PPP solutions with SSR corrections for a single frequency receiver

In my last post, I demonstrated some PPP solutions using real-time SSR corrections from the CLK93 data stream with various dual frequency receivers.  Results were quite good with errors after four hours generally below six centimeters in each axis.

Many of my experiments are done using RTK solutions with short baselines where the differences between dual frequency receivers and single frequency receivers can be fairly small.  With PPP solutions however there are significant differences between single frequency and dual frequency receivers.  This is because the ionospheric errors increase with increasing baseline and the dual frequency measurements are much more effective at coping with these errors.

To demonstrate this, I duplicated the previous experiment with a single frequency receiver, collecting twelve hours of raw observations with a u-blox M8T receiver along with the CLK93 real-time data stream for SSR corrections.   I processed the data with the same configuration settings as last time, except using “L1” instead of “L1+L2” for frequencies and ionospheric corrections set to  “broadcast” instead of “ionospheric-free” (dual-freq) corrections.

As expected, the errors, even after 12 hours were fairly large.  Here is the result.   Even after 12 hours, the vertical error was nearly two meters.

ssr4

Presumably the single frequency PPP/SSR results will improve if RTKLIB is extended to support the RTCM SSR Phase 2 and Phase 3 ionospheric correction messages.

Advertisements

Using SSR corrections with RTKLIB for PPP solutions

If you have been following recent announcements in precision GNSS, you may have been hearing a lot about SSR (State State Representation).  SwiftNav recently announced their Skylark corrections service, and u-blox announced a partnership with Sapcorda to provide correction service for their upcoming F9 receivers.  Both of these services are based on SSR corrections.

So, what is SSR?  Very briefly, it refers to the form of the corrections.  In traditional RTK with physical base stations or virtual reference stations (VRS), the corrections come in the form of observations in which all of the different error source are lumped together as part of the observation.  This is referred to as OSR (Observation Space Representation).  In SSR corrections, the different error source (satellite clocks,  satellite orbits, satellite signal biases, ionospheric delay, and tropospheric delay) are modeled and distributed separately.  There are many advantages to this form but what seems to be driving industry towards it now is that it allows the current VRS model where each user requires a unique data stream with observations tailored for their location to be replaced with a single universal stream that can be used by all observers.  This is a requirement if the technology is going to be adopted for the mass-market automotive industry for self-driving cars, since it is not practical to provide every car with it’s own data stream.

For more detailed information on SSR, Geo++ has a one page summary here and IGS has an 18 minute video presentation on the topic here.  Both are excellent.

Below is an image I borrowed from the IGS presentation which shows the flexibility of the SSR format.  It is intended to show how the same SSR data stream can be used globally for PPP quality corrections and also regionally for RTK quality corrections but it is also a good visual for understanding the message details I describe below.

ssr1

The RTCM standards committee is still in the process of finalizing the messages used to send the different correction components.  They have split the work into three phases.  Phase 1 includes the satellite clock, orbit, and code biases.  Phase 2 includes satellite phase biases and vertical ionosphere corrections, and phase 3 includes ionospheric slant corrections and tropospheric corrections.

There are several real-time SSR streams accessible for free today.  Unlike the paid services, they do not contain enough detailed regional atmospheric corrections to use as a replacement for a VRS base but they can easily be used for static PPP solutions.

I used the CLK93 stream from CNES for an experiment to test how well RTKLIB handled the SSR corrections.  It includes the Phase 1 and Phase 2 RTCM messages but not the Phase 3 messages.  Here is the format of the the messages in the CLK93 data stream:

clk93

You can register for free access to the CLK93 (and other) streams from any of these locations:

Unfortunately, RTKLIB currently only supports the Phase 1 RTCM messages and even this is not complete in the release version.  I have gone through the code and made a few changes to make the Phase 1 SSR functional and have checked those changes into the demo5 Github repository.  I also added some code to handle the mixed L2 and L2C observations from the ComNav and Tersus receivers.  After a little more testing I plan to release this code into the demo5 executables, hopefully in the next week or two.

With only phase 1 measurements, the RTKLIB PPP solutions will work much better with dual frequency receivers than with single frequency receivers.  This is because single frequency receivers require ionospheric corrections for longer baselines.  For this reason, I did not bother with collecting any single frequency data.  Instead, I collected both L1/L2C data with a Swiftnav Piksi Multi receiver and L1/L2/L2C data with a ComNav K708 receiver and a Tersus BX306 receiver.

RTKLIB is usually used to calculate PPP solutions without SSR corrections but this requires downloading multiple correction files for orbital errors, clock errors, and code bias errors and it is usually done with post-processing rather than real-time, after the corrections are available.  With SSR, the process is simpler because the solution can be done real-time and there is no need to download any additional files.  It does, however, require access to the internet to receive the real-time SSR data stream from an NTRIP caster.  The solution can be calculated real-time or the SSR corrections and receiver observation streams can be recorded and the solution post-processed.

To enable the use of SSR corrections in RTKLIB, you need to set the “Satellite Ephemeris/Clock (pos1-sateph) input parameter to either “Broadcast+SSR APC” or “Broadcast+SSR CoM”.  Note that CoM stands for Center of Mass and APC for Antenna Phase Center.  They refer to the reference point for the corrections.  The CLK93 corrections are based on antenna phase centers.

To generate my PPP solution I set the solution mode to “PPP-Static”,  ephemeris/clock (pos1-sateph) to “brdc+ssrapc”, ionosphere correction (pos1-ionopt) to “dual-freq”, and troposphere correction (pos1-tropopt) to “est-ztd”.  I also enabled most of the other PPP options including  earth tides,  satellite PCVs, receiver PCVs, phase windup, and eclipse rejection.

