be, as the morelikely it is that the atmospherethrough which the signals pass to thetwo receivers will be identical.


8/42


9/42

9

Depending on the accuracy required,the user should be prepared to checknewly surveyed points. This isparticularly important if observationtimes have been cut to a minimumand recommendations regardingGDOP ignored.

For a completely independent check:

Occupy a point a second time in a

different window. This ensures thatthe set-up, the satellite constellat-ion, and the atmosphericconditions are different.

Close a traverse loop with abaseline from the last point to thestarting point.

Measure independent baselinesbetween points in networks

A partial check can be obtained byusing two reference stations insteadof one. You will then have two fixesfor each point but each will be basedon the same roving-receiverobservations and set-up.

In all types of survey work it is soundpractice to cross check using inde-pendent measurements. In classicalsurvey you check for inaccurate orwrong control points, wronginstrument orientation, incorrectinstrument and target heights, etc.You close traverses and level loops,you fix points twice, you measurecheck distances! Depending on thejob and accuracy needed it is wellworthwhile applying the same

principles to GPS surveying.

One should be particularly carefulwith Rapid Static with shortobservation times. If the observationtime is too short, or the satellitegeometry (GDOP) is poor, or theionospheric disturbances are verysevere, it can happen that the post-

processing software will resolveambiguities but the results mayexceed the quoted specifications.


10/42

10

For baselines up to about 20 km, onewill usually attempt to resolve theambiguities using the Rapid Staticalgorithm in SKI-Pro post-processingsoftware.

For baselines over 20 km, it isusually not advisable to resolveambiguities. In this case a differentpost-processing algorithm is used inSKI-Pro. This algorithm eliminatesionospheric influences to a large

degree but destroys the integernature of the ambiguities.

Generally speaking, the longer thebaseline the longer one has toobserve.

The ionosphere is activated by solarradiation. Thus ionosphericdisturbance is much more severe byday than by night. As a result, thebaseline range for night observationswith Rapid Static can be roughlydouble that of day observations. Or,put another way, observation times

for a baseline can often be halved atnight.

At the present time ionosphericactivity is increasing in an 11-yearcycle.

The table on page 16 provides aguide to baseline lengths and

observation times under the currentionospheric conditions.

When you inspect the satellitesummary and GDOP plots, you willusually see several good windows(see page 14) distributed through a24 hour period. You should try towork with Rapid Static during goodwindows, and plan your schedulecarefully.

It is impossible to plan GPSobservations to the minute. Ratherthan trying to squeeze the maximum

number of points into a window bycutting observation times to the bareminimum, it is usually better tomeasure one point less and toobserve for a few minutes longer.Particularly for high-accuracy work, itpays to be conservative and not torisk poor results.


11/42

11

The common points should bespread evenly throughout the projectarea. For a correct computation of alltransformation parameters (shifts,rotations, scale), at least three - butpreferably four or more - points haveto be used.

Read the Guidelines to Datum/Map inthe SKI-Pro Documentation fordetails on transformation usingDatum/ Map.

System 500 and System 300 provideaccurate relative positions of pointsthat are observed in a GPS networkand linked in post-processing. Thecoordinates are based on the WGS84 datum.

For most projects it will be necessaryto transform the WGS 84 coordinatesobtained from GPS survey into localgrid coordinates, i.e. into gridcoordinates on the local projection

based on the local ellipsoid.

In order to be able to compute thistransformation, known points withlocal coordinates have to be includedin the GPS network. These commonpoints, with WGS 84 and localcoordinates, are used to determinethe transformation parameters and to

check the consistency of the localsystem.


12/42


13/42

13

Poor windows should only be used tobridge between two or more goodwindows when observing for longperiods of time, e.g. at referencestations and for long lines.

If there are obstructions near a point,use the sky plot to find out if thesignals from a satellite could beblocked. This could cause the GDOPto deteriorate. Check the GDOP byclicking the satellite "off" in theSurvey Design component. A carefulreconnaissance of such sites is wellworthwhile.

