My solution in exam in edge(esp-32) to app(react native) This repo is a collection off the reapos i created in my exam EdgeToApp
Greenhouse system
The client was pleased with your Greenhouse prototype (Obligatory Submission 1), but they want a custom solution that is not limited by the cloud platform you used. Use your experience from the prototype to build a custom solution for the customer. They’ll expand the solution later, but for now they only need a Dashboard with a graph of the sensor data. The client has just started building greenhouses in 3 locations and want to use your new system to monitor all of them.
The requirements are:
- You must register and store Temperature, Humidity and Light from your ESP32 Edge device
- The client wants the solution to run on their own servers, so you must make a Node API to store the sensor data
- They want a Dashboard written in React that shows a graph of the last 24 hours of sensor data
- Deliver a video showing the final solution, all required code and written documentation. The reviewer should be able to replicate your solution.
You can build the Dashboard as either a web App or a mobile App written in Expo. If you are able to come up reasons to use other sensors in the Edge device, please do so. If you think the Dashboard needs more than just a graph, you are free to add features to the solution. Draw on your experience from the obligatory submissions and lectures to get ideas for good additions
The EdgeToAppExam solution is designed to facilitate greenhouse monitoring by seamlessly connecting edge devices (ESP-32 microcontrollers) to a user-friendly React Native mobile application. It collects essential environmental data, including temperature, humidity, and light levels, and securely transmits it to a Node.js server via MQTT. The server manages user authentication, real-time data visualization, and notifications, ensuring users can monitor greenhouse conditions effortlessly. With features like alarms, warnings, and firmware updates for edge devices, the solution empowers clients to maintain optimal plant growth conditions across multiple greenhouse locations, all while providing flexibility for future expansions and customizations.
When making the video i forgot to mention that i also send email notification on alarm and that the grow light is yellow and turn on when there is too low lux, the total hours are calculated at end of the day and sent, if it changes from sunlight/lamplight and less then half an hour has passed it will register 0h and above 30 min will be seen as 1h,so 1,23h will be 1 h and 2.44h will be 3h.
I looked in to how to find how much sun plants need and lux is most used value so added some information about lux values, also looked in to what the plants needed for best possible growing condition. I look over the data and write what values I will use and why then create a list of requirements that will be needed and a list of what we will monitor.
- Bright Summer Day: 100,000 Lux (~10,000 footcandles)
- Full Daylight: 10,000 Lux (~1,000 footcandles)
- Overcast Day: 1,000 Lux (~100 footcandles)
- Traditional Office Lighting: 300-500 Lux (30-50 footcandles)
- Common Stairway: 50-100 Lux (5-10 footcandles)
- Twilight: 10 Lux (1 footcandle)
- Full Moon: <1 Lux (<0.1 footcandle)
(Crane & Balerdi, 2020; Philips, n.d.)
(arbediskrav1)
- least 12 hours of sunlight. or 6 to 8 hours.
- well drained soil and water it good
- lot of humidity; at least 50% or higher
- Optimal temp 26 to 32 Celsius, stop growing at 15 and dies under 0
(Crane & Balerdi, 2020)
(arbediskrav1)
- Pineapple plants do not tolerate temperatures below -2 degrees Celsius and grow slower below 15.5 or over 32 degrees Celsius.
- The optimal temperature is between 20-30 degrees Celsius.
- Pineapples grow naturally in consistent humidity from 40 to 60 percent.
- The plants require at least 6 hours of direct sunlight per day.
(Crane, 2020; Gillette, 2023)
(arbediskrav1)
Optimal conditions for plant growth in the greenhouse include maintaining temperatures between 24 and 30 degrees Celsius. Plant growth slows significantly around 14 degrees and ceases entirely below 0. Therefore, 24 degrees should serve as the threshold for a warning signal. If temperatures fall below this, the heater should engage to restore optimal conditions.
Humidity levels should ideally be kept constant, between 60% and 90%. If levels fluctuate beyond a 5% range, a warning should be issued. Extreme conditions below 50% or above 90% should trigger an alarm.
