Voice Controlled Smart Spaces Florian Gratzer Advisors: Marc-Oliver - - PowerPoint PPT Presentation

voice controlled smart spaces
SMART_READER_LITE
LIVE PREVIEW

Voice Controlled Smart Spaces Florian Gratzer Advisors: Marc-Oliver - - PowerPoint PPT Presentation

Chair of Network Architecures and Services Departments of Informatics Technical University of Munich Voice Controlled Smart Spaces Florian Gratzer Advisors: Marc-Oliver Pahl Stefan Liebald Garching, January 16 th 2017 Overview


slide-1
SLIDE 1 Chair of Network Architecures and Services Departments of Informatics Technical University of Munich

Florian Gratzer Advisors: Marc-Oliver Pahl Stefan Liebald Garching, January 16th 2017

Voice Controlled Smart Spaces

slide-2
SLIDE 2
  • Controlling the environment via

voice commands

  • VSL as middleware
  • Mapping of voice input to actions

within the VSL

  • Arbitrary devices
  • Arbitrary commands
  • Configurable mapping

2 Florian Gratzer | Voice Controlled Smart Spaces

Overview

[http://kreyosimages.s3.amazonaws.com/ks_voice_illustration.jpg, http://media.guitarcenter.com/is/image/MMGS7/V250-Condenser-Microphone/H77071000000000-00-500x500.jpg, http://diziusa.com/wp-content/uploads/2014/12/lampe.jpg, http://i.ebayimg.com/00/s/NjY3WDEwMDA=/z/aFsAAMXQyY1TQp3s/$_32.JPG]
slide-3
SLIDE 3
  • Short response time
  • Low error rate
  • Offline functionality
  • Voice output
  • Support of custom devices
  • Easy configuration

Key Requirements

Florian Gratzer | Voice Controlled Smart Spaces 3

slide-4
SLIDE 4
  • Framework for home automation
  • Available for iOS devices
  • Can be used by Siri
  • Custom built devices not supported

Apple HomeKit

Florian Gratzer | Voice Controlled Smart Spaces 4

[http://www.apple.com/ios/home/]
slide-5
SLIDE 5
  • Voice controllable speaker
  • Amazon Alexa for voice recognition
  • Custom built devices not supported

Amazon Echo

Florian Gratzer | Voice Controlled Smart Spaces 5

[http://www.giga.de/wp-content/uploads/2016/05/Amazon-Echo.jpg]
slide-6
SLIDE 6
  • Built for IoT contest
  • Voice control of custom built

devices

  • Pocketsphinx for voice

recognition

  • Flite for speech synthesis

Voice Controlled Alarm Clock

Florian Gratzer | Voice Controlled Smart Spaces 6

[https://www.element14.com/community/community/design-challenges/pi-iot/blog/2016/08/15/pi-iot-alarm-clock-16-wiring]
slide-7
SLIDE 7

Related work

Florian Gratzer | Voice Controlled Smart Spaces 7

HomeKit Echo Alarm Clock Offline functionality

  • +

Easy configuration

+ +

  • Custom devices

supported

  • +
slide-8
SLIDE 8

Design

Florian Gratzer | Voice Controlled Smart Spaces 8

slide-9
SLIDE 9
  • Human speech waveform contains a large amount of information
  • Dependent on
  • Speaker
  • Speaking rate
  • Acoustic conditions
  • Hardly possible to match samples directly
  • Multiple processing steps are required
  • Using “Features” for matching

9 Florian Gratzer | Voice Controlled Smart Spaces

Voice recognition

slide-10
SLIDE 10
  • Start and end time detection
  • Manual
  • (Semi-)automatic
  • Feature Extraction
  • Filtering
  • Windowing
  • Extracting features
  • Feature Matching
  • Mapping the recording to a sample

Processing steps

Florian Gratzer | Voice Controlled Smart Spaces 10

[B. Pfister – Sprachverarbeitung ISBN: 9788578110796]
slide-11
SLIDE 11
  • Evaluated
  • Used
  • CMU Sphinx
  • Training required
  • Kaldi
  • HTK
  • Designed as front end
  • Jasper
  • CMU Sphinx
  • Requires
  • Acoustic Model
  • Dictionary
  • Language Model

Speech to text engines

Florian Gratzer | Voice Controlled Smart Spaces 11

slide-12
SLIDE 12

Design

Florian Gratzer | Voice Controlled Smart Spaces 12

slide-13
SLIDE 13
  • 2 phases
  • Transcription phase
  • Phonoacoustic phase
  • Text to speech engines
  • MaryTTS
  • espeak

Speech Synthesis

Florian Gratzer | Voice Controlled Smart Spaces 13

slide-14
SLIDE 14

Design

Florian Gratzer | Voice Controlled Smart Spaces 14

slide-15
SLIDE 15
  • Multiple actions per voice command
  • Voice output support
  • GET and SET commands

Mapping Format

Florian Gratzer | Voice Controlled Smart Spaces 15

slide-16
SLIDE 16

Design

Florian Gratzer | Voice Controlled Smart Spaces 16

slide-17
SLIDE 17

Configration Interface

Florian Gratzer | Voice Controlled Smart Spaces 17

slide-18
SLIDE 18

Design

Florian Gratzer | Voice Controlled Smart Spaces 18

slide-19
SLIDE 19

Florian Gratzer | Voice Controlled Smart Spaces 19

Demo

slide-20
SLIDE 20
  • Scenario

 Light control  Custom dictionary

  • Error rate
  • Target error rate: < 5%
  • Actual error rate: 2%

Evaluation

Florian Gratzer | Voice Controlled Smart Spaces 20

slide-21
SLIDE 21
  • Response time:
  • Remote control as

reference

  • Assumption: RC in

reach, but not in hand

  • Target: < 3s

Evaluation

Florian Gratzer | Voice Controlled Smart Spaces 21

slide-22
SLIDE 22

Contributions

Florian Gratzer | Voice Controlled Smart Spaces 22

slide-23
SLIDE 23
  • Integrate the system in the lab room
  • Test other voice recognition engines
  • WhatsApp interface
  • Further evaluations of the system

Future work

Florian Gratzer | Voice Controlled Smart Spaces 23

[https://i0.wp.com/thegadgetox.net/wp-content/uploads/2016/02/whatsapp-logo-vector.png? fit=1150%2C1163&ssl=1]
slide-24
SLIDE 24

Thank you for your attention!

Florian Gratzer | Voice Controlled Smart Spaces 24