RTKLIB PPP solutions don’t support ambiguity resolution so the ambiguity resolution settings are ignored.  I specified the satellite antenna file as “ngs14.atx” which is the standard antenna correction file and is available as part of the demo5 executable package.  I also needed to include the CLK93 data stream as one of the inputs in addition to the receiver observations (and navigation file if post-processing).

I collected a couple hundred hours of observations with the SwiftNav receiver connected to a ComNav AT-330 antenna on my roof with moderately good sky visibility.  I then ran many four hour static solutions over randomly selected data windows.  I also collected a small amount of raw data from a ComNav K708 receiver and a Tersus BX306 receiver.

Below is a typical 12 hour static solution for a SwiftNav and a ComNav receiver.  The SwiftNav solution is in green and the ComNav solution is in purple.  Zero in these plots represents an online PPP solution from CSRS from data collected over a month earlier.  In this particular example, the SwiftNav solution is slightly better but this was not always the case.

 

ssr2

Here is a shorter data set from a Tersus BX306 receiver.  With the relatively small amount of Tersus and ComNav data I collected, I did not notice any differences in convergence rates or final accuracy between any of the three receivers.

ssr3

The solutions generally all converged to less than 6 cm of error in each axis after 4 hours with one or two minor exceptions that seemed to involve small anomalies at the day boundary.  Since the RTKLIB PPP solutions don’t include ambiguity resolution they do take longer to converge but the eventual accuracy should be similar.

I’ve uploaded some of the raw observation data for the different receivers and the CLK93 data stream as well as the config file that I used for the solution here.

This seems like a good start and I hope that RTKLIB will support phase 2 and phase 3 corrections in the future.

Swiftnav experiment: Improvements to the SNR

In my previous couple of posts, I evaluated the performance of a pair of dual freqeuncy SwiftNav Piksi multi receivers in a moving rover with local base scenario.  I used a pair of single frequency u-blox M8T receivers fed with the same antenna signals as a baseline reference.

It was pointed out to me that the signal to noise ratio (SNR) measurements of the rovers were noticeably lower than the bases, especially the L2 measurements and that this might be affecting the validity of the comparison.  This seemed to be a valid concern so I spent some time digging into this discrepancy and did indeed find some issues.  I will describe the issues as well as the process of tracking them down since I think it could be a useful exercise for any RTK/PPK user to potentially improve their signal quality.

Previously , in another post, I described a somewhat similar exercise tracking down some signal quality issues caused by EMI from the motor controllers on a drone.  In that case, though, the degradation was more severe and I was able to track it down by monitoring cycle slips.  In this case, the degradation is more subtle and does not directly show up in the cycle slips.

Every raw observation from the receiver generally includes a signal strength measurement as well as pseudorange and carrier phase measurements.  The SwiftNav and u-blox receivers both actually report carrier to noise density ratio (C/NO), rather than signal to noise ratio (SNR) but both are measures of signal strength.  They are labelled as SNR in the RTKLIB output, so to avoid confusion I will refer to them as SNR as well.  I will only be using them to compare relative values so the difference isn’t important for this exercise, but for anyone interested, there is a good explanation of the difference between them here.  Both are logarithmic values measured in dB or dB-Hz so 6 dB represents a factor of two in signal strength.

Since the base and rover have very similar configurations we would expect similar SNR numbers between the two, at least when the rover antenna is not obstructed by trees or other objects.  I selected an interval of a few minutes when the rover was on the open highway and plotted SNR by receiver and frequency for base and rover.  Here are the results, base on the left and rover on the right.  The Swift L1 is on the top, L2 in the middle, and the u-blox L1 on the bottom.  To avoid too much clutter on the plots, I show only the GLONASS SNR values, but the other constellations look similar.

snr1

Notice that the L1 SNR for both rovers is at least 6 dB (a factor of 2) lower than the base, and the Swift L2 SNR is more like 10 dB lower.  These are significant enough losses in the rover to possibly affect the quality of the measurement.

The next step was to try and isolate where the losses were coming from.  I set up the receiver configurations as before and used the “Obs Data” selection in the “RTK Monitor” window in RTKNAVI to monitor the SNR values in real time for both base and rover as well as the C/NO tracking window in the Swift console app.  I then started changing the configuration to see if the SNR values changed.  The base and rover antennas were similar but not identical so I first swapped out the rover antenna but this did not make a difference.  I then moved the rover antenna off of the car roof and onto a nearby tripod to see if the large ground plane (car roof) was affecting the antenna but this also did not make a difference.  I then removed the antenna splitter, but again no change.

Next, I started modifying the cable configuration between the receivers and my laptop.  To conveniently be able to both collect solution data and be able to collect and run a real-time solution on the raw Swift observations, I have been connecting both a USB serial cable and an ethernet cable between the Swift board and my laptop.  My laptop is an ultra-slim model and uses an etherent->USB adapter cable to avoid the need for a high profile ethernet connector.  So, with two receivers and my wireless mouse, I end up having more USB cables than USB ports on my computer and had to plug some into a USB hub that was then plugged into my laptop.

The first change in SNR occured when I unplugged the ethernet cable from the laptop and plugged it into the USB hub.  This didn’t affect the L1 measurements much but caused the Swift L2 SNR to drop another 10 dB!  Wrong direction, but at least I had a clue here.

By moving all of the data streams between Swift receiver and laptop (base data to Swift, raw data to laptop, internal solution to laptop) over to the ethernet connection I was able to eliminate one USB serial port cable.  This was enough to eliminate the USB hub entirely and plug both the USB serial cable from the u-blox receiver and the ethernet->USB cable from the Swift receiver directly into the laptop.  I also plugged the two cables into opposite sides of the laptop and wrapped the ethernet->USB adapter with aluminum foil which may have improved things slightly more.

