NTU CE/CZ4171 Internet of Things: Communications and Networking

Part 1: Term Paper - Low Power Wide-Area Network

Term Paper Description

The objective of the term paper is to:

• Conduct a literature review from a range of pre-selected topics
• To write a 4500 word survey paper to summarize the related research works

Assessment Criteria

The term paper should fulfill the following criteria:

• Topic selected: To be significant, timely and impactful
• Content of the paper: Scope to be well defined, and contents should be comprehensive in terms of broadness and depth
• Analytical and critical information: Paper to provide insightful analysis
• Presentation: Paper should be clear and logically organized, while easy to understand

Contents

In this submission folder, the following files are available:

• Term Paper - LPWAN.pdf: Submitted Term Paper based on the current Low-Power Wide Area Network technologies.

Part 2: StarSearch - Flutter Application for Animals Detection

Project Description

The objective of the project is to:

• Host a Virtual Machine with convolutional neural network capability to do animal detection
• Create Android application ("StarSearch") to communicate with the Virtual Machine

Demonstration Video

Contents

In this submission folder, the following files are available:

• Image_Recognition.ipynb: Google Colab Notebook used for training.
  • Dataset used: Animals-10 dataset. Source: Dataset
  • Statistics:
    • Training Loss: 0.2541
    • Training Accuracy: 91.55%
    • Validation Loss: 0.5997
    • Validation Accuracy: 81.16%
• /Android App Code/starsearch_iot: Folder containing the source code for the Flutter Application
  • pubspec.yaml: File used to import modules in the Flutter Application.
  • AndroidManifest.xml: File required for permission controls in Android.
    • To be stored in "/android/app/build/src/main" folder
  • /lib: Folder containing all the dart files to build the Flutter application
  • README.md: File containing the instruction for installation of the Flutter application
• /Server Code: Folder containing the source code to be implemented in the Virtual Machine
  • app.py: Code used to run the Flask Application. TensorFlow Lite is imported in this application as well.
  • /model/imagereg.tflite: TensorFlow Lite Model used for inference. Exported from the Image_Recognition.ipynb file.
  • /templates/index.html: HTML file to allow users to upload the image to the server through the web browser
  • requirements.txt: Text file containing the list of dependencies for the server application.
• README.md: Readme file.

Additional Materials

Available for reference (until 31 October 2022):

• OneDrive for Business:
  • Includes the following files:
    • PowerPoint Slides
    • Presentation Video
    • TensorFlow Model
    • StarSearch Application
  • Folder Name: CE4171 IoT Communications and Networking Submission
  • OneDrive Link (Requires NTU Account): OneDrive

Running the Project

For Server

• Run the Flask server in the root folder by executing the command:

$ python3 app.py

• Record the IP address and port allocated by the Flask server. It will be used for connection to the client.

Note: Firewall permissions and port forwarding must be allowed for incoming connections. Otherwise, the client cannot connect even with the correct settings.

For Client (Demonstration Only)

• Download and install the "app-debug.apk" file.

Link in Additional Materials section.

Note: The application will only work when the StarSearch Server is operational.

Additionally, while iOS, macOS and Web files exist in the folder, the application was tested only on Android.

• Alternatively, follow the README.md in /Android App Code/starsearch_iot to install into your device.

Note: Please replace the StarSearch URI with the IP address allocated by your service provider.

Modify the MainScreen.dart in /Android App Code/starsearch_iot/lib/Main.