Thanks to those who attended my webinar last month entitled A Buyer’s Guide to GPS/GNSS Survey Equipment. I received many questions during the webinar and answered a few during the event. As with my webinar last October , I’ll post the questions and my comments here.
Question #1: When using GPS/GLONASS I understand you need at least two GLONASS SVs in order to gain any benefit from the GLONASS SVs, because one SV is required to compute the time difference between GLONASS and GPS time. However, I have heard that if you have an L2C-enabled receiver, then only one GLONASS SV is required as the L2C message has facility for the time difference. Can you (or any of the members) confirm this?
I just checked with (a colleague) who is an electrical engineer. We quickly Googled GGTO (I think) which is a message format contained within the new L2C signal, and it turns out that what I have suggested is true! I wish I had a good reference for you (and me). So if you have an L2C-enabled Rx and you are tracking at least one GPS L2C signal, then the time-offset message should be there and only one extra GLONASS satellite would contribute to the solution. Of course, this time offset would drift, but given that we are talking about atomic time standards, the time offset should be valid for at least a few hours, probably more. This is a pretty complicated reason for getting an L2C-capable receiver for now, but will become increasingly advantageous in the future as more L2C SVs go up.
Gakstatter: Craig actually asked this question right before the webinar (and also during the webinar) and we swapped a few e-mails. I have to check further into this but I don’t think it’s the case at this point because there are no L2C codes (messages) being broadcast now. The benefit of L2C now is the just pilot carrier. Last time I checked with the GPS Wing, they weren’t going to begin broadcasting the code on L2C until 2011 or so.
Question #2: 1) If you use OPUS and one receiver on site, how do you get redundancy between the on-site control points? 2) What software is available to convert epoch dates that actually works?
Gakstatter: Well, I consulted with my geodesist friend Michael Dennis, an Arizona PLS. He was presenting at the Alaska Surveying & Mapping Conference as well.
My first inclination was to suggest to use OPUS (assuming you have a L1/L2 GPS receiver) to establish the on-site control. Then, all of your control will be tied to the same reference frame…albeit no active baselines between the on-site control points.
I would occupy each monument twice at different times of the day. This should be sufficient to flush out blunders. If two of the sessions differ surprisingly or if the quality indicators on one are poor, I’d occupy a third time.
I ran my suggestion by Michael and he added some valuable insight and details that I glossed over (or downright omitted):
“I agree with your answer that a minimum of two occupations (of sufficient duration) be used to provide redundancy (but more occupations are, of course, better). “Sufficient duration” depends on whether OPUS Static (S) or Rapid Static (RS) was used. I usually work in areas far from CORS, so I cannot make reliable use of OPUS-RS, and so I typically want at least three hours (for OPUS-S). But for either type of OPUS, I recommend that the maximum peak-to-peak errors be less than the desired accuracies for the project. The peak-to-peak errors can also be used to compute a weighted mean final OPUS position. Waiting the ~two weeks for final IGS orbits is also recommended, if possible, but be sure to wait at least for the rapid orbits, which are supposed to be available in 17 hours. If three OPUS occupations are made, a sufficiently motivated individual could actually calculate the horizontal error ellipse and height error (scaled, of course, to 95% confidence).”
Michael had great comments on OPUS-S vs. OPUS-RS. If you’ve got gobs of CORS near you, then OPUS-RS might work, but I’d prefer to use 2+-hour (Michael suggests 3-hour) occupation times and run it through OPUS-S.
Some details on orbits. There are three grades of orbits used by OPUS.
- Broadcast orbits (available immediately).
- IGS rapid orbits (available the day after collection).
- IGS precise orbits (available 10-14 days after collection).
Which orbits to use is a bit of a challenge due to the time lag. Two weeks can be a long time to wait for a solution depending on the reason for setting the control. Submitting your data from the job site wouldn’t be the best move for a couple of reasons. The first is that you’d be using the least precise orbits, but more importantly data from many CORS aren’t posted until the next day. If you attempt to process the immediately after the data collection session, the selection of available CORS data might be limited. If you really require processing the data immediately, you should also process a day later and then again two weeks later to benefit from improved orbits.
Michael had a further comment about the lack of on-site ties in the example above.
“Having said all that, I must confess I’m not completely comfortable with the idea of using OPUS alone for establishing control. Maybe I’m being old-fashioned, but I would much prefer to have ties between all the stations on the project. Despite that, I must admit that OPUS has always given me good results (as long as I paid attention to the peak errors and minimum 3 hour occupation times for OPUS-S).”
