Info:

1. siren ML model uses tenserflow yamnet
2. speech detection model uses vosk
3. the passenger speech detection is bad because you have to yell.
4. tensorflow cannot be used on the raspberry pi. we must use tensorflow LITE

Branches

main - contains code that can run on your laptop

raspberry_pi - contains code that can run on the pi / mimic running on the pi

How to run:

the code is in the /src directory*

clone this github repo on your local desktop

make sure you have python 3 installed. Run this to check

python --version

create a virtual environment

python3 -m venv venv

NOTE: Python Tkinter must be installed for the UI -> brew install python-tk@3.12 if using tkinter: /opt/homebrew/opt/python@3.12/bin/python3.12 -m venv venv

enter the virtual python environment. do this by running:

source venv/bin/activate

install packages

pip install -r requirements.txt

run the file

python src/main.py

if you get an error that a package isnt installed, install it:

pip install some-package

Output

Will look something like

Auto Startup

to stop the program, run systemctl --user stop alertrider.service