ludwa6
commented
4 years ago Sense-Able Farming with farmOS
The Context: Given that the farm is by its very nature a complex open system in a dynamic environment, it is essential for all operations under management to incorporate strong feedback loops into their decision-making processes. As owner/operator of a quite diversified farm operation, I have struggled for years with the challenge of how to do this in a systematic way, before concluding that what is really needed is an Operating System for our holistic farm enterprise -and moreover a responsive one (vs a stiff "black box" solution whose logic we could never fully understand, let alone control). Having concluded after some exploration of accessible options that farmOS appeared to fit this description best of all, at least on the face of it, the first step would be to see how well it could listen to what our operation had to say about itself, record those inputs, and give back answers to our questions on-demand, all in a minimally frictional way.
To that end, i see different listening modes, which point to two specific UseCases:
Remote Sensing from fixed locations
The essential problem, as stated by Chris Rowe in documentation of his GrovePi/Node-RED remote sensing solution, really nails it for me: "I am a farmer and I want a tool to remotely know what is happening in the environment because I can't be everywhere all the time." exactly! .
As to solutions: FarmOS has affordances, including a well-documented API (Application Programming Interface) to support the integration of environmental sensors, as described on the farmOS.org site. In the Documentation section under "Assets," see the "Sensors" page for information about how to connect sensors -for example the aforementioned solution from Chris Rowe. In our case, we decided to trial the Quahog solution from Edge Collective, mainly because developer Don Blair was so accessible and helpful... And so it is that you can see on our farmOS instance, under Assets:Sensors menu, the page for "Quahog1" sensor (login as: guest/guest1 for read-only view of all gives a pretty good view (including line graph plots) of the time-series data we've got flowing in from both gateway and remote node version1 (presently located at our greenhouse).
NB: This is a limited apparatus -just 2 sensors in one location for soil moisture and temperature- that still in R&D mode (Am just seeing now, as it happens, that the sensor has been offline for some time; must go check gateway power and maybe push the reset button) but good enough to prove the concept that we can pipe data from pretty much any sensor(s -and there are many accessible options!) via cheap commodity hardware into farmOS with little ado and nary a hiccup, once you get the hardware sorted out.
Mobile FarmWorker Observations
As stated above, Remote Sensing being essentially a workaround for the problem of scarce farmer-bandwidth, the primary feedback mechanism remains the farmers presence, for two most important reasons:
- Technology being, as Douglas Adams famously quoted Bran Ferren "stuff that doesn't work yet, it is important for the farmer to monitor that the technology is working as intended, observing its impacts (including the unintentional) on production; and
- As per the ancient proverb: "Best fertilizer is shadow of the farmer." For this reason, the point of Remote Sensing is to free up farmer to give attention to those things that the machine cannot do (e.g. examining leaves for potential health issues, making ad-hoc interventions as appropriate, etc.)
So, as a handy supplement to our Remote Sensing apparatus, we use FieldKit on iOS and/or Android mobile platforms as the human interface for recording ad-hoc observations, which -tho not yet out of beta- is already proving useful (login as above) for the one fairly limited UseCase that I've outlined in the Issues tracker, along with related bug(s -or enhancements, as developer Jamie Gaehring has classified this set: fair enough). As evidenced by this comprehensive list of issues, the countdown to version 1.0 release is well underway, and I am cautiously optimistic (based on the rhythm and quality of beta releases that JG has released since I joined the program last summer) that we'll soon have a good & handy tool for anyone with a smartphone to pipe observations from the field (including GeoLocated photos) directly into farmOS. Go Jamie! :-)
Conclusions
Inconclusive, in a word, yet Promising (ok that's two words :-). Though I cannot yet say that we've got a production-ready solution to either of the aforementioned UseCases, we do have a solid proof-of-concept on both fronts, and a few questions for further research w/r/t each, i.e.:
- Remote Sensing: As we will be needing a good few sensors, and quickly accumulating quite a pile of data thereby: can we efficiently & effectively digest these dataflows in such a way as to derive sufficient insight to make investment in same worthwhile?
- FarmWorker Observations: Can FieldKit evolve in such a way as to be both agile AND perceptive enough that it will achieve critical mass of adoption by our field workers?
Those are the questions I will be working with my team to answer, the next little while.. And for you, farmOS user in your particular context, i would ask:
- Can you relate to either or both of these UseCases? Have you another take on the overall concept of "Sense-Able Farming," and can you share experience/ ideas for using farmOS to enable it?
If so, then we'd love to hear from you; please join us in discussion of this topic on the farmOS Discourse system... And have a Sense-Able Day! :-)
mstenta
This request is really a test of the workflow to submit proposed blog-posts via pull request.