The GDOP value helps you to judgethe geometry of the satelliteconstellation. A low GDOP indicatesgood geometry. A high GDOP tellsyou that the satellite constellation ispoor. The better (lower) the GDOPthe more likely it is that you willachieve good results.

Poor satellite geometry can becompared with the "danger circle" ina classical resection. If the geometryis poor, the solution in post-processing will be weak.

For Rapid Static you should observewhen the GDOP is less than or equal

to 8. A GDOP of 5 or lower is ideal.

For successful, high-accuracy GPSsurveying it is advisable to take theobservations in good windows.Provided that you know the latitudeand longitude to about 1, thesatellite summary, GDOP, elevation,and sky-plot panels in the SurveyDesign component of SKI-Pro willhelp you to select good windows inwhich to observe.

You should take particular care whenselecting windows for Rapid Staticobservations.

A suitable observation window forRapid Static must have four or more

satellites, with GDOP 8, above acut-off angle of 15 at both thereference and roving receiver.


14/42

14

Selecting Good Windows

Window for Rapid Static:

4 or more satellites above 15 cut-off angle.

GDOP 8.

Whenever possible:

5 or more satellites.

GDOP 5.

Satellites above 20.

Always:

Use sky plot to check for obstructions.

Recompute GDOP if a satellite is obstructed.

Be wary if 2 out of 4 or 5 satellites are low (


15/42


16/42

16

Times and Baseline Lengths

Observation time depends upon:

Baseline length Number of satellites

Satellite geometry (GDOP)

IonosphereIonospheric disturbance varies with time, day/night, month, year, positionon earth's surface.

The table provides an approximate guide to baseline lengths and observationtimes for mid latitudes under the current levels of ionospheric activity whenusing a dual frequency Sensor.

Obs.Method

No. sats.GDOP

8BaselineLength

Approximate time

observation

By day By night

RapidStatic

4 or more4 or more5 or more

Up to 5 km5 to 10 km

10 to 15 km

5 to 10 mins10 to 20 minsOver 20 mins

5 mins5 to 10 mins5 to 20 mins

Static 4 or more4 or more

15 to 30 kmOver 30 km

1 to 2 hours2 to 3 hours

1 hour2 hours


17/42

17

Note that the reference

receiver does not have to beset up on a known point. It isfar better to establishtemporary reference stationsat sites that fulfill therequirements listed abovethan to set up the referencereceiver on known points thatare not suitable for GPSobservations.

For computing the transformationfrom WGS 84 to the local system,known points with local coordinateshave to be included in the GPSnetwork. These points do not have tobe used as reference stations. They

can be measured with the rovingreceiver.

GPS surveying is a differential

technique with baselines being"observed" and computed from thereference to the rover. As manybaselines will often be measuredfrom the same reference station, thechoice and reliability of referencestations are of particular importance.

Sites for reference stations should bechosen for their suitability for GPSobservations. A good site shouldhave the following characteristics:

No obstructions above the 15 cut-off angle.

No reflecting surfaces that couldcause multipath.

Safe, away from traffic andpassers-by. Possible to leave thereceiver unattended.

No powerful transmitters (radio,TV antennas, etc.) in the vicinity.

The results for all roving points will

depend on the performance of thereference receiver! Thus thereference receiver must operatereliably:

Power supply must be ensured.Use a fully-charged battery.Consider connecting two batteries.When possible, consider a

transformer connected to themains.

Check that there is ample capacityleft in the memory device forstoring all observations.

Double-check the antenna heightand offset.

Make sure that the missionparameters (observation type,recording rate etc.) are correctlyset and match those of the rovingreceiver.


18/42

18

The computation of a baseline indata processing requires that thecoordinates of one point (reference)are held fixed. The coordinates of theother point (rover) are computedrelative to the "fixed" point.