Plants thrive under direct sunlight, so when it's sunny and hot outside, users should be advised to open the greenhouse windows. However, the indoor temperature should remain stable, and if it's cold outside, the windows should be closed to conserve heat. The door's state, open or closed, should also be monitored.
A minimum of 8 hours of sunlight is essential for optimal growth. If sunlight is insufficient, artificial grow lights should supplement it.
The window should be open when the outdoor temperature exceeds 24 degrees and it's sunny, to harness as much natural sunlight as possible. When temperatures fall below 24, the window should be closed.
Although there isn't a specific soil moisture sensor, the humidity sensor can provide an indication of when the plants need watering. If the humidity level drops below 50%, it's likely that the plants require watering.
Based on research (Crane & Balerdi, 2020; Philips, n.d.), a sunny day typically has a lux level over 10,000, and lower-end grow lamps provide about 5,000 lux. Indoor measurements showed a lux level of 300 with the lights on. In a production scenario, the grow lamps would turn on when lux levels drop below 5,000. For indoor applications, the threshold is set to 300 lux.
- Maintain temperature between 24 and 30 degrees Celsius.
- Issue a warning and push notification if temperature falls below 24 degrees.
- Trigger an alarm, send an email, and push notification if temperature falls below 14 degrees.
- Keep humidity constant.
- Issue a warning if humidity falls below 55% or rises above 65%.
- Trigger an alarm if humidity falls below 50% or rises above 90%.
- Try to ensure that plants get 8 hours of sunlight or artificial light per day.
- Monitor outdoor weather to manage window states.
- Monitor heater functionality.
- Utilize grow lights as necessary.
- Warn the user of sensor failure or if the device disconnects.
- All alarms should trigger an email and push notification (functional on real phones only, not simulators).
The following parameters should be monitored:
- Temperature (°C)
- Humidity (%)
- Light (lux)
- Heater (on/off)
- Window position (degrees)
- Grow light status (on/off)
- WIFI/MQTT status indicator (on/off)
- Door status (open/close)
- Connectivity status (online/offline)
The door status will be determined using a hall sensor and magnet. The window position will be evaluated using an accelerometer, converting the value into degrees to assess whether the window is open or closed.
The OpenWeatherAPI will be used to monitor outdoor temperatures. If temperatures exceed 24 degrees, an alarm will be triggered if the window is closed. Conversely, if the temperature is below 24 and the window is open, an alarm will be triggered.
An RGB light will be used to simulate a grow lamp (yellow) and to indicate alarms (red) when WiFi, MQTT, or sensor failures occur.
This will have 3 projects, a server, client project(greenleaf) inside solution folder and edge project.
you can run it local or connect the client and edge device to my live solution on azure, for details on setup go to Setup and look at local, remote or private remote setup.
you will find all topic for mqtt under the server routes folder in the wsmqtt.js file
device/devices are referencing esp32 devices and not the client app.
- devices // can be used by any device even when not logged in
- devices//profile to update device profile like location and password or FW to use and if auto update is on
- update : server will send an array whit latest FW for all builds, now its “dev” and “prod” this is for any device that are set to auto update
- locations//alarm : will send warning or alarm values on change it will send if alarm is on/off, type and alarm text.
- locations//live : used to send live humidity, temperature and calculated lux level(server will only store every 10 min)
- locations//light : will only be sent once pr day, it will send hours of sunlight and hours of lamp light per day
- locations/update: devices sends a location and gets back a city for that location, this is for when the client app updates the location of a device it needs to find what city the location is in so the weather API can get the correct temperatures
Are bit many to name them all but the main ones are
- locations
- Devices
- Login
Each endpoint has its own file in the server project under routes named after the main endpoints.
Our server, built on Node.js and Express, incorporates several additional libraries and services used beyond the standard ones. These include:
- nodemailer
- aedes
- bcrypt (for generating hashed and salted user passwords)
- crypto (to encrypt cookies, thereby preventing plain text signing)
- expo-server-sdk (for push notifications)
- mongoose
- websocket-stream and ws (for running aedes over websocket)
- this project runs both local and on Azure cloud whit mongodb cluster
The server operates three services: MQTT over TCP, MQTT over WebSocket, and HTTP. MQTT caters to the edge device, whereas MQTT over WebSocket serves the client project as I haven't identified a reliable MQTT library for react native that supports MQTT over TCP.
