This project is part of EG2605 Undergraduate Research Opportunities Programme (UROP) undertaken during Year 2 Semester 2.
Features:
- View live raw sound data from sensors via Bluetooth Low Energy (BLE)
- Display processed data
- Heart Rate
- Blood Pressure
- Respiratory Rate
- Store and view historic data
- Stored as ranges and averages over different intervals
More detailed information can be found in the report. Link to initial Figma design can be found here.
Note: This react native project has been recently migrated to a bare React Native CLI project due to the need for Turbo Modules to run C++ code (which Expo lacks support for).
Firstly, install the necessary packages to build the project using Yarn
yarn install
Due to the need for the Apple Developer license to run on iOS, our primary target will be on Android devices for development. However, you can still run the application software on iOS devices.
Install Android NDK (version mentioned below) via Android Studio for Skia to build natively using C++. It can be found in this path: Settings > Languages and Frameworks > Android SDK > SDK Tools (Tab).
Next, specify the Android SDK path. You can either add the $ANDROID_HOME environmental variable or create a file named local.properties in the android folder and adding the following entry:
// Windows
sdk.dir = C://Users//<USERNAME>//AppData//Local//Android//Sdk
// MacOS
sdk.dir = /Users/<USERNAME>/Library/Android/sdk
The following are the recomended software versions used for a successful build:
| Software | Version |
|---|---|
| Java | 17 |
| Android SDK | 35 |
| Android SDK build tools | 35 |
| Android NDK | 25.1.8937393 |
Lastly, to build and run the project, simply run:
yarn start
If you run into an error that states: Cannot read property 'makeMutable' of undefined, perform
yarn start --reset-cacheinstead.
The summary page offers an overview of all the processed data obtained from the sensor. Currently, the application is able to show single value metrics (such as heart and respiratory rate) and double value metrics (like blood pressure).
Implementing the line graphs was achieved using Skia, a 2D graphics engine which can render animations on a separate UI thread, distinct from the main JavaScript thread. The D3 library was used to process the data into a format conducive to Skia operations.
The listen page offers medical personnel a comprehensive view of raw sound data measured by the sensor, presented in a cardiogram format.
In order to display incoming data at rate greater than the display refresh rate of 60 fps/hz, the function handling incoming values follows these steps:
- Data received from Bluetooth is stored in an array buffer.
- With each frame drawn approximately every 17 milliseconds (1/60 Hz), we calculate the number of pixels the pointer will cover in each frame rerender (17 ms).
- In cases where the buffer contains multiple data points during a frame rerender, we evenly distribute the incoming data into groups. The number of groups corresponds to the number of pixels displayed in each rerender.
- Subsequently, we compute the average value of each group and display the resulting value for each corresponding pixel.
This page functions as a repository for long-term data storage, drawing inspiration from Apple Health's interface. Data is presented using bars that represent the range of values observed during a specific interval.
Data is stored in the device's internal storage. To facilitate file storage operations within the internal storage of the device, the Expo File System library is used.