In order to avoid that the results areinfluenced by systematic errors, thecoordinates for the "fixed" point haveto be known to within about 20meters in the WGS 84 coordinate

system. Whenever possible, theWGS 84 coordinates for the "fixed"point should be known to within about10 meters otherwise scale errors ofabout 1 to 2 ppm will be introduced.

This means that for any precise GPSsurvey the absolute coordinates ofone site in the network have to be

known in WGS 84 to about 10meters. WGS 84 coordinates for onesite will often be available or can beeasily derived as explained on page23.

If WGS 84 coordinates for one siteare not known or cannot be derived,the Single Point Position computationin SKI-Pro can be used. Remember,however, that Selective Availability(SA) may be switched on. The onlyway to overcome SA is to observe forsufficient time for the effects of SA tobe averaged out in the Single PointPosition computation.

The reference receiver will usually

observe for several hours as therover moves from point to point. Insuch a case, the Single Point Positi-on for the reference receivercomputed in SKI-Pro should berelatively free from the effects of SA.If a Single Point Position is computedfrom only a few minutes ofobservations, the effects of Selective

Availability will not be averaged out.The result could be wrong by 100mor more due to SA.

When computing the Single PointPosition for the starting point of anetwork, always compute for a sitefor which you have several hours ofobservations. The resulting WGS 84coordinates should then be correct towithin about 10 meters.

The minimum observation for thecomputation of a reliable Single PointPosition is probably about 2 to 3hours with four or more satellites and

good GDOP. The longer theobservation time, the better theSingle Point Position will be.


19/42

19

As the Stop and Go Indicator canonly monitor the roving receiver it canonly provide an estimate for therequired measuring time. It should beused only as a guide.

The operator of the roving receivershould also pay attention to certainpoints. This is particularly importantfor Rapid Static surveys with shortmeasuring times.

Make sure that the configurationparameters (e.g. recording rateetc.) are correctly set and matchthose of the reference receiver.

Check the antenna height andoffset.

Watch the GDOP when observingfor only a short time at a point.

For 5 to 10mm + 1 ppm accuracywith Rapid Static, only takemeasurements with GDOP 8.

The Stop and Go Indicator on thesensor provides the roving-receiveroperator with an approximate guideto measuring times for Rapid Staticobservations with four or moresatellites and GDOP less than orequal to 8. It estimates whensufficient observations should havebeen taken for successful post-processing (ambiguity resolution) tobe possible.

At the present time estimates arecalculated for two baseline ranges,0 to 5 km and 5 to 10 km. Theestimates are based approximatelyon the current situation for GPSobservations in mid latitudes andassume that the reference and rovingreceiver are tracking the samesatellites.


20/42

20

Reference Stations

No obstructions above 15 cut-off angle.

No reflecting surfaces (multipath).

Safe, can leave equipment unattended.

No transmitters in vicinity.

Reliable power supply.

Ample memory capacity.

Correct configuration parameters (e.g. recording rate).

Check antenna height and offset.

Does not have to be a known point.

It is better to establish temporary reference stations atgood sites rather than at unsuitable known points.

For precise GPS surveying, WGS 84 coordinates for onepoint have to be known to about 10 meters.

Roving Receiver

15 cut-off angle.

Obstructions should not block signals.

No reflecting surfaces (multipath).

No transmitters in vicinity.

Fully-charged battery.

Sufficient memory capacity.

Correct configuration parameters (e.g. data-recordingrate).

Check antenna height and offset.

Observe in good windows.

Watch the GDOP 8.

Use Stop and Go Indicator as a guide.

Fill out a field sheet.

As with all survey work, it is well worthwhile filling out afield sheet for each site when taking GPS observations.Field sheets facilitate checking and editing at the data-processing stage.


21/42

21

Practical Hints

Tribrachs: check the bubble and optical plummet.

Level and center the tribrach and tripod correctly.

Check the height reading and antenna offset.

An error in height affects the entire solution!

