Mixologist

Cocktail machine project for Ambient Intelligence class @ Instituto Superior Técnico, Lisbon, Portugal.

Report: Mixologist_AI_Project_Report_Group10.pdf

More information: https://fenix.tecnico.ulisboa.pt/disciplinas/AI514/2019-2020/2-semestre

Group 10

Features

• Cup filling up to a specified weight/liquid quantity
• Beverage pouring only if a cup is present and if the cup has the needed capacity
• Cocktail making only if the quantities of the required ingredients are available
• Alerts on empty bottles / full glass
• Storage, management of recipes in a database
• Google Assistant voice control (recipe search)
• Notifications on status update
• Computation of the number of cocktails that it is possible to make with the current bottles

Enhancements

• Add authentication
• Add ingredients / recipes editing

Specifications

This application is written in Python 3, built around the Flask framework for routing.
Data is stored in a MySQL database and can be updated using HTTP REST methods.

Database specification

Prerequisites

Hardware

• Orange Pi PC or other GPIO-capable board (preferably supporting the pyA20 library to make it work right away)
• 4 (peristaltic) pumps
• 4 relays module to drive the pumps
• An Arduino
• 5 load cells modules
• 5 HX711 ADC modules

Software

• Arduino IDE
• Python 3
• Python dependencies (flask, pymysql, flask-socketio, pyA20, spidev, pyserial, requests, pyopenssl - can be installed with the Pipfile, please read How to build)
• MySQL/MariaDB server
• IFTTT account

How to build

Wiring / Hardware

Default pin connections are the following:

Relay board

Relay input	Orange Pi PC pin
1	PA10
2	PA9
3	PA8
4	PA7

HX711 modules

HX711 #	HX711 pin	Arduino pin
1	DOUT SCK	4 5
2	DOUT SCK	6 7
3	DOUT SCK	8 9
4	DOUT SCK	10 11
5	DOUT SCK	12 13

Software

Arduino

Upload the arduino/calibration/calibration.ino sketch to your Arduino and note down the calibration value for each load cell in the provided Serial assistant.
Edit the arduino/json_read_5x/json_read_5x.ino sketch:

float calibrationValue_b1 = -420.00; // calibration value load cell 1          !!!
float calibrationValue_b2 = -418.53; // calibration value load cell 2
float calibrationValue_g = -951.84;  // calibration value load cell 3
float calibrationValue_b3 = -420.80; // calibration value load cell 4
float calibrationValue_b4 = -416.38; // calibration value load cell 5

And update the values for each load cell.
Upload this sketch to your Arduino and connect it to your Orange Pi.

IFTTT (enable Google Assistant and notifications)

Create an IFTTT account on https://ifttt.com/, then enable the Google Assistant service and the Webhooks service.
Once done, note down your API key, and create two applets:

MySQL database

Load the init.sql file to populate your database and note down your connection settings.

Orange Pi

Replace database, Arduino, WS2812 and IFTTT connection settings in webapp/config.ini.
If needed, replace the relay GPIO pins in webapp/Pump.py.

sudo apt install python3 python3-pip python3-dev python3-setuptools mariadb-server mariadb-client libffi-dev
pip install --user pipenv
git clone https://github.com/geckoflume/mixologist.git
cd mixologist
pipenv install
python3 webapp/app.py

The webserver is now started on https://ORANGE_PI_IP:5000/.