IoT

We implemented an IoT by Arduino, Firebase and Android Studio. For Arduino, we use temperature sensor (DS18B20)、pH meter and CO₂ sensor (MG811) to monitor our device. Also we use LCD to print data and use Wi-Fi sensor (NodeMcu) to upload our records to database (Firebase) as well. We use Wi-Fi sensor to get current time and we send it to the Arduino so that it can print it out on Arduino. Also we send the temperature, pH and temperature data, which is on the Arduino, to the NodeMcu. Then we use the NodeMcu to send the data to firebase.

For user-friendly, user can read the data easily from the App. When the data is abnormal, the App would have a caution to let the user know, so that the user can adjust the device in time.

Arduino code

We use temperature sensor (DS18B20), pH meter and CO₂ sensor (MG811) to monitor our device. Besides, we use LCD to print datum and use Wi-Fi sensor (ESP8266) to upload our records to database as well. (You can see more information about Arduino in hardware)

Temperature experimental curve fitting

Data Fitting

1. Given a set of data (including input and output), create a mathematical model to approximate the input and output characteristics of this data.
2. If the data contains one-dimensional inputs and outputs, the mathematical model can be expressed as a curve, which in this case is also called curve fitting.

Therefore, we use the temperature of the thermostat (incubation) on the market and the analog signal sensed by Arduino as the basis for the temperature calibration curve.

Since the error value of the temperature sensing device is not necessarily, we use multiple sampling methods, and then average the values measured in each temperature interval to reduce the error value without excessive error. We use the above method to take 150 strokes as training data, and use the basic accessories of MATLAB to find the appropriate nine temperature model formulas.

From the graph we can see that the more parameters the model contains, the smaller the square error will be. Therefore, if the number of parameters is equal to the number of data points, the squared error will be equal to zero, but this does not mean that the prediction will be the most accurate, because the data points contain noise. The model that completely matches the data also means that the model is most affected by noise. The accuracy will also be poor. In addition, the more model parameters, the higher the time complexity, so we use the last image (cubic) as the curve of analog signal and temperature conversion. Because the instantaneous temperature does not only contain the exact value of the time complexity can not be too high.

Database

Why we need to use database?

Save data from our device, including carbon dioxide concentration, pH value and temperature obtained by sensors in our device. Besides, the App can fetch the information from our database.

What database we use?

We use Firebase to be our database. Because it is a realtime database and a 3rd party Authentication (Google Signin, Facebook Login, Github, Twitter). Last but not least, its cloud messaging can send push messages to any device, so we can easily send it to our App. Based on above reason, we choose Firebase.

How to create a Firebase?

We divide the following instructions into two parts, register firebase and create Firebase project.

Register firebase

1. The quickest way to use Firebase is to sign in with your old Google Account and get started. Sign up and then go to the website.
   Register Firebase account
   Go to the Firebase homepage: http://firebase.google.com and find the "Login" link registered with Google as shown below in the homepage.
2. After signing in to your Google Account, you will be prompted to confirm your permissions. Click "Allow" to complete the registration of Firebase, then click "GET STARTED" on the page to start using Firebase, as shown below:

Create Firebase project

1. The first thing to enter the Firebase console is to create a Firebase project and try out the free program. After entering the console, please click on “Add Project” to create a new project, as shown below:

2. Enter the project name of the App, such as “project”, and then select the country/region, then click “Create Project” as shown below:


Finally, you can wait for Arduino to send data by Wi-Fi to the database.

App

To make our device more user-friendly, we designed an App to get data from our database. The manager of the factory can easily see the data immediately. Besides, we set up a notification system. As long as users connect their smart phone to the internet, they can immediately get the warning notification from the device if the data goes wrong.