I’m new Sparkfun, but loving the possibilities. I handle the IT/Technology for an extremely small dirt work company (15 employees), so I’m a tech-savvy bulldozer operator. Most of the Big GNSS companies in our industry offer extremely expensive and complicated systems that offer way more options than we use or want. I currently run an Emlid RS3 base and rover for our jobs and drone work, but I’m looking at building a RTK kit for our crews that contains a base, either with a tripod or on a temporary site post, and a rover unit that would be in the equipment. My idea is to have an antenna/receiver, like the SparkPNT FPX-T, on the dozer mounted on a magnetic base and have it broadcasting the elevation to an enclosure with a display.
I have an understanding of mapping and survey work and I have been learning Python programming. So my question for this esteemed group is what would be the best use of my time in developing my skills for this project? And if you have any experience or advice as for the best setups for this project? Thank you
One of our users has an extremely interesting thread about using RTK x3 on a bulldozer blade to adjust road-leveling (which is quite a bit more than what you’ve proposed, but it has a ton of good info about caveats, issues, successes, etc) Mounting a Surveying Antenna to a Bulldozer Blade
How accurate do you want/need the measurements to be?
I would love to be sub-cm, but this project is proof of concept and usually within 1-2 inches is adequate.
Great - for sub-cm, we have a few options…some aren’t currently in stock so my recommendations are based on what you can order now
For the most seamless experience I want to point you toward the SparkPNT FPX - SparkFun Electronics - this is an all-in-one device (battery, screen, antenna, etc) that offers 0.6cm + 1PPM accuracy. It’s one of our latest “future-proof” models that would allow you to simply change the receiver board/module in the future (saves about 60% of the unit cost) should you want to upgrade in a few years.
We also have some options that are less expensive but are more susceptible to the potential dangers of a construction environment
This alternative model SparkPNT FPM - SparkFun Electronics also has L-band capability which allows it to receive RTK corrections via satellite signal (it acts as both base and rover in this case). You can also use a phone, a 2nd RTK module, or PC as a base.
Are there any other specs you need for sure?
Those were the models that I was looking at. I plan on using one of these with the tilt capabilities for the rover and a second without as the base. I’m wanting to view the RTK information on a screen either attached to a survey pole or in the cab of the piece of equipment. I would prefer to do this wireless using bluetooth or wifi. And I’m currently learning Python in hopes of building and programming this display. Am I on the right track?
Machine Control will be a great application for the Tilt Function in RTK Everywhere Firmware. The cutting blade (dozer, grader, etc) would have a defined “rod height” based on the antenna mount. So the operator’s blade controls (up/down, tilt left/right) would still be tracked at the cutting surface - with a GNSS antenna on both sides of the blade. Basically, you’ll want the RTK devices to Remain in Tilt Mode.
@Anthony_Carson , yes you are on the right track. Going wireless to the cab will be easy. Your in-cab CPU/Screen will receive standard NMEA positions and decide the cut/fill for the operator.
Python is a quick way to get started. I suspect you’ll spend more time getting the proposed grades ready for your data collector (in-cab CPU/Screen). That part of the workflow seems to be the most tedious. As you already know, it can change from job to job.
I’d say to decide early on how you will consume and prepare your proposed Terrain.
[ TERRAIN MODEL ] <-- the ground surface
/ \
[ Raster ] [ Vector ] <-- The format choice
/ \
[ DEM / DTM ] [ TIN ] <-- The actual data
TIN is the easiest.
Your data collector will wirelessly receive the NMEA position from the RTK device, apply the Geoid for an accurate Ortho Height, find the corresponding X,Y location in the proposed TIN, then calculate the ΔZ for the operator.
Note: do not use the orthometric heights from the NMEA sentences. Start with the Ellipsoid height and apply your appropriate Geoid. That way, your logic/system works with any RTK device. The GNSS chip’s built-in decimated Geoid wont screw-up your elevation accuracy.
I’ve oversimplified… but naturally your app will need to model the blade and handle the NMEA positions from both sides/ends of the cutting surface. You can break the system into 2 modes: rough grading and fine grading. Rough Grading gives the operator a basic cut/fill from the average across the blade. Fine Grading would give the operator an independent cut/fill for each side/end of the blade for “bluetop”. You can have a user defined variable for the threshold to automatically switch between Rough and Fine Grading depending on how close to finished grade you are.
I appreciate the information, it is an amazing help! Currently, I’m using my Emlid survey rig, and an Autel Evo II RTK to gather my data and using Virtual Surveyor to create my photogrammetry products. My owner is an ‘old school’ operator but has said that all he wants is an accurate realtime display of the elevation. We learn to grade or build sites by the feel of the equipment as it is working, we call it learning use the ‘bubble in your ####’ method. The instant verification of what you’re feeling would be amazing. My first goal is to mount one receiver on the top of a dozer and set the offset for the height. I would then let the operator set their target elevation and the display would show the elevation with cut or fill values beside the number. I’ll include a rough sketch of my dream.
OK, then you are way ahead of the game - since you control the output format of the existing ground surface. You wont have to fight/guess the Reference frames and coordinate systems (depending on how you deploy the Base).
You likely already know this, but it might save someone some trouble:
Whatever correction you use to setup a Base (or collect your GCP’s, control, etc) decides what reference frame & System you are in. If we stay consistent It’s easy to do it right, but it’s also easy to get it wrong 
That’s the #1 reason why people say “the elevations are not accurate from this RTK”.
That feels like a legitimate Phase 1. You’ll be able to display the approximate elevation (and cut/fill) of the ground under the dozer without too much fuss.