Eva-Assistant-Python

EVA is a social 🗣 and affective ❤️ robot aimed at assisting older adults in conducting Instrumental Activities of Daily Living (IADLs). It is not only a passive assistant, but an active one: proactive behaviour is incorporated in the robot. It can start conversations and show concern about the user, making the interaction more natural and affective.
This repo contains the brain 🧠 structure of the robot, the proactive and interaction behavior themselves.

Be different 😎

With EVA, you can have conversations in the more natural way. It is activated by a novel method called wakeface, in which the robot activates/listens the user by looking at it. Also, it is able to start conversations by using proactive questions.

Highlighted proactive questions ✨ how are you, who are you
Other hightlighted (cool) services reading incoming telegram messages, sending telegram messages

What are you waiting for to meet it? Construct your own EVA today! 🔝

Main components 🤖🛠️

EVA hardware structure is based on the existing EVA robotic platform. This EVA affective and proactive version is constructed using the following elements:

• 🖥️ Screen Waveshare 5.5" HDMI AMOLED for 👀 displaying
• 📷 Intel® RealSense™ Depth Camera D435i
• 🎤💡 Matrix Voice
• 🔊 Tempest M10 RGB Hero 2.0 Gaming Speakers
• Raspberry Pi 4 Model B
• Plastic structure printed using 3D printer

It is not mantadory to have exactly the same components, as long as they have the same functionalities. If you change any of them, make sure you have the correct drivers. In particular, if camera is changed, camera frames capturing code must be rewritten to use the suitable library for your camera management (it will probably not be pyrealsense2).

Installation ⚙️

Requirements

Use the package manager pip to install all the pre-requisites.

pip3 install -r requirements.txt

Google and IBM services

Ensure you have IBM Cloud and Google Cloud accounts.

It is necessary to have apikeys 🔑 for the following services:

• Google Speech to Text
• Google Text to Speech
• Google Translator
• IBM Natural Language Understanding
• IBM Watson Assistant

Both files (Google credentials and IBM credentials) must be stored in a credentials/ directory, located outside the main project directory:

$ any_directory
.
├── credentials
│   ├── google_credentials.json
│   ├── assistant_credentials.json
│   └── ibm_credentials.env
└── Eva-Assistant-Python/

In this case, by-default environment variables provided by Google and IBM (GOOGLE_APPLICATION_CREDENTIALS and IBM_CREDENTIALS_FILE) are used for automatic services apikey authentication (easier!). Also, a WATSON_ASSISTANT_CREDENTIALS variable containing assistant_credentials.json path is used for obtaining the assistant id.

WARNING ⚠️: IBM Watson Assistant dialog tree can be freely designed by the user. This makes more flexible and customized the user interaction with your own robot 😉

Telegram service 💬

Follow these steps:

1. Get your own api_id and api_hash from Telegram (click here and follow the instructions), under API Development.
2. Store them in a telegram_credentials.json file. This file must be located in credentials/ directory, and should be like this:

{
  "api_id" : 12345,
  "api_hash" : "0123456789abcdef0123456789abcdef"
}

3. Create your telegram.session file, located also in credentials/ directory. Using Telethon library is very easy.
4. Create two environment variables TELEGRAM_CREDENTIALS and TELEGRAM_SESSION which contain their corresponding paths.
5. Enjoy the service!

Usage 🚀

For executing EVA, you only have to run from root repo directory:

python3 main.py

➡️Note: proactive_phrases.json contains phrases totally in spanish, so if you want EVA to speak in a different language, teach it your language by changing this file and developing the IBM Watson Assistant dialog tree in your language. It will be happy to learn it 😊

And that's how you construct your own affective robot! 🤖❤️👩🏻

Authors 📝

• Laura Villa 🦁