Integrated in the Embedded Systems course the goal of this project is to study, architect and develop a system capable of simulating a supermarket like cashier, that can weigh objects identified by a unique Radio-Frequency identification tag (RFID).
The system is composed by four key components: 1 Arduino responsible for weighing and reading the RFID tag for product identification, a Raspeberry Pi as database where the relashionship between PRODUCT - RFID and the previous contnt of cashiers is saved when a system shutdown happens and an Android smartphone that serves as display for the content of the cashiers.
In practice, the system will read the tag from the product while weigbing it. The product is identified by its RFID tag and its name appears on the Android device, as well as, its weight. The Android device is used to keep a list of the weighted products and to reset the list.
This project helped the group develop a whole new set of skills and apply Computer Science knowledge to the development of Embedded systems.
We started the project by preparing the GitHub repository where the code would be uploaded and changed as went made progress. By using a version control system we got the desired flexibility to really work as a group, where everyone was free to make suggestions, change things and get feedback of the colleagues.
As the semester when by we tried to guide ourselves as much as we could by our Gantt diagram. It showed us what we had purposed ourselves to in the last five weeks of the semester as the dealine to deliver the project drew near.
As a quickstart to the whole development effort we configured the Google Firebase Real-time database. By using Firebase we were not required to be physically together to be able to test the system as a whole. This was a great help because we were not always available to go to the faculty. The Firebase configuration went by smoothly. However we had to troubleshoot some problems as they appeared. With each others supoort and with the help of the Professors we were able to correctly configure Firebase.
As the weeks progressed we worked in smaller groups to develp the independent modules of the system. One group was in charge of developing the Arduino Modulo, another group took change of the Raspeberry Pi module and another worked in the Android app. As we worked we touched bases frequently and as necessary we tried to help each other out. The first step was to make sure that each component could connect to Firebase and from there the develpment of each component went by smoothly.
However, there was a major delay with the Arduino module due to the unavailablity of the scale to connet to it. We had to way almost four weeks until one was provided for the group to work. Due to this situation we had to delay the devleopment of the Arduino modulo and the testing of the system as a whole. Since the scale is the cornerstone of the entire system, we could not test the whole system until the last week before the due date.
On the other hand, on the Android app and Raspberry Pi sides there were no major delays and every task was performed on schedule.
- Install Android Studio
- Install git version control software
- Clone the code from this repository
- Have with you an Android phone or activate the emulation software
- Have access to Google's Firebase
- To be able to authenticate with Firebase you must compile the project using the debug.keystore provided. The easiest way to do this it to copy and replace the debug.keystore into the ~/.android/folder
- Open the project with Android Studio
- Run the app or build the APK and install
- For the realtime cashier system and history we used directly the Firebase Realtime databse standard SDK. Here we used a query linked with a listener for all the changes that comes fro the cashier
- For the Restart Cashier button, we used a REST endpoing in the Raspberry Pi, where we sent a requent wirh the ID of the cashier to be restarted
On the entry screen you must press the Start Cashiers button and go through the Google login to authenticate in the system.
By clicking the Humburger icon button on the top-left corner of the screen you are shown the app menu. From here you can get to the Cashiers, History or About screens.
In the Cashiers screen you are presented with the list of the cashiers available. By clicking on each cashier you have real-time access to the items that cashier has read. When you finish weighting the items in the cashier, you can pres the Restart Cashier button to send those items to History and start weighting a new set of items.
On the other hand, in the History screen. Click the cashier you want to check the History for and select the date and time to see the set of items read for that period in time.
- Paste the folder CashierServer into /home/pi of the Raspberry Pi
- Open Terminal
- cd CashierServer
- chmod 777 -R scripts
- sed -i -e ''s/\r$//' *.+'
- ./scripts/install.sh
-
Cashierserver.py: This file contains the connections that we stablished between the other systems of the project with firebase. Basically, is the server for the web.
-
scripts.js: Here we placed the functions that dynamically implements the web page.
-
index.html: It’s the main page for the server.
- Open the terminal and type: ./CashierServer/scripts/start.sh
- Open your favortie Web Browser and access your local IP address on port: 8080. For example: 192.168.1.127:8080
- To see the contents of the page you must login with a Google account. However, this only works if your Google account is on the list of authorized domains
For the Realtime cashier and history we directly used the Firebase Realtime Database SDK. Here we used a query linked with a listener for all the changes that comes form the cashier.
We used Cherrypy which is a web server serving as general station. We enableded a REST service for the communication of the Android app.
First click the LOG IN USING GOOGLE button to authenticate with Firebase Realtime database.
After this first step you have access to the information in the databse and the cashier history.
By clicking in History you have access to the content of the cashiers by date and time. Clicking on any date you can observe the products that where read on the selected day.
The scale cables are soldered are soldered to the HX/11 conversor module. Ten connect the HX711 pins to the Arduino pins following the above sequence.
| HX711 | Arduino |
|---|---|
| GND Pin | GND Pin |
| DT Pin | A1 Pin |
| SCK Pin | A0 Pin |
| VCC Pin | 5V Pin |
Link the Arduino to a computer via the USB cable. Then put the object on the scale with the RFID tag:
- Once all the cable links are done, the next step is to upload the scale code to the Arduino. First, we identify the product to be weighted by putting the RFID tag close to the NFC shield.
- Then we put the object on top of the scale and wait until the value appears on the console.
- Finally, to connect the Arduino with the Firebase it is necessary to run the command python CashierServer.py that starts the local server.
- Connect the Wi-Fi shield and the NFC shield on top of the Arduino board.
- Then connect the Arduino to the computer.
- The weight sensor is already attached to the scale.