Both MQTT servers have been configured to use the same event listeners. This design allows the client and edge device to communicate effectively, regardless of whether they operate on different ports or protocols. When an alarm event occurs or the edge device stops transmitting data (triggering a keepalive event after 30 seconds), the server sends push notifications and emails to the client app/email. Push notification only works on a real phone so test it you need expo go on your phone.
We use signed and encrypted cookies for persistent sign-ins, once a username and password have been provided. Passwords are hashed with unique salts, ensuring identical passwords do not result in the same hash. This is facilitated through the bcrypt and crypto libraries.
Certain routes require login, and some even necessitate admin privileges. MQTT also employs username and password authentication, controlling who can publish and subscribe to various topics.
Every 10 seconds, the server scans the firmware (FW) folder for files in the format fw/dev/fw-dev-vX.bin or fw/prod/fw-prod-vX.bin (where X is a version number). It then identifies the latest version for each build and publishes it on the "update" topic, allowing devices to identify the newest version.
I use dev for local testing and then prod for when the device should connect to the azure VM running.(still need to setup the edge device from the hotspot, see Edge device for details)
Push notifications can fail at two points, and if not properly handled, your app's push notification functionality may be disabled. To counter this, we remove a token associated with a user if it fails during execution. We also record the ticket ID and token for that ticket, checking 10 minutes later (or up to 30 minutes for production) to verify if the push notification was successful.
Considering each user may have multiple devices, we store tokens in an array within the userModel. Simultaneously, we store tickets and tokens for verification in the validTokenModel.
The device will transmit the follow data:
- lux, temp and humidity will be transmitted live
- sun light data for each day will be sent once each day.
- there will be total of 17 alarms/warnings and status messages sent
- they will only be published on value changed.
- door alarm will also need too be open for a set amount of time before it get sent as an alarm
The edge device will transmit live data pertaining to light (lux), temperature, and humidity. It will issue alarms or warnings for high or low levels of these measurements, with specific thresholds detailed in the Monitoring section.
The device calculates lux based on the input from both light sensors. It will also interface with the Weather API to fetch external temperature data every 10 minutes.
The device uses an accelerometer to check the window's state, determining whether it's open or closed. It calculates the angle based on gravitational force, with an angle above 3 degrees indicating the window is open. If the window is open and the outside temperature is below 24 degrees, it sends an alarm. Similarly, if the temperature exceeds 24 degrees with clear skies and the window is closed, it sends an alarm. The device also triggers an alarm if the door remains open for an extended period.
The device broadcasts data over MQTT and uses HTTP for firmware downloads. If the device is set to auto-update, it will automatically update its firmware as soon as a new version is uploaded to the server. More information on this can be found in the Server Project section.
The edge device will fist start a hotspot called “ESP-WIFI-MANAGER “ on ip 192.168.4.1 where you setup your WIFI details and give the device a static IP for your network, then add inn the IP for where the Server is running. MQTT and http server should be the same IP.
The device will then connect too the given WIFI network and connect to the server, it will also start a new server on the static IP you gave where you can restart the hotspot if you need too change the settings.
If the device loses connection to the MQTT server over long period of time it will disconnect from the WIFI and restart the hotspot. When ever a device connects or reconnects due to a restart or a new device is connecting. the device will run thru all alarms and switch them on/off to test that all alarms are working and that the correct values get set. This will mean email and push notification will be triggered for each Alarm.
This project is built using React Native.
Key libraries utilized, in addition to standard React/React Native Expo libraries, include:
- react-native-picker/picker
- react-native-paho-mqtt
- expo-notifications
- react-native-chart-kit
Primary application screens include:
- Login: User authentication.
- Devices: Manage device settings, such as location or firmware.
- Locations: Create, modify, or delete locations.