Regarding software that converts epoch dates, I’d refer you to HTDP (Horizontal Time Dependent Positioning) offered by the National Geodetic Survey (NGS). You can use it to convert between reference frames and epoch dates. I think some manufacturers may have incorporated this into their software, but I would still do a spot check to make sure they both provide the same answer.
Question #3: Please comment on the limitations of GPS survey in challenging environments (canopy, terrain, etc).
Gakstatter: GPS will always be challenged by tree canopy and terrain due to the nature of the technology. Terrain is easier to deal with than tree canopy. With terrain, it’s just a matter of tracking enough satellites. You either track them or you don’t. An open-pit mine is a good example of that. Even when combined with GLONASS satellites, an open-pit mine of sufficient depth and steep enough slopes will prevent a receiver from tracking a sufficient number of satellites for a good-quality position. This environment is one of the reasons why pseudolite technology was developed. However, over time this will change as more GLONASS and other satellite systems (such as Galileo and Compass) are deployed. A fully populated dual constellation (GPS, GLONASS) will result in an average of ~20 satellites in view as opposed to half that (or less) with only GPS. If you add a fully populated Galileo constellation into the mix, now you have 90 satellites to choose from.
Tree canopy is a different story because it’s not a &ldq
uo;hit or miss” proposition.
The receiver will pick-up and drop a satellite dynamically when tracking under tree canopy. For centimeter-level positioning, your receiver needs to consistently track the satellites it is using in order to provide a reliable position. The temptation is to push a receiver into an environment where it can’t provide a reliable solution to “just get the last shot.” The risk is that the receiver will report good quality indicators (fixed solution with low RMS values) but record a poor position. Even worse are the scenarios where the position is reasonably close to the actual position (within a few feet), but it’s not easy to detect the blunder since the quality indicators are good. You’d rather the position be grossly incorrect so the blunder is obvious.
I think the long-term solution to precise positioning in that environment is the integration of several technologies like GNSS, inertial navigation, laser rangefinding, and other technologies. All of these technologies exist today, but they aren’t integrated into a small enough and user-friendly enough package at reasonable enough prices. That problem will be solved with time.
One thing I believe for sure is that GPS/GNSS will not solve that problem completely even with the modernized GPS signals (L2C, L5, L1C) and the addition of other satellites from systems like GLONASS, Galileo, and Compass. Yes, there will be a marked improvement in that environment, but not completely solved.
Question #4: Is the survey GPS industry responding to the challenges of the oncoming solar maximum event? If so, how are they responding?
Gakstatter: I think you’ve got to define which GPS technology is most venerable. That would be the users who are trying to optimize the accuracy of single-frequency GPS (L1) by modeling the Total Electron Count (TEC) — particularly, real-time correction systems like DGPS, SBAS (WAAS, EGNOS, MSAS, GAGAN), and commercial DGPS services. Dual-frequency receivers, although not immune to the effects of an extreme event, are much better equipped to deal with dynamically changing TEC within the ionosphere due to the known frequency dependence of the delay.
This subject is worthy of another article by itself (I published one last fall), so I won’t go into much detail here but rather save most of the detail for another day.
The GPS industry isn’t doing anything at this point except keeping an eye on sunspot activity. Keep in mind that extreme solar events typically happen on the downside of the solar cycle, which is 11 years long. The first four years of the solar cycle are the ramp up. We are starting the ramp up so the solar maximum will be in the 2012 timeframe. The last extreme solar events occurred about two years after the solar maximum, so if we use similar timing, the extreme events of the next cycle will occur five to seven years from now. There’s much debate though. Some experts are suggesting that maybe this cycle will be a dud, and so far it has been tame.
Everyone seems to be in monitoring mode, and experts don’t even agree on how severe this cycle will be. The National Geodetic Survey says, “We’ll know when we get there.” In essence, nothing is being done to prepare and I’m not sure there is anything to do.
In the October 2003 extreme event, DGPS accuracy blew out to 15-20 meters and WAAS accuracy blew out to 25 meters. Commercial DGPS users complained about accuracy blowouts also. WAAS is the only system that actually monitors and warns users of the accuracy blowouts (if the receiver is designed to utilize the warning that WAAS provides).
The good news is that this should be the last solar cycle where we have to worry about this as much as we are. By the time the next solar events might happen (2025), we will have all the GPS modernized signals deployed to mitigate it (primarily L5 and L1C).
Question #5: I’m a surveying engineer from Romania. What can you tell us about VRS? Recommendations?
Gakstatter: Briefly, RTK networks are experiencing explosive growth around the world. It’s a topic one cannot avoid when discussing GPS/GNSS today.