Use a radio to maintain contact between reference androver.

Consider orienting the antennas for the most precisework.

Field Sheet

Point Id.: Date:

Receiver Serial No.: Operator:

Memory card No.:

Type of set up:

Height reading:

Time started tracking:

Time stopped tracking:

Number of epochs:

Number of satellites:

GDOP:

Navigation position: Lat. Long. Height

Notes:


22/42

22

X

X

X

X

X

Data can be transferred to SKI-Prodirectly via a PC-card slot, or via acard reader, from the controller(System 300) or receiver (System500), or from a disk with backed-upraw data. During data transfer, theoperator has the opportunity to checkand edit certain data. It is particularlyadvisable to check the following:

Point identification: Check spelling,upper and lower case letters,spaces etc.

Make sure that points that havebeen observed twice have thesame point identification. Makesure that different points in the

same project have different pointidentifications.

Height reading: Compare with fieldsheets.

Note that some of the abovesite-related parameters canbe changed in somecomponents of SKI-Pro.However, the affectedbaselines have then to berecomputed.

After reading in a data set alwaysmake a back-up on either a disketteor on the hard disk. You can thenerase and reuse the memory cardbut you still have the raw data. Whenbacking up data from severalmemory cards, it is advisable tocreate a directory for each card.

After importing all the data related tothe project it is often worthwhilemaking a backup of the wholedirectory where the project is locatedbefore starting to process the data.


23/42

6X

7X

13X

15X

17X

23

As explained on page 18, thecomputation of a baseline requiresthat the coordinates of one point areheld fixed. The coordinates of theother point are computed relative tothe "fixed" point.

For any precise GPS survey theabsolute coordinates of ONE site inthe network have to be known inWGS 84 to about 10 meters. WGS84 coordinates for one site will often

be available or can be easily derived.

Using SKI-Pro it is easy to convertthe grid coordinates of a known pointto geodetic or Cartesian coordinateson the local ellipsoid. If theapproximate shifts between the localdatum and WGS 84 are known,WGS 84 coordinates to well within

the required accuracy can bederived. The local Survey Depart-ment or University will usually be ableto provide approximatetransformation parameters.

As explained on page 17, thereference receiver does not have tobe on a known point. If the referencereceiver was on a new (unknown)point and a known point wasobserved with the roving receiver,simply compute the first baselinefrom the known point (rover) to theunknown point (reference) in order toobtain and store the required initialWGS 84 coordinates for thereference receiver.

If good initial WGS 84 coordinates forthe reference site are not known orcannot be derived as explained in thelast two paragraphs, the Single PointPosition computation in SKI-Pro canbe used. When using the SinglePoint Position computation alwayscompute for a site for which there are

several hours of observations. Theeffects of Selective Availability shouldthen average out and the resultingWGS 84 coordinates should becorrect to within the required 10meters.

See section"Need for one knownpoint in WGS 84" on page 18 forfurther details.

Always keep in mind that poor initialcoordinates for the reference receiverwill affect the baseline computationand can lead to results outside thequoted specifications.


24/42

24

In the vast majority of cases, thedefault settings for data-processingmay be accepted and may never bealtered by the operator. On somerare occasions the operator mayneed to modify one or more of thedata processing parameters. Themost common ones are describedbelow.

It is common practice in GPSsurveying to set a 15 cut-off angle inthe receiver. 15 is also the systemdefault value in data processing.Avoid cut-off angles less than 15 ifprecise results are to be obtained.

Although you can increase the cut-offangle you should be cautious whendoing so. If the cut-off angle for dataprocessing is set higher than in thereceiver some observations will not

be used for the baseline computationand you may "lose" a satellite. Itcould happen that only threesatellites would be used in thecomputation instead of four. Youcannot expect a reliable answer withonly three satellites.

It can sometimes be advantageous,however, to increase the cut-off angleto about 20 in case of a disturbedionosphere and provided thatsufficient satellites above 20 withgood GDOP have been observed(use the Satellite Availabilitycomponent in SKI-Pro to check theGDOP).