- Main: Access main application features.
- Create User: Register new device credentials or create a new user (with potential for admin access after DB edit).
- Main Dashboard: Monitor statistics for a chosen location.
- Lux/Temp/Humidity Dashboard: View 1h, 6h, and 24h graphs for each dataset.
- Sun Dashboard: Display the past 3 days of sunlight hours versus lamp light for plants.
- Logs: View history and current status (ON/OFF) of all alarms and warnings that have been activated at least once.
- Error: Redirect to this page in case of errors like missing or incorrect cookies, or wrong password/username.
The application communicates with the server using HTTP and employs MQTT for receiving live data and alarm statuses from the edge device at a particular location.
When loaded onto a physical device via Expo Go, the application can also receive push notifications. If an alarm or warning occurs and the user clicks on the push notification, they will be directed to the dashboard of the location where the alarm was triggered to check the logs. Notifications function even when the app isn't actively being used.
- Node and Npm
- MongoDB installed if using Local Setup
- Mail account if testing email alerts
- Edit Arduino_secrets.h in edge project
- Edit .env in server project and make sure you change or comment out DB_URL if you are using localhost for db
- if you want too use your own username and password for client app and device you need to edit the start of server.js, we inject a admin and user at the start there if none are found.
The device will take some time under startup, it needs to wait for light sensors to be all the way active then it will calibrate all sensors.
When you connect a device or restart a device the setup will turn on and off all the alarms so you will get a total off 11 emails and push notification if you have them active while doing so. you might get the window open alarm on/off all the time it means your device is just at the threshold for open window(3degrees) so just turn it a bit and it will stop.
you need to go in to .env and change anything that has too do whit mail and DB to match your email and DB connection. also need to make sure the PORT is set too 3000
commands to run
npm i to start the server whit hotreaload
npm run devor run
node server.jsThe server will create 2 user, one to use when logging in to the client(both will work)
and one to set on the edge device once all is up and running.
the Admin account is test/test for username/password
It will also add dummy data for the past 24h for live data and 1 week for the sunlight data.
If you plan to connect the device later to the Azure server running you will need to select a unique name, but it cant be too long. reason I set the name is so i can easily see what device is doing what
You will need to build and upload the FW, then run “build filesystem image” and “upload filesystem image” restart the device and and then connect to ESP-WIFI-MANAGER hot spot and connect to 192.168.4.1. input your WIFI info and what IP you want the device to have(for resetting WIFIMANANGER) Then add the Ip that you run the server on for http and MQTT server IP. Next add the ports, 3000 for HTTP and 4000 for MQTT.
You need expo Go for android or iphone to get full use of the app. or you can setup an android/iphone emulator.
You need to change some files
in appcontex.js you need too change the IP to your node server IP and port to 3000 in app.js you need to change IP to your node server IP, port is the same for both local and cloud.
commands to start:
npm installthen
npx expo install // not needed but i run it since it gives feedback on things then start it using
npx expo startnow press a for android or scan QR code for your phone.
this mobile app has been design on an android pixel 3a phone running on a emulator so If things look bad or dont work on your phone I would recommend setting up Android emulator by follow this guide. have tested it on a iphone 7 but dose not look as good there, but all thing work.
https://docs.expo.dev/workflow/android-studio-emulator/then start the client project again.
Users created by the server on startup if missing.
admin:
//userName:”test”
//password:”test”
//user:
//userName:”edge-01”
//password:”edge”once your client, server and edge device is setup and connected to the server you can go in too you client app and log in as test/test then go too locations and add a location. Next go back to main menu and select devices, then the device you have connected and add the “edge-01” as username and “edge” as password and select a location for the device. once you click save it will start sending data to that location and you can go locations again and then select dashboard for the location you send to too and see the live values.
You can create new locations and move the device or connect new devices
Once you are done testing locally you can move the device over to production, and start broadcasting to the live Azure server. continue to “Remote setup” for setting this up.
If you want to connect your device and app to a live environment you can use this option.
