Samsung Innovation Campus: Final Project
Source Code File w/o Sound Source Localization Function
This project aims to gather and classify the traffic sounds with Raspberry Pi, and later display them visually from a mobile app to the driver with hearing loss. I hope this will be beneficial for future work for providing safe driving for everyone.
URBANSOUND8K DATASET: Link to Dataset
Yarasa - Android.zip is to-be-created .apk file may be installed on an Android device.
Contents of Yarasa - Pi should be placed on a directory of your choice on your Raspberry Pi.
These project elements are yet to be set for an automatic connection. Following steps shall be followed for a smooth operation.
You need a Raspberry Pi 3+, a ReSpeaker 4-Mic Array for Raspberry Pi, and an Android smartphone for this system.
Firstly, your Raspberry needs to know how to operate a ReSpeaker. Simply follow the instructions on ReSpeaker Wiki.
Hint: This process might be a little challenging, but remember that instead of using "sudo ./install.sh", using the following can save your life:
sudo ./install.sh --compat-kernel
Not my favorite solution, but gets things done.
On your Raspberry Pi, you should follow the instructions given on Yurockkk's Repistory to obtain a successful Bluetooth connection between Android & Raspberry Pi. Shout-out to them for the help of their project for establishing the Bluetooth connection between these parts.
Simply run Yarasa-Pi.py on your Raspberry Pi. If this is your first time running such a program on your Pi, you might struggle with libraries for some time, thanks to the limited resources of the Rasbian environment.
The Android device that has the Yarasa - Android.rar on it should be connected to your Pi via Bluetooth.
And that should do it. This project is a very basic first step, and certainly open for improvements, especially for sound source localization. Together we can build an equally safe, easy, and enjoyable living for everyone.
For this great dataset, URBANSOUND8K
Bluetooth connection, Pi side: Yurockkk's Bluetooth-RPi-Python
Bluetooth connection, Android side: Yurockkk's Bluetooth-RPi
Signal processing: Shubham Gupta's work