Susi Linux

SUSI AI on Linux

This project aims at creating an implementation of Susi, capable to run on Linux Devices in a headless mode. It will enable you to bring Susi AI intelligence to all devices you may think like a Speaker, Car etc.

Important: Tests Before Releasing New Images

1. The SUSI Server must be building on the pi on bootup
2. The hotword detection should have a decent accuracy
3. SUSI Linux shouldn't crash when switching from online to offline and vice versa (failing as of now)
4. SUSI Linux should be able to boot offline when no internet connection available (failing as of now)

Mobile App (Android)

This app is for configuration. Download Here

Smart speaker assembly tutorial

How to assembly a smart speaker: Video

Current Status

• Voice Detection working with Google Speech API/ IBM Watson Speech to Text API.
• Voice Output working with Google TTS/ IBM Watson TTS/ Flite TTS.
• Susi AI response working through Susi AI API Python Wrapper
• Hotword Detection works for hotword "Susi"
• SUSI Wake Button in Raspberry Pi is working.
• Audio parsing working through SOX
• Youtube audio working MPV player
• Connection between the mobile devices and the smart speaker is done through ssh by using Access Point Mode

Roadmap

• Offline Voice Detection (if possible with satisfactory results)
• Make Update Daemon check for the updates at regular intervals

General working of SUSI

• SUSI.AI follows a finite state system for the code architecture.
• Multiple daemons are created during installation (Media Daemon, Factory Daemon, Update Daemon)
  • The Update Daemon updates the repo every time the Raspberry Pi is restarted up
  • The Media Daemon detects for USB connection and scans the songs present in it to play from Smart Speaker
  • The factory reset Daemon is a button on Port 17 which delete the repo completely and restores it from the backup
• A local Server is deployed every time the Rasperry Pi starts and automatically switches to online server if there are any issues with the local server
• Google TTS and STT services are used as default services but if the internet fails, a switch to offline services PocketSphinx (STT) and Flite (TTS) is made automatically

Setting up Susi on Linux

Setting up Susi on Linux is pretty easy.

Note:

For the app to work properly, repo must be cloned inside the folder /home/pi/SUSI.AI/ , i.e. the path of your repo will look like /home/pi/SUSI.AI/susi_linux/

Minimum Requirements

• A hardware device capable to run Linux. Currently on Raspberry Pi 3 is supported. Other embedded computers, like BeagleBone Black, Orange Pi, will be supported in the future.
• A Debian based Linux Distribution. Tested on
  • Raspbian on Raspberry Pi 3
  • Ubuntu 64bit on x64 architecture
• A microphone for input. Currently the development team is using ReSpeaker 2-Mics Pi HAT for Raspberry Pi.
• A speaker for output. On development boards like Raspberry Pi, you can use a portable speaker that connects through 3.5mm audio jack. If you are using ReSpeaker 2-Mics Pi HAT, the speaker should be plugged to this board.

Installation on Raspberry Pi

For installation on Raspberry Pi, read Raspberry Pi setup guide.

Configuring a connection through SSH

Step 1: Initial Setup

• Both the raspberry Pi with raspbian installed and the mobile device should be on a same wireless network
• One should have an SSH viewer like JuiceSSH(Android) and iTerminal(IOS) installed on their mobile devices
• Now we must enable SSH on our raspberry Pi

Step 2: Enabling SSH on Raspberry Pi

• To enable SSH on your Pi , follow the following steps:
  Menu > Preferences > Raspberry Pi Configuration.
  Choose the interfaces tab and enable SSH Step 3:Setting Up the client
• Login to your raspberry pi as the root user (pi by default)
• Type the following command to know the broadcasting ip address
  pi@raspberrypi:hostname -I
• Now , open the client on your mobile device and add the configurations
  

By default the username of the system is ‘pi’ and the password is ‘raspberry’

Configuring SUSI.AI through the Android App

• After Running the installation script , you'll have a RasPi in access point mode. With a Flask Server running at port 5000.
• You can use the mobile clients to configure the device automatically.
  

Configuring SUSI.AI Manually (in case of any failures in the mobile clients)

• First, you have to configure the room name by hitting the url : localhost:5000/speaker_config?room_name=roomName
  • Where roomName is the name of the room in which the speaker has too be placed
• Second, to configure the wifi credentials, use the URL : localhost:5000/wifi_credentials?wifissid=WIFISSID&wifipassd=WIFIPASSD
  • Where WIFISSID is the name of your Wifi SSID and Wifi Password
• Third, to configure your SUSI credentials use the URL :localhost:5000/auth?auth=y&email=EMAIL&password=PASSWORD.
  • Where EMAIL and PASSWORD are your login credentials of SUSI.AI
• Finally use the URL : localhost:5000/config?stt=google&tts=google&hotword=y&wake=n

• Your RaspberryPi should restart automatically, if not execute the script rwap.sh in the access_point folder.