You will need to go in too arduino_secrets.h and change NAME to a unique name other then “edge-01”
if you all ready have local setup running for server, client and edge you can use the client to change from dev to prod build on the edge device and then select version 1. Then the edge device should be sending too remote server instead.
If you don't have local setup you will need to build and upload the fw, then run “build filesystem image” and “upload filesystem image” restart the device and then connect to ESP-WIFI-MANAGER hot spot and connect to 192.168.4.1.
input your wifi info and add 20.253.246.222 as ip for mqtt/http server http port 80 and mqtt port 4000.
You need expo Go for android or iphone.
go to https://expo.dev/@freebattie/greenleaf and go to advance, select “ESA update”, “Expo Pro”, set channel to “master” and sdk to “48.0.0” or go here : https://expo.dev/%40freebattie/greenleaf?serviceType=eas&distribution=expo-go&scheme=exp%2Bgreenleaf&channel=master&sdkVersion=48.0.0 on IOS use cam, on android use expo app and scan QR code.
Or you can build it from source code, then you need to make sure the correct port and IP is setup.
in appcontex.js you need too change the IP to20.253.246.222 and port to 80
in app.js you need to change IP to 20.253.246.222
No need to do anything since you are connecting to the azure remote server.
- Azure Tools
- Remote -SSH
easiest way is to install azure tools in VSC and then from the server project create a VM (ubuntu)
if you installed azure tools you will have a tab for Azure form there you right click “Virtual Machine”
and select Create.
Follow the steps, you might need to try different regions if you are using free tier
you need to then connect too the azure VM over SSH (should have a exaction icon for it)
right click the VM and select connect and enter password you made when setting up the VM
right now NODE and NPM have breaking issues and the official package is to old and on some ubuntu version so you need to download it from a PPA (personal package archive), and when I tested you needed to use node 16.x and not 18.x
first you need to install PPA
cd ~
curl -sL https://deb.nodesource.com/setup_16.x -o /tmp/nodesource_setup.shyou can inspect the script if you like using nano or other editors like vsc
nano /tmp/nodesource_setup.sh
code /tmp/nodesource_setup.shnext you need to install Node and NPM (used apt-get instead since apt didn't work for me)
sudo apt-get install nodejs npmcheck that you got the correct version
node -v
v16.20then you need to upload the code to a folder in your vm, if you are using VCS you can just drag the server folder over and it will upload it, I set it up whit git and Github where i created a SSH RSA public and private key, then added the public key to Github and used SSH download link from Github to clone the repo since my repo is private. and i followed this example
https://linuxhint.com/clone-repo-with-ssh-key-in-git/once you got the code uploaded you need to add your mongodb cloud connection string too the server.js file.
next you need to connect too Azure and find you VM and go to networking and open inbound and outbound ports for port 3000,4000 and 4001 if you don't change them.
also write down public IP of your VM, you will need to change your Client and Edge project files to use this IP. again you need to uncomment the code that creates the first user in server.js
to start the server go in too you VSC SSH connection, from terminal go to the server folder and run
npm i
then start the server
npm run devthen stop the server using
ctrl Cthen install mp2 for running the server
npm install pm2 -gstart the server whit watch so it will restart on file changes, —log will make console.log() in app write to a file up one level in logs folder, this to stop watch form restarting when file gets written too:
pm2 start server.js --name greenleaf --log ../logs/file.log --merge-logs --watchsave changes
pm2 saveyou can now try and use MQTT explorer app to connect too the server and see if it works.
once you are happy with the setup and don't need too change the server app files any more you can turn off watch and or logging
pm2 stop greenleaf
pm2 delete greenleaf
pm2 save
pm2 pm2 start server.js --name greenleaf --log ../logs/file.log --merge-logs
//or
pm2 start server.js --name greenleaf
//then
pm2 saveif you want to run server on port 80 your app needs higher privileges, to avoid doing that i redirect from incoming on port 80 to 8080 on the server so server runs on 8080
sudo iptables -A PREROUTING -t nat -i eth0 -p tcp --dport 80 -j REDIRECT --to-port 8080 or you will need to give SUDO privileges when running server.