R is a programming language for statistical computing and data visualization. It has been adopted in the fields of data mining, bioinformatics, and data analysis.
We assume R isn’t useful for flipping switches and reading sensors. But that’s an assumption, not a fact.
Let’s be honest. Computer languages are just an abstraction layer on top of the metal. Languages provide constructs to make the expression of logic easier, tailored for the different ways people approach logic and data. But all languages eventually drive the behavior of transistors, and the logic gates and microcircuits built on top of those transistors.
We assume R doesn’t care about transistors. Possibly true, but I don’t believe that’s set in stone. In fact, I’ve proved it’s not true.
My urban garden needs irrigation during the summer months. The simple solution is to use commercially available irrigation timers; they turn on, then turn off. Simple enough. They cost about thirty dollars.
If you forget to bring them in during the winter, they freeze and break. After a few years, the plastic valves wear out and they jam closed (bad for plants) or jam open (bad for water bills). They don’t adjust to rain or hot weather.
You can buy more robust timers. Some of them have Internet connectivity and look up the weather forecast. They cost about $100. If you leave them outside, they freeze and break.
Or… You can use your knowledge of raspberry pi and the assortment of parts in the electronics box to build your own.
Raspberry Pi Zero
I chose a Raspberry Pi for three reasons:
I need internet connectivity if I am going to look up weather reports.
My preferred language is R, which I can run from a linux operating system, but not from micropython or C
I have several Raspberry Pi’s gathering dust in my office. I should use them.
I am using a Raspberry Pi Zero WH. The wireless makes it easy to connect to the internet and the headers provide a convenient way to connect relays and buttons.
The Bits
This project requires a mix of a lot of parts:
The Raspberry Pi controller
Designing the logic to convert the desired amount of water balanced against rainfall (measured in inches) into seconds to hold the valves open to deliver irrigation water (measured in gallons per hour)
Programming the logic in R
Controlling relays to control the voltage required by the irrigation valves
Plumbing the flow of water from house lines through irrigation valves and out to the drip irrigation system
This photo shows the development board. It’s just a piece of scrap wood with components temporarily screwed in place. At the top is the power for Raspberry Pi and irrigation valves. In the middle is the Raspberry Pi Zero, below that is the two servos used to control the irrigation voltage, and below that is the irrigation valves.
There isn’t any water connected to the system at this point – I’m only trying to test the electronics and develop and test the code. In particular, I’m trying to see if the software & hardware are providing realistic results. I don’t want to give it access to my water bill until I’ve reality checked its performance.
Here is an overall picture of the three parts on the finished board. Left: Irrigation valves and plumbing, Center: Raspberry Pi and relays, Right: power.
Notice I’ve moved the project to a larger and more formal board. This shape was more suited to the layout of the parts. I wanted the plumbing to be physically below power and the Raspberry Pi in case there was a leak.
This is a closeup of the raspberry pi. I’ve housed this in a mason craft food preservation container. Costs about $3 and is waterproof – hopefully something I don’t need.
The board chassis attaches the base of the container and includes a hole for feeding wires through a notch in the base and up to the RPi. Access is by pulling off the jar – it’s held on by a gasket. Because the RPi Zero uses so little energy, cooling is not a problem.
Software
I run this on linux – my programming language of choice is R. The code constantly changes – thank god for git and github. I do a lot of development on a Macintosh, then push to github. I ssh into the Raspberry Pi and pull from github to update the running code. This gives me a slight buffer from the online system, and github provides me with a backup should the memory chip in the RPi fail.
“Eeee-VAP-oooo-TRANS-PURR-ation,” I savor the word as I release it into our conversation. I’m still at the party with Marsha and Bob. We’re trying to determine why anyone (such as me) would want to use R on their Raspberry Pi.
Photo by Tima Miroshnichenko
“Big word,” says Bob. “What’s it mean?”
“Water evaporation from the earth and transpiration from plants,” I respond. “It’s a sum of the water escaping from my irrigation system. Look it up on Wikipedia.”
Marsha interrupts grumpy Bob; “So – That means, um…desired amount of water – rainfall + evapotranspiration equals the amount of water your irrigation system needs to supply.”
“Precisely,” I agree. “Until I found out about evapotranspiration, I was unsure how to account for temperature. I knew hot days would require more water because of increased evaporation; but was stumped how to translate temperature into increased inches of necessary water.”
“Never heard of it,” says Bob.
“Me neither,” I agree. “Evapotranspiration is handy, but doesn’t show up in all weather forecasts. Open-Meteo makes it available.”
“Say you’ve got seven days worth of this miracle number,” says Bob. “What does the R code look like?”
I’m at this party where Bob and Marsha and I are discussing the best languages for programming a Raspberry Pi. Bob advocates for Python, Marsha is a devout student of C. I’m defending my use of R. After all, Raspberry Pi starts with R. We have chased all the other guests out of the room with our conversation.
“With R, I have all sorts of built-in data management,” I say. “Manipulating matrices is in R’s basic DNA.”
Steve wanders in from the other room and joins our conversation. “Matrices aren’t a proper data strategy. You should be using a database. You can run SQLite on a Raspberry Pi with hardly any effort.”
Bob and Marsha simultaneously turn to stare me down. They are curious about how I’m going to get around this supposition.
“Sure. SQL with R–in particular SQLite, would have been easy to implement,” I pontificate. “Just call up RSQLite, push a few buttons, and Bob’s Your Uncle.”
“And that’s not what you did?” Steve is incredulous.
“I store the R object on disk and pull it into memory when I need it.”
“What kind of knucklehead stores data as a file on disk?”
I’m at a party, and the topic of programming languages comes up. A quarter of the room politely leaves, another half rudely leaves, and the remaining three people banter about the proper language for a certain project. Bob, Marsha, and I have the room to ourselves.
Tonight we’re talking microcontrollers: Arduino or Raspberry Pi. We like to connect things and aren’t afraid to release the magic white smoke with an ill-advised application of excess voltage. Bob assumes programming is done with micropython. Marsha prefers C.
“I use R to water my garden,” I say. “Yep, I installed R on a Raspberry Pi and use it to turn the water valves on and off.”
I’m working on my irrigation system. This requires a controller to turn it on and off. Cheap controllers are just timers. More expensive controllers factor in the current weather and avoid watering the garden during a rainstorm. Very expensive controllers (weather-based irrigation controllers) look at weather forecasts to avoid irrigating the garden when a rainstorm is going to happen in the next few days.
I’m satisfied with simple timers, but I forget to bring them in before freezing weather. They freeze, they break. Time to buy a new timer.
I’d like to be more sophisticated about my watering schedule. And I have all these Raspberry Pi and electronics parts laying around the house. It’s time to put together a Raspberry Pi Powered irrigation system with look-ahead scheduling based on the actual water needs of my plants. That’s a lot of technology to put together and get to work reliably.
