Tilt compensation for GNSS rovers has evolved — a lot. It has been less than a decade since the first tilt compensation features were integrated into GNSS rovers. However, compared to the latest systems, that functionality seemed primitive, and often rather frustrating to operate.
The first wave of tilt compensation used magnetic compass orientation, which required (often complex) calibration routines, and suffered from the inherent inconsistencies associated with relying on magnetometers. While such features represented gains in efficiency, the reception by end users was lukewarm and adoption was not broad. But the next wave, integrating IMU’s and GNSS for orientation, finally delivered on the promise of precise and painless tilt compensation.
The steps leading to the R12i
On September 1, 2020, Trimble has announced the addition of integrated, calibration-free tilt compensation to its flagship rovers — the R12i
There were several carefully planned steps leading up to this development, and to put the discussion of the new feature below into context, let’s look at the evolution of said flagship rover.
First announced at the INTERGEO exhibition in 2012, the R10 was a distinct departure from their legacy line of rovers. It was a compact unit, hardened for harsh environments (and user rough handling), and designed to accommodate multiple (and emerging) constellations. It also included magnetic oriented tilt compensation.
Tilt was, and is seen as, a real gain in efficiency, especially when seeking to put positions on points that could not accommodate the placement of a rover leveled on a pole, like against a building, under dense canopy or under a building overhang. Or where safety was a consideration; like getting a shot out on the edge of a busy highway, or at the edge of pit. And one of the biggest sources of error is in levelling the pole. Tilt sought to set the user free from such constraints.
Tilt on rovers was not unique to Trimble. There were many of these early implementations. I had tried many of these, and while most were decent, they all shared the same limitations, especially having to calibrate — and re-calibrate. This is due to inconsistencies in the magnetic environment of a site or area, and the loss of repeatability over time. But even with such limitations, production gains could be substantial. For example, on a survey of a gravel road around a reservoir I had performed with an early R10, we contrasted its performance with a legacy rover that did not have tilt.
I headed off in one direction, shooting the crown and road edges, and another surveyor headed off in the other. I had set the R10 to take a shot when I tilted it up level (within a preset tolerance), and the other surveyor levelled their rover up manually with the bubble on the pole. When we met up on the other side, I had covered about three-fourth of the loop, to the other about a one-fourth. I had heard from other users (of various systems) that the tilt did come in handy for such rough topo and stakeout and getting difficult shots. But of course, we all wanted better.
The next models in the R10 family, were the R10-2 (2018) and R12 (2019). The R10-2 was a hardware revision in preparation for the next advances. It was a complete upgrade of the processing power, and added certain other efficiencies, but the big leap forward was the introduction of a new RTK engine (2019), one that the additional processing capabilities enabled. The original R10 (as with many rovers of that era) fell a bit short on the processing power needed to deftly and optimally integrate all of the newer constellations and the multiple signals they provide into RTK solutions. Trimble calls this the “ProPoint™ GNSS positioning engine,” and owners of R10-2’s can do a software upgrade to implement it, or one can buy an R12 (launched in 2019) with it installed as standard.
The big gain from such updated RTK engines is the ability to get the most out of the full range of satellites and signals. I had tried out the new engine, and while there were several other rovers that performed nearly on par with the R12 (and on par for a few), most did not. The claim of up to 30% productivity gains some early adopters reported seemed quite reasonable. More satellites have brought a lot of potential, and measurable (no pun intended) benefits.
This equates to being able to work in more environments, where legacy rovers had always struggled — where one would often also want to use tilt compensation. Once all these pieces were in place, the conditions were ripe for calibration-free tilt compensation.
Also Read: Improved gravity data brings geospatial greatness
Q&A
I spoke with Chris Trevillian, a surveyor I have known for many years, who is now a Product Marketing Director for Trimble Geospatial GNSS, about the new rover, and the tilt feature they call the “Trimble TIP™ technology.”
I noticed the housing looks the same as the R10.
We wanted to keep the same housing, the same form factor. We did this to keep some consistency with the accessories we have with the top end of the portfolio, like the quick release and the adapter to set it up as a base station on a tripod. The form factor is hard to beat in the industry—small and lightweight. Customers love it.
