Activity Feed › Discussion Forums › GNSS & Geodesy › How does GPS tilt compensation work?
How does GPS tilt compensation work?
Posted by jt50 on October 21, 2019 at 12:24 amI was wondering how does a gps tilt compensator work? I have a vague idea that it uses compass o gps coordinates & a kind of electronic pendulum to get deviation from vertical but how does it know where the tip of the rod is pointing on the ground? Or are there 2 compensator inside? Near the top & bottom?
Just thinking out loud after reading a lot of new gps receivers have this new feature.
rover83 replied 2 years, 9 months ago 28 Members · 48 Replies- 48 Replies
Gonna follow this as I’ve been curious about this also and I too have a couple of ideas on how it works.
If you can measure tilt angle (liquid surface or other sensor?) and antenna azimuth (compass or MEMS gyro or ?), you can combine those with rod height to get the ARP-to-tip vector.
The Trimble R-10 has some calibrations you have to do on a regular basis. I can’t think of the names right now. Then the face of the unit has to face you so that it has orientation not just for tilt but for the electronic bubble. I am sure there is way more than that inside.
My concern is how accurate is the MEMS in translating the tilt to ground coordinates? If an upright receiver would give you 5mm+1ppm. Would the slant computed coordinates still retain this 5mm+1ppm accuracy?
Natester or Shawn help me out here. Doesn’t Javad offer equipment with automatic tilt compensation that doesn’t have constantly have to be calibrated?
I am preparing to buy two new sets of Javad equipment prior to the sale of the business as as condition of the sale.
@flga-pls-2-2
The internal compass has to be calibrated when the magnetic enviroment changes. This could be a couple times a day or couple times a week. The Level sensors need calibrated but it hold for quite a while.
The calibration process can halt the use of the tilt compensator. We use it to get points under fence lines for the most part, tilting the receiver and grabbing a point can save time and increase accuracy, instead of setting offset points or line points to locate a monument, simply tilt the receiver and put the tip of the rod on the dimple.
The problem was that at the discovery of a monument that needed the function, the receiver has often needed to be re-calibrated. So you end up using offset points and line points to locate the corner, faster than the calibration process.
If it is calibrated it works great, and if you know you are going to need the feature, calibrate and work that job using it. Big metal fences,,,,,be careful.
It is easy to test to make sure it is doing what you want, simple really.
We were impressed how accurate it is, the key is the HI, any error in that will make a corresponding error in the horizontal location for the point. Of course if you are plumb over the point the horizontal location will not change if there is an error in the HI, not so when tilted. We don’t usually care much about the elevation of a property corner, but when in tilt mode it’s critical.
Thanks, per msg. I will.
As has already been stated, tilt correction is based on the amount of tilt from plumb, the direction of tilt related to the coordinate system, and the distance from the ARP to the pole tip.
The direction is usually determined by one of two methods: magnetometer (digital compass) and inertial measurement. The direction from a magnetometer must be corrected for magnetic declination and is subject to magnetic interference (just like any other compass). The magnetometer needs to be calibrated somewhat routinely and often depends on location. Typically I calibrate about once per week, sometimes more frequently, sometimes less. The software should include a model of the magnetic field to account for magnetic declination, since magnetic North varies (sometimes quite significantly) from geodetic North. The inertial systems are not susceptible to magnetic interference. Calibrations occur constantly during movement. I have some questions regarding inertial systems. I believe in general they are more precise and more robust than magnetic based sensors, but I believe they may still have some limitations. IMU’s are subject to drift over time. The longer the time from the zero up, the more error the IMU will have. From what I’ve read, the current IMU’s used for tilt correction require calibration every 60 seconds. Calibration is based on the rover moving with a fixed RTK solution. In the open, this is not difficult. In canopy, this could be problematic. What about collecting points with a leaning offset that require several minutes to acquire a fix? Perhaps the best system is a hybrid that uses the IMU in a dynamic situation or magnetic in a static situation. Perhaps the IMU could be calibrated by means other than fixed RTK positions. As I said, I suspect IMU is generally more robust than compass, but from the white papers I’ve read, I suspect it isn’t the best answer to all situations.