Here is the same plot as above after the changes to the cabling from a drive around the neighborhood.

snr2

I wasn’t able to eliminate the differences entirely, but the results are much closer now.  The biggest difference now between the base configuration and the rover configuration is that I am using a USB serial cable for the base, and a ethernet->USB adapter cable for the rover so I suspect that cable is still generating some interference and that is causing the remaining signal loss in the rover.  Unfortunately I can not run all three streams I need for this experiment over the serial cable, so I am not able to get rid of the ethernet cable.

I did two driving tests with the new configuration, similar to the ones I described in the previous posts.   One was through the city of Boulder and again included going underneath underpasses and a parking garage.  The second run was through the older and more challenging residential neighborhood.  Both runs looked pretty good, a little better than the previous runs but it is not really fair to compare run to run since the satellite geometry and atmospheric conditions will be different between runs.  The relative solutions between Swift and u-blox didn’t change much though, which is probably expected since the cable changes improved both rovers by fairly similar amounts.

Here’s a quick summary of the fix rates for the two runs.  The fix rates for the residential neighborhood look a little low relative to last time but in this run I only included the most difficult neighborhood so it was a more challenging run than last time.

Fix rates

City/highway Residential
Swift internal RTK 93.60% 67.50%
Swift RTKLIB PPK 93.70% 87.90%
U-blox RTKLIB RTK 95.70% 92.80%
U-blox RTKLIB PPK 96.10% 91.10%

Here are the city/highway runs,  real-time on the top, post-process on the bottom with Swift on the left and u-blox on the right.  For the most part all solutions had near 100% fix except when recovering from going underneath the overpasses and parking garage.

snr4

Here are the same sequence of solutions for the older residential neighborhood.  This was more challenging than the city driving because of the overhanging trees and caused some amount of loss of fix in all solutions.

snr5

Here’s the same images of the recovery after driving under an underpass and underneath a parking garage that I showed in the previous post.  Again, the relative differences between Swift and u-blox didn’t change much, although the Swift may have improved a little.

snr1

Overall, the improvements from better SNR were incremental rather than dramatic, but still important for maximizing the robustness of the solutions.  This exercise of comparing base SNR to rover SNR and tracking down any discrepancies could be a useful exercise for anyone trying to improve their RTK or PPK results.

Initial look at the ComNav K708 receiver

ComNav was kind enough to recently lend me two of their K708 receivers for evaluation.   I also have a Tersus BX306 receiver that was given to me earlier by Tersus for evaluation.  Both of these are relatively low-cost dual frequency receivers that offer full GPS L2 support., unlike the SwiftNav receiver I evaluated in my previous posts which is GPS L2C only.  I have described the Tersus BX306 before in a previous post but last time I was not able to evaluate it with a local base since I did not have a second dual frequency receiver that supported L2.  Tersus has also just recently released their new V1_19 firmware so I included that in this evaluation.   As usual I’ve also included  a pair of u-blox M8T receivers to use as a baseline.

Here’s a photo that shows the three receivers each with their associated serial port and power cabling.  The u-blox M8T is on the left, Tersus BX306 in the center, and ComNav K708 on the right.  The ComNav receiver is actually only the smaller daughter board in the center of the larger board, everything else is part of the very sturdy but rather clunky dev kit.

rcvrs3

The Tersus BX306 is priced at $1699 but lower priced versions are available. For example, the BX305 supports GPS L1/L2 but Glonass G1 only, and the BX316R is GPS L1/L2 and Glonass G1/G2 but provides only raw observations for post-processing.  Both of these options are priced at $999.

The ComNav K708 is similar to the better known K501G but newer and more capable.  ComNav doesn’t list their prices on their website but they have told me that both the K501G and the K708 configured to be equivalent to the K501G (GPS L1/L2 and GLO G1/G2) are available for less than $1000.

Both the Tersus and the ComNav receivers come with GUI console apps which are good for initially getting familiar with the receivers.  However each had their unique quirks and I found myself fairly quickly abandoning them for the more familiar quirks of the RTKLIB apps.  Managing three simultaneous real-time solutions involving five separate receivers while also logging raw observations for all five was actually quite challenging and I made a couple of unsuccessful runs before I got everything working at the same time.

I found that the key to turning this into a manageable and automated process was replacing each of the different manufacturer’s GUIs with an RTKLIB stream server (STRSVR) and a plotter (RTKPLOT) each with it’s own dedicated .ini file.  Eliminating the GUIs also gave me a better understanding of exactly what the receivers were doing and what the GUIs were doing.

STRSVR provides a standardized, always visible red/yellow/green indicator for each stream along with a continuously updated bps number that indicates not only that the connection is alive, but that data is flowing.  This allowed me to tell at a glance that all streams were flowing and that all the log files were being updated.  Using the “-t” option in the command line to specify a title for each window also helped keep things straight.

Both receivers are configured by sending Novatel-like ASCII commands over the serial port and these can be added to the STRSVR Serial “Cmd” window and saved to a “.cmd” file, similar to configuring the u-blox receiver.  Notice in this example, I also sent a reset to the receiver every three minutes which was a convenient way to automate the testing of acquisition times.

strsvr1

I connected both dual frequency rover receivers to my laptop, using two COM ports for each one and using a USB hub to get enough ports.  I set up both receivers to output NMEA solution messages and raw RTCM observation messages on COM1 at 5 Hz and accept RTCM base station data on COM2.  Both receivers have decent reference manuals to describe their command set but I also found this Hackers Guide to the K501G from Deep South Robotics quite useful for getting started.