You may sometimes find that abaseline result is outside

specifications even though fivesatellites have been observed. If oneof the satellites never rises aboveabout 20 the observations to thissatellite may be badly affected by theionosphere. Raising the cut-off angleand computing with only four high-elevation satellites can sometimesproduce a better result.


25/42


26/42

26

With this parameter you candetermine how SKI-Pro shouldcompute baselines. The systemdefault value is 20 km.

For baselines up to this limitationvalue, L1 and L2 measurements areintroduced as individual observationsinto the least-squares adjustment.The Lambda search developed byProf. Teunissen and his co-workersat the TU Delft is used as an efficient

approach to find possible candidatesets of integer ambiguities. Thestatistical decision criteria used hasbeen published previously togetherwith a different search algorithm, theFast Ambiguity Resolution Approach(FARA) by Dr. E. Frei and is nowcalled FARA statistics.

For baselines above this limitationvalue, a so-called L3 solution isperformed. The L3 observable is alinear combination of the L1 and L2

measurements. The advantage of theL3 solution is that it eliminates theinfluence of the ionosphere.

However, it also destroys the integernature of the ambiguities, thereforeno ambiguity resolution can becarried out. This is not important,

however, as successful ambiguityresolution over long distances is inany case hardly feasible.

The Rms threshold is used tominimize the possibility of unreliablebaseline results.

During the computation of a baseline,the least-squares adjustmentcomputes the root mean square(rms) of a single-difference phaseobservation (i.e. the rms of unitweight). This value is compared withthe Rms threshold.

For most GPS surveying applicationsone will usually accept the systemdefault "Automatic". This willautomatically select an appropriaterms parameter depending on theduration of your occupation.

The rms of a single-difference phaseobservation is largely dependent on

the baseline length, observation time,and ionospheric disturbance.Ionospheric disturbance is less atnight.


27/42

27

The following table provides a very approximate guide to the rms of asingle difference that a user could expect:

If the rms of a single-difference observation exceeds the rms threshold,the baseline solution with fixed ambiguities will be rejected and only thefloat solution will be presented (ambiguities not resolved).

Note, that the advanced parameter "Use stochastic modelling"(see page 29) will additionally reduce the rms values of a singledifference.

For Rapid Static observations with up to 10 minutes of measurementtime, one should be cautious about increasing the rms threshold becausean unreasonably high rms value could lead to a weak solution beingaccepted.

For longer observation times - let ussay about 30 minutes or more - therms threshold can be set higherwithout undue risk.

Note that the rms thresholdapplies only to baselines up

to the limitation value (see page 26).For baselines above the limitationvalue ambiguity resolution is notattempted.

Distance Day Observation Night Observation

10 min

10 min

10 min

10 min

Up to 5 km


28/42

28

The solution type parameter appliesto all baseline up to whichambiguities are attempted to be fixed(see page 26). If solution type "Stan-dard"is chosen, SKI-Pro will attemptto fix ambiguities and applyionospheric corrections as defined inthe parameter "Ionospheric model".

If solution type "Iono free fixed" ischosen then the baselinecomputation is done in two steps.

First ambiguities are attempted to befixed, then in the second step anionospheric free solution is calculatedusing fixed L1 and L2 ambiguities.

The advantage of this approach isthat any ionospheric disturbance iseliminated while fixed ambiguities areused; it is recommended to choose

this solution type for all baselinesbetween 5 km and 20 km, inparticular if daylight observationshave been taken.

This parameter is only used forbaselines up to the limitation value(see page 26, "Fix ambiguity up to"),that is for baselines for which SKI-Pro will try to resolve ambiguities.

The default parameter is"Automatic", which will automaticallyselect the best possible choice. Ifsufficient observation time isavailable on the reference, this willbe the "Computed model". In any

other case the "Klobuchar model"will be taken provided that almanacdata is available. Typically there isno need to change the default.