The tilt is determined by tilt sensors in the receiver head. Consider the “system” of a rover. You’ve got the receiver head and the pole. In the receiver head are PCB’s (printed circuit boards) on which the tilt sensors are mounted. These sensor are extremely tiny. The receiver housing has a threaded female receiver or a bayonet mount to connect to the pole. What is the likelihood that the tilt sensors are perfectly mounted to the PCB (rotationally), or that the PCB is slightly rotated in the housing, or the threads in the connection are slightly out of perpendicular to the housing? Whatever the magnitude of these errors may be, there is certain to be some error. So a proper calibration should include a “taring” of sorts, like zeroing a balance scale to remove the systematic errors. It’s simple enough. Put the receiver on the pole with a calibrated level bubble that will be used in the field and force the tilt sensor to zero when the pole is plumb. Without this calibration, there is not guarantee that the coincident with the tilt sensors.
The pole height must be accurate for tilt compensation to work properly as the process, in essence, is creating a 3D vector from the receiver’s ARP to the pole tip. If the height is not correct, then the vector is not correct and the point will not be correct.
As a conscientious surveyor, I would recommend always locating a check point from two directions if possible or perhaps rotating the head and repeating the observation to have some confidence that the tilt correction is correctly being applied to the point.
Well stated, Shawn.
Nate
- Posted by: @shawn-billings
As a conscientious surveyor, I would recommend always locating a check point from two directions if possible or perhaps rotating the head and repeating the observation to have some confidence that the tilt correction is correctly being applied to the point.
There lies the dilemma . Are there any deviation values or accuracy readings from the tilt sensors included in the position files to show an error circle for the point coordinate? How will the user know that the calibration is off on a particular reading? We can’t have the field guy calibrate the unit for each point or every 10 or 20 points reading because then it would defeat the purpose of using a GPS unit to make work a bit faster.
With the invention of modern and more sensitive equipment comes the necessity for the operators to be more capable in order to be released into the public and using the equipment.
I for one have not been able to allow my helpers to use javad equipment and other rtk GPS because they will not first demonstrate they can properly operate the equipment.
I was not allowed to use a transit until I could show that it was set up properly and how to adjust the vials if needed and turn angles and double them and care for the tripod and accessories and bring them back day after day intsct.
Training is an important part of our profession as should be. The learning curve for new equipment is part of what we do.
There is no foolproof equipment being made that will constantly give surveyors the correct absolute answer every time and every agency that regulates surveyors does not judge us on the same statistics that the equipment is being evaluated on.
We as surveyors make our choices based upon what we can live with, equipment and personal and liability and the bottom end.
0.02
I thought tilt compensation would be handy. Over 1/2 of my offset shots are near fences, buildings, bridges etc. I’ve decided to stay plumb, and do my own offset data. Just to keep the “goof” out of my data.
Now, using the tilt sensor, out in the open, to auto-start-stop, now that’s the bomb!
Nate
The tilt sensor works well, it gives us results that are very close to plump measurements. It is simple to test, if you are worried about error, it’s the same as anything, redundant measurements are required.
If you are looking for a statement of accuracy (5mm+1ppm for example) I can’t help you, all I can say is our checks have been between 0.00-0.02′.
The R10 uses tilt sensors (accelerometers) as well as magnetic orientation, and they can be used independently of one another. If you only want to use the e-bubble to ensure that you’re always plumb you can do so. I calibrate the tilt sensor on a tripod and tribrach, and then rotate it as a check on the calibration of both the R10 and the tribrach. Calibrations must be performed every 30 days maximum.
If you want to use the magnetometers to collect compensated points, there are two additional calibrations that must be done and that should be performed whenever you go somewhere that the earth’s electromagnetic field might be different than where you calibrated last – the instructions on this are pretty vague. When you measure a compensated point, the data collector stores values for electromagnetic interference, and Trimble has guidelines on how to interpret these values. This is essentially your QC / reliability check.
From my experience, the compensated points aren’t worth the effort – most of the situations where I’d want to use it are in places that have too much interference. But I like the e-bubble; not only is it more precise than the pole bubble, but you never have to take you eyes off of the data collector screen. Also when performing an integrated survey you still get the e-bubble when you switch over to the robot, which is absolutely fantastic when collecting topo features.