For reference, here are the commands I used to configure the Tersus rover:

fix none
unlogall
log com1 gpgga ontime 1 nohold
rtkcommand reset
log com1 gpgga ontime 0.2

log com1 rtcm1004 ontime 0.2
log com1 rtcm1012 ontime 0.2
log com1 rtcm1019 ontime 1
log com1 rtcm1020 ontime 1
interfacemode com2 auto auto on

saveconfig

and here are the commands I used for the ComNav rover:

interfacemode compass compass on
unlogall com1
fix none
refautosetup off
set cpufreq 624
rtkobsmode 0
rtkquality normal
set pvtfreq 5
set rtkfreq 5
log com1 gpgga ontime 0.2 0 nohold
log com1 gprmc ontime 2 0 nohold
log com1 rtcm1005b ontime 10
log com1 rtcm1004b ontime 0.2
log com1 rtcm1012b ontime 0.2
log com1 rtcm1019b ontime 2
log com1 rtcm1020b ontime 2
interfacemode com2 auto auto on

saveconfig

My intent was to setup the receivers in default RTK mode with a 5 Hz output for NMEA solution messages and RTCM raw observation and navigation messages.  The one exception to default was that I found the “rtkquality” setting on the ComNav receiver defaulted to “quick” which was giving me false fixes, so I changed this to “normal” and that seemed to fix the problem.

By setting things up this way, I only need to click on the correct combination of icons (each tied to it’s own .ini file) from my RTKLIB menu to bring up the correct windows and a few more clicks to start the streams in a simple and repeatable way.

dualFreq3

I’m jumping ahead a little bit, but here is a screen capture of the rover-connected laptop streaming two NTRIP sets of base station data to the rovers while simultaneously logging and plotting the computed solutions for all three rovers along with raw observations for all five receivers,  and also computing an RTK solution for the M8T receivers with RTKNAVI.

Capture3

I should mention that there was one very annoying bug that was introduced to STRSVR in one of the recent RTKLIB releases that gives an error if a data file already exists instead of an overwrite dialog but I did fix this and add it to a new demo5 b29b code release available at the download page on rtkexplorer.com.  The new release also includes a fix for another bug that prevented the “-i” command line option to specify a config file for RTKPLOT from working properly.

I then setup the second ComNav receiver as a base station for both dual frequency rovers and used a single COM port to stream RTCM messages from the receiver to a PC.  I used an STRSVR window on the PC to stream the messages to a NTRIP caster using the free RTK2GO NTRIP caster service as I have previously described.  I used ComNav AT330 antennas for both the base and rovers with the rover antenna shared by all three rover receivers.   I did not have enough connector hardware to share the base antenna so used a separate u-blox antenna for the M8T base receiver.

The next step was to collect some data.  I started with a relatively simple challenge, a static rover with a reasonably open sky view and a short baseline.  The ComNav and Tersus solutions both assume the rover may be moving so I set up the M8T solution as kinematic as well.

Let’s first look first at the ComNav solution compared to the M8T solution.  Both solutions were computed real-time.  RTKPLOT will plot NMEA data but it did not seem to like the mix of NMEA and RTCM data in the same file.  To deal with this, I wrote a simple matlab script to strip the NMEA messages from the log file and put them in a separate file.  Below I have plotted only the Up/Down axis for both receivers just to avoid too much data,  the M8T is on top, and the ComNav below.  Each of the larger breaks in the fix was caused by me disconnecting then reconnecting the antenna to force a re-acquire.

comnav1

The M8T configuration was identical in the left and right plots, but the ComNav “rtkquality” parameter was set to “quick” in the left plot, and “normal” in the right plot.  It’s not as obvious here as it is in the other axes but the third ComNav fix in the left plot is a false fix and had over 0.2 meters of error in the N/S axis.  Changing the “rtkquality” parameter to “normal” seemed to help and I did not notice any more false fixes after making that change.

The ComNav receiver typically achieved a fix very quickly regardless of the “rtkquality” setting, usually in less than 30 sec although in one case it took a minute and a half.  This was noticeably faster than the M8T receiver, which took from 1 to 3 minutes each time in this example to achieve a first fix.

The scales are the same in the two sets of plots, so as you can see, the ComNav fixes are a fair bit noisier than the M8T fixes.  I don’t know why this is but it is something that I hope to investigate more.

Unfortunately I got a mix of good and not so good results from the Tersus receiver.   I did not see this behavior in my previous evaluation so I’m fairly certain this is not a problem with the hardware.  I suspect it has something to do either with my setup or with the new firmware.  I am going to hold off on sharing any of the Tersus data until I understand better what is going on.

Next, for a more challenging test, I moved the rover antenna to a spot with fairly poor sky views located between several large trees.  The sky view directly above the antenna was clear but a large percent of the overall view was blocked.   Again, I just plotted the Up/Down axis with the M8T position solution on the top and the ComNav solution on the bottom.

comnav2

I disconnected and reconnected the antenna three times in this experiment.  The M8T did not get a fix in the first try before I gave up after 12 minutes, but it did after 13 and 11 minutes in the second two tries after briefly getting a false fix in the second try.  Definitely marginal conditions for the M8T.  The ComNav receiver did significantly better with two fixes in less than 3 minutes and one in 9 minutes.  The errors were relatively large in the first fix but based on the other two axes it was not a false fix.  You can also see that the ComNav third fix was noticeably noisier than any of the other fixes on either receiver, again for unknown reasons.

For the third part of the experiment I moved the receivers into my car and attached the antenna to the roof and collected data for three spins around the neighborhood.  The results are plotted below.  In each case the M8T real-time solution is on the left, and the ComNav is on the right.  In the data in the first row, I shared a single antenna for all three receivers.  For the data in the second and third row I used separate antennas.  I did not change any of the config settings for any of the receivers between these runs and the above runs except that the rtkquality setting was still set to “quick” for the ComNav receiver for the second and third rows.