I’ve used various GPS/GNSS equipment on networks operated by Trimble, Topcon, and Leica software and receivers. They are very, very convenient.
It’s a complex subject. Look forward to my next column that will delve into RTK networks.
Question #6: Do you know of any studies of real time accuracy obtained using CORS base-station networks (with the cell-phone data link)?
Gakstatter: I assume you are referring to RTK networks. I’ll write more about this next month, but I’ll say a little here.
Like I mentioned above, I’ve used several different receivers on several different RTK networks. My general feeling is that traditional base/rover configuration gives you better control over accuracy (especially vertical) than RTK networks, primarily due to control over the baseline distance. Of course, if you are using a traditional base/rover configuration and start roving 10-12 km from your base, you’ll run into the same problem. The idea is that you have control over the baseline when you operate your own base station and you don’t when you’re tied into an RTK network.
But one can’t dismiss the robustness of the RTK network solution using many reference stations versus the vulnerability of a single baseline base/rover configuration. More later on this…
Question #7: I’ve read somewhere L1 receivers will not be usable after 2020. Is this true?
Gakstatter: Not at all. I’ve written quite a bit about the Department of Defense’s intent to discontinue supporting semicodeless techniques after December 31, 2020.
It only affects L1/L2 receivers that use semicodeless techniques (about 300,000 of them). If your receiver can utilize L2C, then it is fine.
L1 receivers will not be affected at all.
Question #8: Is cycle slip a problem when trying to use an L1 RTK system in a real-time application?
Gakstatter: My experience with L1 RTK says that it’s a useful tool for clear-sky environments when there are enough satellites available and you use a base/rover configuration of the same brand. It performs especially well when you have SBAS satellites (WAAS, EGNOS, MSAS) within view because it uses them like another GPS observable.
When used in the environment it was designed for (as described above), cycle slips aren’t an issue in my opinion.
Question #9: Are you guys planning any webinars on using RTK networks? That would be a good topic!
Gakstatter: In fact, my next webinar (in April) will cover this very topic.
Question #10: When do you plan to retire your Ashtech system?
Gakstatter: When it stops working J. I think no one will be able to fix it when it does.
Interestingly enough, I’ve been able to utilize it as a base station with the new Magellan PM-500 (without GLONASS).
Question #11: What are typical price ranges of each class of receivers?
Gakstatter: Here are my guesstimates based on U.S. prices. My prices are the entry level for the category:
- GPS L1: US$7,000 and up for a pair of receivers and post-processing software. L1 survey units really work together the best in pairs due to l
imited baseline distance.
- GPS L1 RTK: US$12,000 and up for a pair of receivers, spread-spectrum radios, and data collector.
- GPS L1/L2: US$8,000 for a single receiver with internal memory and without post-processing software. The assumption is that the user would utilize an online positioning service such as OPUS, PPP, or AUSPOS.
- GPS L1/L2 RTK: US$19,000 and up for a pair of receivers, narrow-band radios, and data collector.
- GPS/GNSS L1/L2/GLONASS RTK: US$27,000 and up for a pair of receivers, narrow-band radios, and data collector. US$15,000 and up for a single receiver and data collector configured for RTK network operations.
Question #12: If they are semi-codeless and will not work after the sunset, does this mean that the modulation scheme will be changing for L2?
Gakstatter: First of all, the GPS Wing has made it clear that the sunset isn’t a hard date, so receivers may work after that date. They just won’t guarantee it.
My understanding is that there will be no change to the modulation scheme for L2. The GPS Wing recommends that civilian receivers utilize the new L2C signal.
Question #13: L5 will improve the precision of positioning in high covered areas? Thank you!
Gakstatter: I sort of covered this in Question #3. L5 will really benefit the civilian high-precision user in a few ways:
- mitigatingthe effects of the ionosphere.
- four times more power than L2C.
- enhanced code structure for more robust positioning.
- resides in the highly protected aeronautical frequency band (1176.45 MHz).
I wouldn’t expect that just because the broadcast power is four times greater than L2C that one can expect L5 to “punch through the trees,” although it will help contribute to a more robust position solution.
Question #14: Any thoughts about L1 GPS/GLONASS/WAAS RTK receivers? The product can do L1 RTK, support network RTK, use online free positioning service, and utilize wireless service for base/rover communication, price is 1/3 to 1/2 of those of GPS L1/L2 RTK systems.
Gakstatter: Honestly, I don’t have any experience with that type of receiver. I’ve used L1/WAAS RTK in a base/rover configuration and on a network. The base/rover configuration worked well within its limits. The RTK network configuration wasn’t so good. I think most of the problem was due to the baseline distance. The nearest reference station in the network was nearly 20 km away.
