GrassFarm

This is my new approach to building the Urban Grass Farmer application. In the first approach, I started with the user interface which was inspired by my frustrations with OpenSprinkler. However, I realized that I want something that 'just works' rather than something that gives me surgical control over creating and managing schedules. In this version, I'm starting with the control mechanism functions that will:

• Launch a 'watering'
• Interrupt a 'watering'
• Log the results of a watering (zones + times)
• Configure the system for the number of zones connected
• Configure watering restrictions and preferences
• Connect to a weather API and configure the location
• Log local daily rainfall reported by the API
• Determination if it is time to launch a watering
• Determination if a watering should be interrupted

Then, I plan to use Fuzzy Logic to implement and tune the watering process.

1. The System will have a default set of assumptions about restrictions or preferences
2. The System will use the watering restrictions/preferences to create a predicted time to start the next watering
3. The System will use the fuzzy logic model to determine if a watering should take place and how long each station should water
4. The System will allow users to provide feedback like "grass is too dry" that will tune the model

Weather API

I've chosen to use Tomorrow.io. It's free API provides the data we need and has plenty of additional options should we want to track more weather variables in the future.

Environment variables

Include a .env in the root directory with the following values. Don't forget to run source .env from the root before mixing and/or deploying to the device.

Variable | Value Description TOMORROW_APIKEY | The API key from your tomorrow.io account NERVES_SSID | SSID of the wifi network you plan to deploy the device in NERVES_PSK | Wifi network password

Targets

Nerves applications produce images for hardware targets based on the MIX_TARGET environment variable. If MIX_TARGET is unset, mix builds an image that runs on the host (e.g., your laptop). This is useful for executing logic tests, running utilities, and debugging. Other targets are represented by a short name like rpi3 that maps to a Nerves system image for that platform. All of this logic is in the generated mix.exs and may be customized. For more information about targets see:

https://hexdocs.pm/nerves/supported-targets.html

Getting Started

To start your Nerves app:

• export MIX_TARGET=my_target or prefix every command with MIX_TARGET=my_target. For example, MIX_TARGET=rpi3
• Install dependencies with mix deps.get
• Create firmware with mix firmware
• Burn to an SD card with mix burn

Learn more

• Official docs: https://hexdocs.pm/nerves/getting-started.html
• Official website: https://nerves-project.org/
• Forum: https://elixirforum.com/c/nerves-forum
• Elixir Slack #nerves channel: https://elixir-slack.community/
• Elixir Discord #nerves channel: https://discord.gg/elixir
• Source: https://github.com/nerves-project/nerves