 

 

 

comnav5

 

comnav6

comnav7

I have not had a chance to look at this data closely but at first glance, from a fix percentage perspective only, I don’t see significant differences between either of the receivers.  The obvious advantages the ComNav receiver demonstrated in faster fixes in the static tests did not seem to carry over to the moving rover case.  I do plan to look at the raw data more carefully to see if I can understand better why this is.  For whatever reason, the Tersus receiver seemed to perform better with a moving rover than it did with a static rover, and was very similar in fix percentage to the other two receivers in this part of the experiment.

Next I planned to post-process the raw data through RTKLIB to better understand what is going on but as usual, nothing is as simple as you hope for, and I ran into another issue.

Both the Tersus and the ComNav receiver report a mix of 2W and 2X  measurements for the raw GPS L2 measurements.  If the satellite supports the newer L2C code it locks to that and reports a 2X code, if not, it locks to the older L2  and reports a 2W code.   You can see this in this example observation epoch from the Rinex conversion of the ComNav receiver RTCM output.  The left three columns are the L1 measurements, the middle three columns are the L2 (2W) measurements and the right three columns are the L2C (2X) measurements.  You can see that all the GLONASS satellites report L2 measurements only but that the GPS satellites are a mix of L2 and L2C measurements.

comnav4

This is new for the Tersus receiver, it did not do this when I evaluated it with the older firmware.  For the ComNav receiver, this is the default behavior but it is possible to change this through a command to specify L2 only, no L2C.  As far as I can tell, the Tersus only supports the mixed L2/L2C mode.  All the data I collected for this experiment was in the mixed L2/L2C mode.

Unfortunately RTKLIB does not like this format and throws away all of the L2C measurements.  It is possible to fool RTKLIB into using all the measurements by changing the 2X’s in the “Obs Types” list in the file header to 2W’s but I haven’t looked yet at to what extent mixing the code types affects the solution or how to avoid throwing away the L2C data without editing the header.

I will leave a more detailed analysis of the data to a future post.  My initial impression from these results though, is that although there are some obvious advantages with the ComNav receivers, replacing a pair of low cost single frequency receivers with a pair of low cost dual frequency receivers does not magically make the challenges of precision GNSS go away and that it will still require close attention to the details and recognition of their limits to get good results with either set of receivers.

 

 

PPK vs RTK: A look at RTKLIB for post-processing solutions

The “RTK” in RTKLIB is an abbreviation for “Real-time Kinematics”, but RTKLIB is probably used at least as often for “PPK” or “Post-Processed Kinematics” as it is for real-time work.  In applications like precision agriculture, where the solution is part of a real-time feedback loop, RTK is obviously a requirement, but in many other applications there is no need for a real-time solution.  For example, drones are often used for collecting photographic or other sensor data but only need precision positions after the fact to process the data.  PPK is simpler than RTK because there is no need for a real-time data link between GPS receivers and so is often preferable if there is a choice.  The downside of course is that if there is something wrong with the collected data, you may not find out until it’s too late.

For the most part, RTKLIB solutions are identical regardless if they are run on real-time data (RTK) or run on previously collected data (PPK).  The most significant exception to this rule is what RTKLIB calls the “Filter Type”.  This is selected in the configuration and can be set to forward, backward, or combined.  Forward is the default and this is the only mode that can be used in real-time solutions.  In forward mode, the observation data is processed through the kalman filter in the forward direction, starting with the beginning of the data and continuing through to the end.  Backward mode is the opposite,  data is run through the filter starting with the end of the data and continuing to the beginning.  In Combined mode, the filter is run both ways and the two results are combined into a single solution.   This mode is set using the “Filter Type” box in the Options menu if using one of the GUI apps, or with the “pos1-solytpe” input parameter in the configuration file if using a CUI app.

There are two advantages to a combined solution over a forward solution.  First of all, it gives two chances to find a fix for each data point.  Let’s say there is an anomaly in the middle of the data set that causes the solution to switch from fix to float and not come back to fix for some period of time.   It may cause both the forward and backward solutions to lose fix but they will lose fix on opposite sides of the anomaly.  By combining the two solutions we are likely to get a fix for everywhere except right at the anomaly.  Another case where it often helps is in recovering the beginning of a data set.  Let’s say the first fix didn’t occur until five minutes into the data set.  With a forward solution, you would need to guarantee that nothing important happened during that five minutes, but with a combined solution, the backward pass will normally provide a fix all the way to the very beginning of the data set so there is no lost data.

The second advantage of the combined solution is that it provides an extra level of validation of the results.  To understand how this happens, it’s important to understand how RTKLIB combines the forward and reverse solutions.  For each solution position point there are three possibilities; both passes are float, one is float and one is fix, or both are fixed.  If both passes generate a float position, then the combined result will be a float with a value equal to the average of the two positions.  If one is float, and the other is fix, the float is thrown away and the fix is used.  In the case where both are fixed, then RTKLIB will attempt to validate the result by comparing the two values.  If they differ by less than four sigma, then the result will be a fix, otherwise it will be downgraded to a float.  Either way, the value will be the average of the two positions.  This degrading the solution type when the answers from opposite directions differ provides an increased confidence in the solution, at least for points for which we got two fixed values.

I will show a couple examples of the differences between forward and combined modes.  The first example is a more typical case and demonstrates how combined mode will normally give you a higher fix percentage while at the same time increasing confidence in the solution.