However, I can only assume that if L1/WAAS RTK works well within its specifications, that L1/WAAS/GLONASS RTK would work that much better with the additional observables in a base/configuration.
Lastly, my experience is that most networks (if not all) don’t support broadcasting SBAS data and some do not even support GLONASS. Maybe this will change in the future.
Question #15: Why do GPS users still think that LI RTK is “high-precision GIS”? A centimeter in a surveying app is still a centimeter in a GIS app. Do you agree that most GIS users expect more than 0.5-meter results?
Gakstatter: Well, I hope I didn’t lead people to think that is the only use for it. I think L1 RTK can be applied to construction staking and topography surveys similar to L1/L2 RTK as long as it’s operated within its stated limits.
I think the value proposition of L1 RTK puts it in a price range that GIS users can afford RTK where they couldn’t before. Just think that 10 years ago, the price tag of a sub-meter GIS receiver was about US$10,000.
Question #16: How soon do you think inertial navigation will be a marketable solution?
Gakstatter: There are some out there now, but not at the right packaging/integration/price-point level. I think we’ll start to see mainstream products in the 3- to 5-year timeframe.
Question #17: Is it worth it to pay more at this time for an L1/L2 RTK GPS system capable of receiving signals that will be available only after 2 or 3 years?
Gakstatter: If you buy a GPS L1/L2 receiver (no L2C) today, there is only one system you need to consider and that is the semicodeless sunset date of December 31, 2020…12 years from now. GPS L1/L2 RTK systems are getting cheaper and cheaper.
Just because new signals are being broadcast in the future (L5 and L1C), it doesn’t mean that your GPS L1/L2 system won’t work any longer.
Question #18: A recent article in Geomatics World (Jan/Feb 2009) suggested that the inclusion of GLONASS signals marginally worsens an RTK position in areas of variable sky view (robust intercomparisons were undertaken it was carried out in the football stadium of Old Trafford in England).
Gakstatter: I haven’t read the article. I would be interested in reading the details.
To me, users select GLONASS to work in environments where using only GPS lacks sufficient satellites. It’s all about productivity and not as much about accuracy. Of course, one would prefer it not to degrade accuracy. This is a good subject to look at in more detail. My experience with GLONASS hasn’t demonstrated this, but I can’t say that I took a scientific approach in comparing the two. It was on a couple of projects where using only GPS was cutting into my efficiency due to GPS “brownouts” because of the terrain. I ended up using a GPS/GLONASS receiver and was pleased with the productivity. There wasn’t a noticeable degradation in accuracy either.
Question #19: What do you know about the quality of Altus receivers?
Gakstatter: I haven’t used the Altus product, although I’ve spoken with them and I know some of the guys who started the company…very experienced GPS people who used to work at Leica and Magnavox. They use a Septentrio OEM receiver. Septentrio has developed a reputation for very good receiver technology.
Question #20: I hear rumors about how different manufacturers of GLONASS receivers process the data differently. I understand that some process, or “handle,” the data significantly differently, and that some don’t handle the data very well. Can you talk about this a little?
Gakstatter: I have some experience with GPS/GLONASS receivers from a couple of different manufacturers. In my experience, the receivers performed in accordance with the product specifications inasmuch as I was using them for RTK.
I wouldn’t doubt that manufacturers are handling GLONASS differently, but it’s difficult to determine who is doing it “better” than other manufacturers.
I think the best way to make the determination is to try it yourself in your environment remembering that the benefit of GLONASS is to increase productivity, not increase accuracy. When there are plenty of GPS satellites in view (6+ with a low PDOP), there is no need to use GLONASS.
Question #21: Considering cost/performance, L1 is the most expensive. What do you think? If a fully loaded state-of-the-art receiver costs $5K more than a simple L1, what is the economic impact over the lifetime of the receiver (5 years) considering all other expenses of a survey company?
Gakstatter: I understand your point. I think it depends on what kind of projects a survey company is participating in. If they are doing large scale topo and construction staking work, then I would agree that they should seri
ously consider a state-of-the-art RTK receiver. In that environment, an L1 receiver would hinder productivity.
However, if it’s a small, low-overhead shop performing residential lot surveys, then an L1 receiver might deliver the maximum efficiency. It’s simple to operate and simple to maintain.
Keep the dialogue going on these comments. I think it’s a great discussion and I’m open for comments and criticisms.
Story filed from 65o 3’ 11’’ north latitude, 146 o 3’ 20’’ west longitude. This is the furthest north I’ve been in North America.
Also in the March newsletter: About Alaska