A "Computed model"may be usedinstead of the standard model. Thisis computed using differences in theL1 and L2 signal as received on the

ground at the Sensor.

The advantage of using this model isthat it is calculated according toconditions prevalent at the time andposition of measurement. At least 45

minutes of data is required for aComputed modelto be used.

The Standard model is based on anempirical ionospheric behaviour andis a function of the hour angle of thesun. When the Standard model ischosen corrections are applied to all

phase observations. The correctionsdepend on the hour angle of the sunat the time of measurement and theelevation of the satellites.

For long lines above the limitationvalue (see page 26), the ionosphericeffects are eliminated by evaluating alinear combination of L1 and L2

measurements, the so-called L3observable. Ambiguity resolution isnot attempted.


29/42

29

Using this option may supportambiguity resolution on medium andlonger lines when you suspect theionosphere to be quite active.

You should, however, be careful withshorter baselines since bad data -e.g. data influenced by multipath orobstructions- may be misinterpretedas being influenced by ionosphericnoise.

This is why by default this setting isonly used for baselines longer than10 km.

Note that in order to ensurereliable results this option

will not be used for the processing ofkinematic data.

SKI-Pro will automatically select toprocess whatever data is available.Thus there is little point in processingwith anything but "Automatic".

Short observation times with RapidStatic are only possible with dual-frequency observations. Long linescan only be processed successfullyusing L1 and L2 data.

Selecting "Iono free float" makes

SKI-Pro compute an L3 solution evenif the baseline length remains underthe limit to fix ambiguities (see p.26).Remember, that for an L3 solutionthe observation time has to be longenough.

It will not make much difference tothe end result as to whether youselect the Hopfieldor Saastamoinenmodel, but you should never work

with "No troposphere". You cannotexpect to achieve good results if notropospheric model is used.


30/42


31/42


32/42

32

When interpreting the results, onehas to distinguish between baselinesup to the limitation value ("Fixambiguities up to") and baselines

above this value (see page 26).

For baselines up to the limitationvalue, ambiguity resolution using theLambda search and the FARAstatistics is always attempted.

For baselines above the limitationvalue, a so-called L3 solution (linearcombination of L1 and L2measurements) is performed. This

eliminates the ionospheric effects butdestroys the integer nature of theambiguities. Thus ambiguityresolution is not carried out.


33/42

33

These will usually be the "truevalues".

However, one should also be aware

that very severe ionosphericdisturbances can cause systematicbiases in the phase observations. Inthis case, although the results of theleast-squares adjustment will bestatistically correct, they could bebiased away from the true values.

The statistical methods implementedin FARA are based on very restrictivecriteria in order to try to ensure thehighest probability of a reliable result.When the ambiguities are resolved,you know that SKI-Pro has found a"most probable" solution with an rmsvalue that is significantly lower thanfor any other possible ambiguity set.

If the guidelines for baseline lengths,observation windows, number ofsatellites, GDOP, and observationtimes are followed (combined

perhaps with your own experience),the results of baselines for which theambiguities are resolved should bewithin the system specifications.

Nevertheless, as explained above, itis simply impossible to eliminatecompletely the possibility of theoccasional biased result.

For baselines up to 20 km (systemdefault for "Fix ambiguities up to"),

ambiguity resolution should alwaysbe successful if good results are tobe achieved.

For baselines up to the limitationvalue, SKI-Pro searches for allpossible combinations of ambiguitiesand evaluates the rms of a single-difference observation for each set of

ambiguities. It then compares the twosolutions with the lowest rms values.If there is a significant differencebetween the two rms values, theambiguity set yielding the lowest rmsvalue is considered as the correctone. This decision is based onstatistical methods.

The reader will realize, of course, thata least-squares adjustment can onlyprovide the "most probable" values.