Wasn’t more room inside needed for the IMU?
The accelerometers and gyroscopes are all mounted directly to boards inside. We can do it without taking up much space at all. We did this also with modifications in the R10-2 and R12; updating processors and capabilities inside the unit. It’s pretty similar on this one. Instead of a magnetometer we used in the past, we are using a suite of accelerometers and gyroscopes.
Speaking of gyroscopes, there has been some misunderstanding (I see in surveying forums) about how IMU and GNSS integrated tilt works; like folks worrying that the gyros are mechanical and will break down, or that they drift too quickly. And that leads some folks to believe that it might not be any better than the magnetic oriented ones they tried (and were often disappointed in).
The gyros are not mechanical; there are no moving parts in the solution. The kind of understanding in the market is that IMU lose heading quickly, that IMU drift. An IMU is very good at understanding relative position to where you just were, but over time they can’t track long movements without drift. Magnetic-oriented solutions are not going to respond to the changes in magnetic fields to be consistent.
What we are doing in fusing the GNSS side of the equation is leveraging the ProPoint strength of the GNSS solution, integrated with the IMU, so we are able to get that repeatable and reliable heading determination, which allows it to be accurate and fast. The IMU is sampling at 200hz, and the GNSS is not — but no one is moving so fast as to need to have GNSS at such a high rate. So, we actually have to sample down the IMU in order to integrate it tightly with the GNSS engine, and those two give us the ability to compute the (pole) tip position in real-time.
This is not entirely new. Mobile mapping systems and UAS that integrate GNSS and IMU have been around for a long time.
Like our POS-AV for airborne, POS-MV for water borne, and POS-AV for mobile mapping systems, as you mentioned, it is a complex process. For instance, on a boat, the position has to take into account heading, heave, yaw, and roll — integrating GNSS with IMU and accelerometers is the solution. Our Applanix division has been making these for many years; adapting and fitting such a solution on a rover took a lot of hard work, but we have succeeded while keeping it in the same housing. As in these other systems, it is the heading, or trajectory of the GNSS antenna, that keeps the IMU from drifting, by syncing it up on a constant basis.
Another misconception I hear is that some people are under the impression that you must take two GNSS measurements and relate it to the vertical and tilted position.
If it worked the way you described, you would lose your heading. With ours you get the orientation from the last heading angle. We are constantly sampling both the IMU and GNSS to always provide the heading accurately and thus the TIP position in real time without two measurements. It one seamless motion. And as long as the GNSS solution is working the TIP compensation is working.
However, you do not have to always keep the GNSS antenna moving. You could lean it on the pole up against the truck for instance, and we put the receiver in static mode — it knows it has stopped moving. We constrain the IMU and keep it from drifting. But when you pick it up and start moving again you continue where you left off, and the solution continues to update as you move.
Can you fall back on standard RTK mode?
You can go into a GNSS only mode, turning off the IMU when you want, if that is looking for what you want to do. And in that mode, the eBubble turns on automatically to still provide you the traceability for your vertical measurements.
How long does it take for the tilt feature to kick in?
The alignment process with the IMU is so fast that typically when you are walking from one point to the next it is aligned before you know it. Initially, it might take 5-10 seconds, sometimes more in a bad environment. But once have initiated, you have compensation on all points; it does not have to do a 5-10 second initialization every time. Though if you lose the tilted solution for some reason you will have to do this process again. But that is under very bad GNSS environments of if you lay your receiver on the ground or something like that, situations when you would expect to lose GNSS positioning capabilities.
What quality indicators does the user have for the tilt mode?
On the data collector, in Trimble Access, we tried to simplify as much as possible. There is a little indicator for orientation, and the eBubble goes away when you are in IMU mode. We played with altering the eBubble to indicate tilt quality. But here is where test user feedback really changed our thinking: they really wanted to focus precision values. Rather than just some color-coded bubble, a dot moving in the bubble, the users wanted to readily see the precision values we are estimating based on the tilt angle and based on the GNSS solution.
When not in IMU mode, can you still set it to take a shot when the pole is level?
