This repository contains the software for an automated cocktailmaker which recommends a drink to the user based on the users emotions and alcohol level using an Artificial Neural Network (ANN).
We're planning to use 8 bottles of liquor and drinks to create 10 different types of drink recipes.
The cocktailmaker consists of the following hardware:
- Raspberry Pi 4 Model B + LCD Touchscreen + Pi Cam
- Arduino Mega + various sensors
- Laptop (any model)
- 8 magnetic valves
The box casing is planned to be made of sheet metal and will hold the 8 liquor and drink bottles as well as the Arduino Mega, Raspberry Pi and magnetic valves.
- User is in front of the machine and is introduced to the procedure via the GUI.
- GUI leads user through every step in the process such as the 10 second emotion detection and 10 second alcohol measurement.
- Measurement data is sent to the ANN.
- ANN processes data and outputs three drink recommendations.
- Three drink recommendations get displayed on the GUI.
- User selects one of the three drinks.
- Routine starts to create the selected drink via control of the valves.
- ANN is trained with the selected drink.
- Program starts from the beginning.
Below you'll find the current architecture of the cocktailmaker and its dataflow.