34/42

34

Note that for baselines upto 20 km it should normallybe possible to resolve theambiguities provided that

sufficient observations havebeen taken (see page 15for a guide to baselinelengths and observationtimes). If the ambiguitiesare not resolved check therms values in the logfile(see next page).

For baselines above the limitationvalue (system default = 20 km), SKI-Pro eliminates the ionospheric effectsbut does not attempt to resolve

ambiguities.

Thus the result will always show"Ambiguities not resolved"(Ambiguitystatus = no).

Note that there is usually nobenefit in trying to resolveambiguities for lines over20km.

As already explained, ambiguityresolution should always besuccessful for baselines up to 20 kmif good results are to be obtained.

If insufficient observations weretaken or the satellite constellationwas poor, SKI-Pro will not be able toresolve the ambiguities. If theambiguities are not resolved it ismost unlikely that the systemspecifications will be achieved.

If the ambiguities are not resolved inRapid Static (short observationtimes) it is difficult to give anindication of accuracy. However, as arough guide, one could multiply thesigma values for each estimatedcoordinate by 10 in order to obtain anapproximate estimate of the accuracy

of the baseline computation.


35/42

35

As explained in section "Rmsthreshold" (see page 26), if the rmsfloat exceeds the rms threshold, thebaseline solution with fixedambiguities will be rejected and onlythe float solution will be presented(ambiguities not resolved). Thus ifambiguities are resolved the rms floatand rms fix have to be lower than therms threshold.

The table on page 27 provides anapproximate guide to the rms values(float and fix) that can be expected.

If the rms threshold is lower than therms float or rms fix one can considermanually increasing the rmsthreshold value. However, as

explained on page 27, one shouldexercise a certain amount of cautionwhen doing this for Rapid Staticobservations with up to 10 minutes ofmeasurement time.

The reason is that this could allowunreasonably high rms float and fixvalues and could therefore lead to aweak solution being accepted.

Manually widening the rms thresholdvalue for successful baselinecomputation requires a certainamount of experience andjudgement.

If baselines of greatly differinglengths have to be computed, it isadvisable to make two or morecomputation runs. In this way youcan select and compute batches ofbaselines which fall into the samecategory of processing parametersets.

For baselines up to the limitationvalue, ambiguity resolution using theLambda search and the FARAstatistics is always attempted.

When you look at the logfile, you willfind a summary of the FARAstatistics at the end of each baselineoutput. You should check thefollowing:

Number of satellites: there shouldalways be at least four.

The rms float: this is the rms valuebefore fixing ambiguities.

The rms fix: this is the rms valueafter fixing ambiguities. The rms

fix will usually be slightly higherthan the rms float.


36/42

36

For baselines above the limitationvalue (system default = 20 km), SKI-Pro eliminates the ionospheric effectsbut does not attempt to resolve

ambiguities.

When inspecting the logfile check thefollowing:

The number of satellites observed.

The rms of unit weight

The rms of unit weight should be lessthan about 20 mm for lines of about20 km to 50 km. For lines over 50 kmthe rms of unit weight will usually behigher due to the minor inaccuraciesin the broadcast ephemeris.

If the results are not as good as youwould expect, it can be wellworthwhile comparing the informationin the logfile with that in the field

sheets. Check if the number ofsatellites used in the baselinecomputation is the same as thatnoted in the field sheets. Rememberto check the reference station as wellas the rover. If the number of thesatellites is not the same, the GDOPvalues could be higher than youexpected. Check the actual GDOPfor the satellites used in thecomputation using the SatelliteAvailability component of SKI-Pro.

If a point was observed twice indifferent windows or two referencereceivers were operatingsimultaneously, you should compare

the resulting coordinates.


37/42

37

After inspecting the summary ofresults and the logfile, store theresults that meet your accuracyrequirements.