The plots below were taken from an M8N receiver on a sailboat using a nearby CORS station as base.  With ambiguity resolution mode set to fix-and-hold, I was able to get a solution with nearly 100% fix except for the initial convergence, but I would prefer to use continuous ambiguity resolution because of the higher confidence of the solution.  In the position plots below, the top was run in forward mode, the middle in backwards mode, and the bottom in combined mode, all in continuous ambiguity resolution mode.

combined1

As you can see the forwards and backwards mode solutions are not bad but both have gaps of float in the middle as well as floats during the initial acquisition.  The combined solution though has almost 100% fix rate and in addition includes the additional confidence knowing that every point found the same solution when running the data in opposite directions.

This second example comes from a data set posted on the Emlid Reach forum with a question on why the combined solution was worse than the forward solution.  In the plots below, the top solution is forward, the middle is backward, and the bottom is combined.

combined2

This data was GPS and SBAS only, so had a fairly low number of satellites, also included a mix of poor observations and the solution was run with full tracking gain (i.e fix-and-hold with the default gain).  Both forward and backward runs found fixed (green) solutions and tracked them all the way through the data set.  However, at least one of them was most likely a false fix, causing the fix to be downgraded to float (yellow) for most of the combined solution as can be seen be seen in the bottom plot.

To confirm this, the plot below shows the difference between the forward and backward solutions.  As you can see, the two differ by a fairly substantial amount and it is not possible from this data to know which one is correct.

combined3

In this case, turning off fix-and-hold and running ambiguity resolution in continuous mode sheds some light on what may be going on.  The plots below are again forward, backward, and combined.  This time the forward solution loses fix early on and never recovers it, whereas the backwards solution maintains a fix through the whole data set and is probably correct since without fix-and-hold enabled, it is very unlikely to stay locked that long to an incorrect solution.  The backward solution is also consistent with the beginning of the forward solution, since the combined solution remains fixed in the early part of the data set where both forward and backward solutions are fixed.

combined4

Again, this can be confirmed by looking at the difference between the forward and backward solutions.  In this case they agree everywhere that both are fixed.

combined5

As this example demonstrates, if post-processing is an option, it often makes sense to run in combined mode with continuous ambiguity resolution instead of forward mode with fix-and-hold enabled.  The additional pass will increase the chances of getting a fixed solution without the risk of locking onto a false fix that fix-and-hold can cause.  Even if you find you can not disable fix-and-hold completely, it may allow you to reduce the tracking gain (pos2-varholdamb)

So one last question is why are there still some float values in the middle of the combined solution? We would expect that since the backwards solution is fixed and the forward solution is float, that the combined solution should just become the backwards solution and all but the very end should be fixed.

The answer to this question turns out to be the way the reverse pass of the kalman filter is initialized.  I have chosen in the demo5 code to not reset the filter between forward and reverse passes if continuous ambiguity resolution is selected.  If fix-and-hold is selected then the demo5 code does re-initialize the kalman filter between passes.  This is different from the release code which always resets the filter between passes.

In this case, the results would have been slightly better if the filter were re-initialized but most of the time I find that allowing the filter to stay converged avoids a large gap in the backwards solution during the active part of the data set where the filter is reconverging. With fix-and-hold enabled I have found the chance of staying locked to an incorrect fix is too high and so it is better to reset the filter.  This is a recent change and hasn’t yet made it into the released version of demo5 but I should get it out soon.  The current version of the demo5 code (b28a) does not reset the filter for either case.

Modifying the if statement in the existing code in postpos.c to match the line below will give you the newest behavior.  Removing the if statement altogether will cause the filter to always be reset and will match the release code.

combined6

The other factor to consider when deciding whether to run the filter type in forward or combined mode is that combined mode will take nearly twice as long to run since it is processing each data point twice.  Most of the time this shouldn’t be an issue since it is not being run in real-time.

So to summarize, my recommendation would be to use combined mode if you do not need a real-time solution as the only real cost is a small amount of additional computation time and it will give you both higher fix percentages and more confidence in those fixes.

Real-time solutions with RTKLIB and NTRIP using a cell phone as data link

As I mentioned in an earlier post, I’ve recently acquired access to some low cost dual frequency receivers, specifically a Tersus Precis BX306 and a pair of Swift Piksi Multis.  I have been playing with them over the past few weeks and plan to share my experiences with them over a series of posts.

Both receivers provide internal RTK solutions as well as raw measurements that can be processed with RTKLIB.  I’m interested in how the RTKLIB solutions compare to the internal solutions as well as how both of these compare to solutions derived from single frequency data collected simultaneously with the dual frequency data.

The first issue I ran into with this experiment, however, is that both receivers will only provide an RTK solution for real-time data, neither have the capability to post-process previously collected data.  This meant that I needed a way to provide a real-time stream of dual frequency base station data to the receivers.  I wanted to be able to  do this while driving a car around the local area so I needed more range than a low cost set of radios would give.

Fortunately, I have fairly good cell phone coverage in this area so I was able to rely on my cell phone for the data link.  In this post I will explain how I did that, both for an external CORS reference station and for my own base station.  In both cases I used  NTRIP server/caster/clients to do this.  NTRIP is a protocol for streaming of DGPS or RTK correction data via the internet using TCP/IP.  The NTRIP server sends out the data to an NTRIP caster and the NTRIP client receives it. For more details, there is a good description here.