Yes, all the same functionality is there, but may be becoming a bit moot, as you do not have to be completely level to get a good position. In both modes, you can set it to take a shot only within tolerances that you preset.
It is true that there have been several implementations of no-calibration tilt by other firms. I have tried some of these, and have found that some perform quite well, some others that take a little getting used to, but all are at least a step up on the older magnetic oriented rovers. What do you feel sets yours apart from the others?
The simple answer is in the performance of our GNSS engine. We recently did tests along with some other new systems. We have a test course we set up to really push the limits of the ProPoint engine: tree canopy, up against buildings, and tough multipath environments. Out of all the points we shot in that test course, the others were able to shoot zero points while in tilt mode, but ours hit all of them.
A tilt solution itself is only as good as the GNSS engine that drives the rover. If you get next to a building or under a tree and you lose your good RTK solution, the tilt function is disabled. That was some of the feedback that was most critical to us in development of this solution. People like tilt features but get frustrated when it does not work when they need it the most. That is why we put so much emphasis around the ProPoint engine being as robust as it is. That was our focus in making the feature worthwhile.
Survey rovers can take a real beating; they are an outdoor, hard work tool. How can you tell if the wear and tear is hurting the tilt solutions?
We have built in the IMU Integrity Monitor. Customers are familiar, like drone pilots, that occasionally, you must take your drone, flip sideways and upside down to recalibrate the IMU. The good news is that we have the GNSS coupled with the IMU that kind of check and balance one another.
With all IMUs, there are inherent sources of bias. Those can be abrupt shock, temperature, ageing of components on chipsets, etc. All bias introduction is monitored inside the receiver, and we call this the integrity monitor. If it ever goes over the threshold, we trigger an alert to the user, and say: you may want to take a close look at what is happening to your positions. In that instance, we shut of the IMU and tell the user to recalibrate. They can do that in the field in a matter of roughly a minute. It’s super easy, but to honest with you, I do not expect many users to ever have to do this.
But the fact is that these pieces of equipment take a beating in the field sometimes, and certain customers do not treat them as well as we think they should, or our engineers think they should. So, we put this in as a failsafe; if we notice something wrong, we tell the user to check and recalibrate if needed.
Key takeaways from this launch, we expended a lot of effort to:
- Make sure the feature works where the customers want it to work.
- Make sure it is reliable and people can trust it.
You decide
I have not completely confirmed this yet, as this only just now got announced (though the social media teasers kind of gave it away), but it looks like adding tilt is not going to be very costly. The price differential between the R12i with the tilt function and the R12 without it, is likely only to be about 5%.
I note that Trimble did release a no-calibration tilt feature on a construction rover more than a year ago, and I kept asking when we would, see it on a survey rover. The answer I got was that no one wanted to rush without extended testing and customer feedback to refine it (surveyors can be quite demanding when it comes to precision). I do remember trying some other unrelated instruments that seemed to just want to get some hot-topic-feature-or-other out there in a hurry. Not always pretty.
There was a lot of skepticism, and some less-than-stellar experiences, about the first wave of tilt (of various kinds). And this latest wave of tilt carries its own triggers for skepticism. The concept of achieving stability from movement seems counterintuitive, but if you think about, for instance, the physics of GNSS, the high-speed orbits make for precise tracking and even prediction. For many years, IMU+GNSS-based solutions have been working wonders for airborne, waterborne, and mobile mapping systems — now you can have that same power right there in your rover.
It is almost surveying blasphemy to state that any kind of gee-whiz tech could be more precise that traditional bubbles (though consider the compensators in your total stations). But I am not alone is saying that the new wave of tilt may just have crossed that frontier. I have test-driven several the new tilt units — yes, against points shot with a total station, and I am impressed. I have test driven the R12, and will do so with the R12i, but like anything, the proof will be in the pudding. It is always my advice that if you are skeptical about what an instrument claims it can do, do not just go in circles on what-ifs and assumptions… try them out. You will be the best judge of whether it meets your standards or not.
Also Watch: What is PPK and how does it help in achieving high-level accuracy in surveying