I’d like to see the IMU based rover in action, but the only way that will happen here is if Trimble comes out with their own version of it.
OK, now I know a little bit more (or less) about tilt sensing with GNSS.
Is there a report that delineates which GNSS instrument uses which technology? Lets start with the Trimble R10, Javad, and Leca GS18t
and did any of the models change technology midway through their product cycle?
I can recall plenty of times questioning the instrument operator how they could write down 190?ø when it obviously is more like a 50?ø angle between the marks. So, I don’t want to say paying attention is a lost art, but it’s still one worth pursuing. If you know your orientation and rough amount of tilt and what it means, then you have checks probably in all equipment. In the LS there is a ppk solution in post processing that tilts don’t apply to (at least in my limited testing). So I located a found rebar using tilt among a thorn apple tree recently, tilted roughly north and east to avoid branches, 1/4 foot or so in each direction, about 3?ø tilt maybe on roughly 5 foot pole. PPK solution shows about that distance and direction to rtk tilt solution and agrees with math of my rough approximation. In addition rtk tilt within .25 of published which agrees with other found evidence while ppk would be half foot. I’m satisfied for purposes of that rural, rough terrain larger parcel. Got to talk with field crews before and after, over and over, and eventually most can be trained to pay attention. FWIW, haven’t calibrated the LS in a few months and it took a nasty fall on the pavement a couple weeks ago. Seems to still be good. Dont’ follow that last practice kids. Calibrate, calibrate, calibrate; redundancy, redundancy, redundancy. And pay attention! Surveying and texting don’t mix; wait for a text break.
The Trimble R10 uses a magnetometer that has to be calibrated. Like Lee said, it’s kind of a pain. I had grand plans/hopes to survey a bunch of building corners with it to skip offset routines, but it pretty consistently says it’s got too much magnetic interference to be of any use. While the ebubble is sorta handy, I’d rather have the screen real estate since we run all TSC3s.
In the advertising banner at the top of the site page, I’m sure I saw one of the paid advertisers here, Tersus, advertising GNSS calibration free tilt compensation.
From Leica’s website (straight copy and paste, no 1st hand user experience here, but am a long time Leica user so follow their stuff closely):
What makes the GS18 T different? How does it work?
The Leica GS18 T is different from any other product in the market because it does not rely on a magnetometer to correct the pole tilt for each measurement. A magnetometer is affected by metallic objects, like vehicles, iron fences and beams or even reinforced concrete, all of which are often found on construction sites. In order to provide accurate and reliable tilt values, Leica Geosystems developed an application-specific, powerful and lightweight inertial micro unit (IMU), which is built into the GNSS antenna. The real -time tilt compensation combines GNSS data with the IMU??s tilt and direction values. Different to magnetometer-based GNSS antennas, the GS18 T??s tilt compensation does not need to be calibrated and is immune to metallic object interference.
And how accurate are the measurements made with tilt compensation?
The accuracy depends on many factors. There is the accuracy of the GNSS position. If we look at the specified RTK solution, it is the same as with the GS16 ?? typically 8 mm + 1 ppm (single baseline length) Hz and 15 mm + 1 ppm (single baseline length) V. The accuracy of the tilt measurement is divided between the tilt accuracy and the direction of tilt accuracy. The tilt accuracy will typically be better than 0?ø12’00” and the direction of tilt accuracy typically better than 1?ø00’00”. This means over a 1.8 metre pole length, the overall tilt accuracy is typically better than 20 mm at the pole tip with a tilt angle of 30?ø. This accuracy is comparable and most of the time better than using a conventional GNSS pole with a 20?? level bubble to make the pole vertical.
If the pole is tilted more than 30?ø, the accuracy of the tilted measurement reduces for two reasons:
- Tilt accuracy declines as the tilt is increased and the direction of tilt accuracy has higher impact
- With increasing the tilt angle, the antenna starts to lose track of some satellites, therefore reducing the GNSS position quality.
In any case, the position quality indicator in Leica Captivate shows the combined GNSS position quality and tilt quality, therefore representing the true pole tip accuracy.
Pole tip accuracy of 0.0175 metres, shown in Leica Captivate v3.0 on the field screen.
Other brands may also have a no calibration needed tilt sensor??
SHG
Log in to reply.