Using this setup I was able to run real-time solutions with RTKLIB as well as with the intenal RTK engines in the Swift and Tersus receivers.  Here’s a diagram from the RTKLIB manual showing the setup I used for running a real-time RTKLIB solution using RTKNAVI.  When I ran a Swift or Tersus solution, the configuration was similar, but the NTRIP caster streamed the base station data to STRSVR instead of RTKNAVI, and STRSVR then streamed it to the receiver where it was combined with the raw receiver observations to create an internal RTK solution.  Also missing in this diagram is the cell phone which should be in between the internet and the rover PC.

ntrip.rtklib

The amount of free base station reference data that is available online on a real-time basis is a fair bit more limited that what is available after the fact for post-processing.  Fortunately I was able to find a CORS reference station about 17 km away that is available real-time through the UNAVCO NTRIP caster.  The service is free if the data is used for educational purposes and appropriately attributed.   Most of their stations are on the west coast of the U.S. but they do have some scattered across the rest of the country as you can see in this map from their site.  There are other networks available in other parts of the world that can be found by searching online.

unavco_map

To access the UNAVCO data I had to request access through email but the process was very simple and within a couple hours of my request I was all setup with an account and password.

Once I had my account set up, I used RTKLIB on my laptop computer to collect the data from the internet and stream it to the rover receiver over a serial port.  If I were doing this experiment within range of a wireless router then I could leave the computer connected to the wireless.  In this case though, I wanted to roam outside the range of my home wireless.  To do this, I enabled a hot spot on my cell phone and logged into that with my computer.

I was able to access the raw observation data stream from the UNAVCO NTRIP caster directly using the NTRIP client option in RTKLIB.  If I had wanted to generate a real-time RTKLIB solution, I would have configured the input streams of RTKNAVI but in this case I want to stream the raw data directly to the receiver so it can use the observation data for it’s internal solution.  I did this using the STRSVR app in RTKLIB.  I specifed the “NTRIP Client” option as input type and then entered the information from my UNAVCO account into the “Ntrip Client Options” as shown below.

ntrip_client

In this case I wanted the data from station P041 in RTCM3 format so I had to specify the Mountpoint as “P041_RTCM3”.  For other networks, the mountpoint details may be a little different.  Most NTRIP casters use Port 2101, and that was the case for this one.  For the STRSVR output type, I specified “Serial” and then configured the serial port options for whichever rover receiver I was using.  Before doing the configuration, I had connected the receiver to the laptop using a USB cable.

I then had to configure the receiver to tell it to get its base station data from the COM port and specify that it is in RTCM3 format.  The details for doing this on the two receivers are a little different but fairly straightforward in both cases.  You may also need to specify the exact base station location manually or the receiver may be able to get it from the data stream depending on the receiver and NTRIP stream details.

And that’s it.  With this configuration, either receiver was able to fairly quickly lock to a fixed RTK solution and continue to receive base data as long as I stayed in range of cell reception.  Any lag in the base station observations appeared to be less than a second.

That worked great for using an existing external reference as base station.  However, I also wanted to run another real-time experiment where I used one Swift receiver as base and the other as rover.   To do this, I needed to set up an NTRIP server to stream the data to  a caster on the internet as well as an NTRIP client to receive it.

I started by connecting the second Swift receiver to an old laptop with a USB cable and then downloading RTKLIB, the Swift console app,  and the right USB drivers.  The base station antenna is on top of my roof and the laptop is in the house so I was able to connect the laptop to the internet using my home wireless.

For the NTRIP caster, I found it convenient to use RTK2GO which is a community caster available for anyone to use at no cost.  To send the data to the caster, I used the “NTRIP Server” as the STRSVR output type and configured it as shown below.

strsvr_server

Again, the port is 2101.  You can choose any name for the mountpoint.  If that name is already in use, then rtk2go will assign a suffix to it, so it is best to choose a name that is unlikely to already be in use.  The password at the current time is BETATEST but that may change from time to time so it’s worth verifying it is still correct.

For the STRSVR input, I selected “Serial” and specified the correct COM port for the base station receiver.  In this case the raw observations are in Swift binary format which RTKLIB does not support so it sends them unaltered.  If they were in a format that RTKLIB did support, then they could be converted to RTCM3 to reduce bandwidth and make them more easily usable by someone else not using a Swift receiver as rover.  You can specify the conversion to RTCM3 using the “Conv” menu on the STRSVR output.

Start STRSVR and your base station observations are now accessible to anyone in the world through RTK2GO.com!

On the rover side, the NTRIP client is set up as I previously described using STRSVR except you want to use the same caster/mountpint/password as you just did on the base station.  In this case the user-id is left blank.  Again, set the STRSVR output to “Serial” to send it to the receiver.   Then set up the receiver to get it’s base station data from the serial port and, in this case, specify that it is in the Swift Binary Protocol (sbp).  Start the receiver and it should fairly quickly get a fix.  If you are seeing baseline data but not a solution, then most likely you have not specified the base station location to the rover.

I was now able to drive around almost anywhere and get continuous real-time RTK solutions using either my own base station or the CORS reference station as base.  In the next post I will discuss some of the data I collected and analyzed.

 

 

 

 

A fix for the RTCM time tag issue

In my last post I described a problem with a loss of some of the raw measurement information caused by the lack of resolution in the time tags in the RTCM format.  Since the RTCM format is typically used to reduce bandwidth requirements in real-time applications, it is causing real-time solutions to fail when post-processing the same raw data without the translation to RTCM gives good results.  In this post I will describe a fix for this problem.

First of all I want to thank Felipe Nievinski, Igor Vereninov from Emlid, and Anthony Woolridge for their comments to the last post that pointed me to the solution.  They make this a collaborative effort between the U.S., Brazil, Russia, and the U.K!  It still amazes me how enabling the internet can be!

