We have a conference paper about CANdroid! You can find it here.Permalink
CANdroid’s poster was at the ASABE 2016 conference. If you have questions, feel free to contact us!
We utilized CANdroid to collect ISOBUS data during a manure spreading session. The session lasted for about 2 hours with a total of 6 passes of manure spreading in the field.
CANdroid sitting in the back of the tractor:
CANdroid collecting ISOBUS data while the tractor is running.
After the data collection, we parsed the some useful data out using the stored ISOBUS data and plotted them.
We will continue testing CANdroid for its data collection and the Cloud integration. Stay tuned!
We have a new ISOBlue video! This video walks through how to create your own ISOBlue, and then demos it!Permalink
We have a conference paper about ISOBlue! You can find it here.Permalink
Here are the slides from ISOBlue day at Purdue on April 29, 2014:
|Jim Krogmeier||Open Ag Technology and Systems Group Introduction|
|Aaron Ault||ISOBlue History and Current Status|
|Alex Layton||Libraries, Buffering, and Yield Monitor Demo|
|Hani Almansouri||RTK GPS|
|Sam Noel||Soil Telemetry and Weather Station|
|Aaron Ault||ISOBlue Future and Funding|
Thanks to all who participated!
Here is a first draft of the instructions outlining how to get it working for youself: in Google Docs.
Please send us feedback if you work on getting this going yourself and are able to help clarify any confusing points for future fellow travelers.
The first ISOBlue meeting will take place April 29, 2014 at Purdue University in West Lafayette. We plan to demonstrate the newly working RTK GPS functionality and a demo yield monitor Android app showcasing new buffering capabilities.
The goal of the meeting is to prioritize the next steps for ISOBlue and identify funding for future development. Some possible future paths include:
If you are interested in attending, please send us a quick email to RSVP so we can plan ahead.Permalink
We’ve been working on getting RTK GPS working on ISOBlue for about a year. We finally had some success a couple of weeks ago. In the chart to the right, see the measured results. The yellow points indicate the estimated position before RTK lock was achieved, and the green points indicate the position after RTK lock was achieved. Notice the scale has each grid box as a 1x1 cm grid. In this test, the box was sitting still.
We used an NVS NV08C-CSM-BRD receiver, an off-the-shelf cheap passive GPS antenna, and ISOBlue. We setup a free account with the Indiana Department of Transportation’s CORS network to get the RTK base station corrections. The output of the NV08C and the streaming base station data were both fed into RTKLib running on ISOBlue. ISOBlue did the corrections in real-time And viola! 1-cm accuracy for under $200.
We were excited enough by this test, we ran another. We found a survey location whose position is known extremely well, and placed our antenna on that spot. Pictures are below, and they are self-explanatory: the red dot is the known true location and the blue dots are the locations computed by ISOBlue. Only 1-cm away from the true known location! Testing is continuing…
The Holy Grail for ISOBlue has been to get yield and moisture data in real-time, and forward it over Bluetooth. We are partially there: we have yield! Using a calibrated yield monitor on the Case 6088, we logged data for several runs and kept track of the wet bushels of corn that the monitor tallied for each run.
The yield messages come through in PGN 65488. The first two bytes represent the mass flow (yield), and can be interpreted as little-endian integers. We summed them up for each run, and then divided the monitor’s wet bushel total by this sum to get a scaling factor. We then used the same scaling factor for subsequent runs and compared the results to the monitor’s wet bushel total for each run. The scaling factor for that particular combine turned out to be 1.895e-05.
The resulting numbers are very close as you can see in the graph to the right:
difference from the monitor tally.
We’re not sure exactly why we’re getting such a small deviation, but for now we’re happy to get less than 1% error from a calibrated yield.
In order to figure out which messages are from the yield sensor, we collected data with the yield sensor plugged in, then collected messages after unplugging it. We did the same for the moisture sensor. The elevator speed was apparently in the same message as the yield, since it remained at 0 while the sensor was unplugged. Unplugging the yield sensor triggered no alarms in the combine, but unplugging the moisture sensor resulted in an error about grain temperature being out of range.
Our next steps are to figure out the moisture messages, and geolocate the mass-flow numbers using the GNSS messages from the ISOBUS, and compare our yield map to the monitor’s yield map to identify which points in particular it is altering non-linearly. The monitor lists a standard delay of 12 seconds for grain flow, so we’ll be using the same metric. Achieving accurate points at the end of each row will be tricky, and we’ll likely need to find the header position ISOBUS message in order to know when the header is picked up and put down.
The NAME of the yield sensor lists the manufacturer as Ag Leader, even though the part number on the sensor is a CNH number. I understand from industry sources that Ag Ledaer is the standard manufacturer of mass flow sensors, so there is a good chance that the same yield message format exists on other manufacturers as well, but we’ll need some volutneers to upload other trace files in order to verify this.Permalink
Since pretty much everyone else out there is going to have the same problem of not knowing which ISOBUS messages are present on which machines, we’re going to start an open, public database of ISOBUS messages from as many machines as we can find.Permalink
We’ve successfully collected multiple runs of data without any issues from two combines: a 2008 Case 8010 and a 2011 Case 6088. Files containing all the received messages are posted here for future inquisitive minds.Permalink
The ISOBlue tutorial is up! Everything you need to know to go from zero to logging ISOBUS data.Permalink
Logged our first data from a 2011 Case 6088 combine for about 4 hours while harvesting wheat. The histogram of PGN’s is here and the mean and standard deviation of timings are here in CSV format format.Permalink
The BeagleBone Black can talk to a skid steer! Video. Watch it in HD quality to see ISOBUS messages on the terminals.Permalink