The coordinates are averaged(weighted mean) if more than onesolution for a point is stored. Forinstance if you store the coordinatesfor point A from one baseline solutionand then you compute and store thecoordinates for point A again from

another baseline solution, the storedcoordinates will be updated to theweighted mean values from the twosolutions. The weighted mean istaken provided the coordinates agreein both height and position to withinthe "Limits for Automatic CoordinateAveraging"set in SKI-Pro (default =0.075m).

It follows that you should exercise acertain amount of care when storingpoints that have been fixed in morethan one baseline computation.

Compare the results before storing.


38/42

38

Baselines above the limitation value(default = 20 km):

L3 solution, ambiguity resolution not attempted.

Results should meet specifications providedsufficient observations are taken.

Long lines need long observation times.

Inspect double fixes, independent baselines etc.

Store results that meet accuracy requirements.

Coordinates averaged if more than one result stored.

Interpreting and Storing the Results

For lines up to 20 km, ambiguity resolution should besuccessful if high-accuracy results are to be obtained.

For long lines over 20 km, the L3 solution withoutambiguity resolution will normally be used.

Baselines up to the limitation value(default = 20 km):

Ambiguity resolution always attempted.

Ambiguities resolved (Ambiguity status = yes):

SKI-Pro has found most probable solution.

Results should normally meet specifications.

Ambiguities not resolved (Ambiguity status = no):

Float solution presented.

Result outside specifications, inspect logfile.

Consider increasing the rms threshold andrecomputing.


39/42

39

6

7

1315

17

22

23

24X

30X

32X

35X

37X

After the observations have beencomputed, you may wish to adjustthe results if multiple observations topoints exist. This provides the bestestimates for the position of thepoints. See SKI-Pro online help"Adjustment"for further details.

The results of the baselinecomputations are coordinates in theWGS 84 system. Using a"Coordinate System"in SKI-Pro,

these coordinates can betransformed into coordinates in anylocal datum or grid system.


40/42

40

X

X

X

X

XX

When measuring with the SR510(System 500) or SR9400 / SR261(System 300) there are severalpoints that should be noted in order

that the measurements aresuccessful and good results can beobtained.

Only observation windows with aminimum of 5 satellites above 15and a good GDOP (< 8) should beused.

The minimum observation time inStatic or Rapid Static should neverbe less than 15 minutes.

As a rule of thumb the baselineobservation time should be 5 minutesper kilometre of the baseline lengthwith a minimum time of 15 minutes.

Recommended (minimum)observation times:

A Rapid Static observation canusually be considered to besuccessful when SKI-Pro can resolvethe ambiguities. Providing an

estimate of the required observationtime is more difficult for singlefrequency receivers than for dualfrequency equipment as considerablyless information is available for thepost processing software. Never theless, the above table should serve asa guide.

By default, SKI-Pro will not attempt toresolve ambiguities if less than 9minutes of (rapid) static, single-frequency data is available. This isdone in order to avoid unreliableresults. Once the ambiguities areresolved correctly the length of thebaseline will normally be accurate to

about 5 - 10 mm plus 2 ppm. Thesedefault settings can be changed inthe Data Processing component ofSKI-Pro, but this is notrecommended.

Baseline-length Observation time1 km 15 min.

2 km 15 min

3 km 15 min

4 km 20 min

5 km 25 min

6 km 30 min7 km 35 min

8 km 40 min

9 km 45 min

10 km 50 min

> 10 km > 60 min


41/42

41

6

7

1315

17

22

23

24X

30X

32X

35X

37X

If the highest possible accuracyshould be achieved it isrecommended to orient the antennasin a common direction.

On long baselines above 10 km theaccuracy which can be achieved withsingle frequency Sensors is inferiorto that which can be achieved withdual frequency Sensors due toionospheric effects which cannot beeliminated with single frequency data.Users who have previously workedwith dual frequency equipmentshould be aware of this fact.


42/42

Printed in Switzerland - Copyright LeicaGeosystems AG, Heerbrugg, Switzerland 2000Original text

712168-2.0.0en

GPS500 General Guide

Documents