• To login in the speaker , start the configuration script by
  python3 config_generator.py <stt> <tts> <hotword_detection> <wake_button> Where

  • stt is the speech to text service
    • You have the following choices
    • 'google' if you want to use google as the stt service
    • 'ibm' if you want to use ibm as the stt service
    • 'sphinx'(offline) if you want to use pocket-sphinx as the stt service
  • tts is the text to speech service
    • Similarly,
    • You have the following choices -'google' if you want to use google as the tts service
    • 'ibm' if you want to use ibm as the tts service
    • 'flite'(offline) if you want to use flite as the tts service
  • hotword_detection is the choice if you want to use snowboy detector as the hotword detection or not
    • 'y' to use snowboy
    • 'n' to use pocket sphinx
  • wake_button is the choice if you want to use an external wake button or not
    • 'y' to use an external wake button
    • 'n' to disable the external wake button

• Eg. To google as the default stt and tts and using snowboy as the default hotword detection engine and no wake button , you have to use the following command python3 config_generator.py google google y n

Using SUSI in a Authenticated Mode

* To use SUSI in an authenticated mode, you have to use the script 'authentication.py'
`python3 authentication.py ` Where - authentication_choice is the choice if you want to use SUSI in an authenticated mode or not - email is your registered email id with SUSI.AI - password is your registered password of the corresponding email registered with SUSI.AI * Eg. to use SUSI.AI in authenticated mode use `python3 authentication.py y [email protected] password`

Installing on Ubuntu and other Debian based distributions

For installation on Ubuntu and other Debian based distributions, read Ubuntu Setup Guide

Update Daemon

At any point of time, we may want to check if the current version of susi linux is updated. Hence we compare against the corresponding remote repository and we update it accordingly every time the raspberry Pi has started. Use the following commands.

• cd update_daemon/
• ./update_check.sh

Factory Reset

To initiate the factory reset command.
Use the following commands.

• cd factory_reset/
• chmod +x factory_reset.sh
• ./factory_reset.sh

Run SUSI Linux

This section is intended for developer.

SUSI Linux application is run automatically by systemd. The main component is run as ss-susi-linux.service. By default, it is ran in production mode, where log messages are limited to error and warning only. In development, you may want to see more logs, to help debugging. You can switch it to "verbose" mode by 2 ways:

1. Run it manually

• Stop systemd service by sudo systemctl stop ss-susi-linux
• Use Terminal, cd to susi_linux directory and run

python3 -m main -v

or repeat v to increase verbosity:

python3 -m main -vv

2. Change command run by systemd

• Edit the /lib/systemd/system/ss-susi-linux.service and change the command in ExecStart parameter:

ExecStart=/usr/bin/python3 -m main -v --short-log

• Reload systemd daemon: sudo systemctl daemon-reload

• Restart the servive: sudo systemctl restart ss-susi-linux

• Now you can read the log via journalctl:

  • journalctl -u ss-susi-linux
  • or journalctl -fu ss-susi-linux to get updated when the log is continuously produced.

The -v option is actually the same as the 1st method. The --short-log option is to exclude some info which is already provided by journalctl. For more info about logging feature, see this GitHub issue.

Setting Up the access point mode

To allow the raspberry Pi to behave as an access point

• Execute the wap.sh script by ./wap.sh
• To convert RasPi back to normal mode use ./rwap.sh

SUSI Smart Speaker - IOS/Android Workflow

To Use the ReSpeaker 2 Mic Array as a default Audio Driver

• User can use pacmd to change the audio card to piHat

• Use the following commands
  • pacmd list-sinks to check the index of the device
  • pacmd set-sink-port <sink name> <port name> '' is generally 1 , but you can choose depending on your list.
  • eg. if you want to use inbuilt speaker ports pacmd set-sink-port alsa_output.platform-soc_sound.analog-stereo analog-output-speaker
  • if you want to use your headphones with it , we use pacmd set-sink-port alsa_output.platform-soc_sound.analog-stereo analog-output-headphones
    
    

If the above approach doesn't work , you can use the following approach

• After running the installation script

• Type this command cd /etc/pulse
• Open the file default.pa
• Replace line 38 by load-module module-alsa-sink device=hw:2,1 (This will disable the default soundacards from loading up)
• To enable default sound cards(usb mic,built-in headphone jack,etc ) , comment/delete out the above line