I’ll start by showing again this example of a RINEX output from an M8T receiver with the official raw measurement output (RXM_RAWX) and the debug raw measurement output (TRK_MEAS) enabled simultaneously.  I think  this provides a good insight to what is going on.  The TRK_MEAS  message is the top 5 lines and the RXM_RAWX message is the bottom 5 lines for a single epoch.  The first line in each message is the time stamp and the following lines are the measurements for each satellite.  In the satellite measurements, the second column contains the pseduorange value.

trkmeas1

The time stamp specifies the receiver time of the received signals and the sixth column is the number of seconds.  For the TRK_MEAS message these values are always aligned to round numbers based on alignment to the sample rate.  For example in this case the measurement rate was 5 Hz and all the time stamps occur on multiples of 0.2.  This is because they are based on the raw receiver clock without any corrections.

The time stamps from the RXM_RAWX messages however often differ from the round numbers by small arbitrary amounts.  This is because the receiver has estimated the error in its own clock and adjusted the measurements to remove this error.  In this case the estimate of clock error is 0.001 seconds and so the time stamp is adjusted by this value (18.8000000 to 18.7990000).

To keep the time stamps consistent with the other parts of the measurement, the clock error also needs to be removed from the psuedorange and carrier phase values since they are based on the difference in time between satellite transmission and receiver reception and will include any errors in the receiver clock.  We see from the above observations that the pseudorange measurement for satellite G24 has been adjusted from 22675327.198 to 22375547.970, a difference of 299779.228 meters.   The speed of light is 299792458 meters per second so the clock error of 0.001 seconds is equivalent to 299792.458 meters,  a value very close to the amount that the pseudorange was adjusted by.

A similar adjustment needs to be made to the carrier phase measurement as well but it is not as easy to see in this example because the carrier phase measurements are relative rather than absolute and the two messages in this case use different references.  The carrier phase measurements are in cycles, not meters, so the frequency of the carrier phase needs to be included in the translation from clock error to carrier phase cycles but is otherwise the same as the pseudorange adjustment.  In equation form, the adjustments are:

P = P -toff*c
L =L – toff*freq

where P=pseudorange, L=carrier phase, c= speed of light, and freq=carrier frequency

So, basically, the receiver is trying to help us out by removing its best estimate of the clock error from the measurements.  This is unnecessary since RTKLIB is quite good at estimating this clock error on its own, but by itself this adjustment does not cause a problem.

It is when the adjusted measurement is translated to RTCM that we get in trouble.  The resolution of the time stamps in the RTCM format is 0.001 seconds.  In this particular example it would not be an issue because the error is exactly 0.001 seconds or one count of the RTCM format.  Most of the time, however, this error is not an exact multiple of 1 millisec.

For example, here is a time stamp for the data set described in the previous posts.

> 2017  1 17 20 31 48.9995584  0  9

And here is the same time stamp after being translated to RTCM and then to RINEX

> 2017  1 17 20 31 49.0000000  0  9

As you can see, the clock adjustment was less than half a millisec so was completely lost in the roundoff to the RTCM format.  However, the adjustments the receiver made to the pseudorange and carrier phase are still present in those measurements.  We now have a problem because the clock correction is in part of the measurement and not the other pieces.  RTKLIB can not correct for this lack of consistency within the measurement.

So, how do we avoid this problem?  Fortunately, RTKLIB has an option to adjust the time stamps to round values using the same equations described above to adjust time stamp, pseudorange, and carrier phase to maintain consistency within the measurement.   I imagine it was put in specifically to solve this problem. We can invoke this option by adding “-TADJ=0.001” in the “Options” box in the “Conversion Options” menu in STRSVR or using the “-opt” option in the command line with STR2STR.  Note that this option needs to be set in the conversion from raw binary format to RTCM format, not the conversion from RTCM to RINEX.  It is possible to set this option when converting from RTCM to RINEX but this won’t help because the damage has already been done in the earlier conversion.

Unfortunately, there is a bug in the implementation of this option in RTKLIB, at least for the u-blox receivers, so by itself, this is not enough.  The problem is that invalid carrier phase measurements are flagged in RTKLIB by setting the carrier phase value to zero.  The time stamp adjustment feature adjusts these zero values slightly so they are no longer recognized as invalid.  They end up getting included in the output as valid measurements and corrupt the solution.

Fortunately, the fix for this bug is very simple.  Here is the code in the decode_rxmrawx() function in ublox.c that makes the adjustment:

/* offset by time tag adjustment */
if (toff!=0.0) {
fcn=(int)U1(p+23)-7;
freq=sys==SYS_CMP?FREQ1_CMP:
(sys==SYS_GLO?FREQ1_GLO+DFRQ1_GLO*fcn:FREQ1);
raw->obs.data[n].P[0]-=toff*CLIGHT;
raw->obs.data[n].L[0]-=toff*freq;
}

If we add a check to the first line of code to skip the adjustment if the carrier phase is zero, then all is fine.

if (toff!=0.0&&cp1!=0) {

Below is the original solution after RTCM conversion on the left and with time tag adjustment and the bug fix on the right.  If you compare the solution on the right to the solution with no  RTCM correction in the previous post you will see they are nearly identical.

timetag

I am still wary of using RTCM because of its other limitations described in the last  post, particularly the loss of the half cycle invalid flag and the doppler information, but I believe this fix eliminates the most serious issue that comes from using RTCM.

I will release a new version of the demo5 code with this fix sometime in the next few days.  It will take a little while because I also want to include some other features that have been waiting in the pipeline.  If you want to try the fix right away, you just need to  modify the one line of code described above and rebuild.

Update 2/2/17:    I have taken Anthony Woolridge’s suggestion and modified the RTCM conversion code to automatically adjust the pseudorange and carrier phase measurements to compensate for any round off done to the time tag.  This means it is not necessary to set the time-tag adjust receiver option.  This change is currently checked into my Github page and I hope to post new executables in the next couple of days.