Speech Recognition for User Interface robotics

1
Speech Recognition for User Interface robotics

• Oratile Leteane
• 203024214
• Supervisor
• Mr Lester Cowley
• 2007

2
(No Transcript)
3
Contents

• Background
• Problem description
• Related work
• Analysis of the new system
• Discussion of tools
• Risks of the project
• Project plan
• Progress report
• Conclusion

4
Background

• Increasing number of robots
• A means of communication need to be established
• Natural language, a desirable communication mode
• Researchers need to find a way.

5
Background (cont.)

• Good thing about studying robots
• iCat research platform is available
• Platform for studying robotic
• User interface robot iCat
• OPPR (Open platform for personal robot)
• Website for supporting iCat research community

6
Background (cont.)

• iCat

7
Background (cont.)

• Different expressions for iCat

Anger
Sad
Happiness
8
Background (cont.)

• Main objectives is to give icat robot
  capabilities
• Ability to recognise constraint speech
• Able to give relevant responce through
• Speech
• Facial expressions

9
Problem description

• Relevance of the project
• Research foundation
• Problem statement
• Providing iCat with a facility that will enable
  it to recognise constraint human speech

10
Problem description (cont.)

• Keywords
• Dragon Naturally speaking 9 (for developers)
• OpenCCG (Natural source language parsing)
• Scope
• Understanding iCat current capabilities
• Using current capabilities, DNS-9 and OpenCCG to
  implement application to enable iCat to recognise
  human constraint speech.

11
Problem description (Cont.)

• Limitation of research
• Follow certain number of commands
• Objects detection and visual recognition not part
  of research

12
Related work

• Literature study
• iCat community website for supporting the iCat
  research community

13
Related work (cont.)

• Existing systems
• Kate (2006) Exploring Social Interaction with
  user interface robots using iCat
• Interst To provide some insight into social
  interaction between human and robots

14
Related work (cont.)

• Human-Robot Dialogue for Joint Construction tasks
• iCat is envolved on construction task
• Robot and human are equal partners

15
Related work (cont.)

• Robust Speech Recognition System for
  Communication Robots in Real Enviroments
• Aim to develop communication robot that can
  naturally interact with humans (high level dialog
  communication)

16
Analysis of new system

• Functional requirements
• Recognise a certain number of specified commands
  
• Respond to the given commands
• Able to do arithmatic calculations

17
Analysis of new system (cont.)

• Non-functional requirements
• Easy of use
• Easy to learn

18
Analysis of new System (cont.)

• Discussion of tools
• OPPR
• Reasons to use
• Contain robot animation editor
• Contain ready to use library with animations for
  speech

19
Analysis of new System (cont.)

• Discussion of tools
• LUA programming
• Dragon Naturally Speaking 9 (DNS-9)
• for Speech Recognition
• OpenCCG library
• Parsing of recognised speech to logical form

20
Project risks

• New project taken at NMMU in the speech domain
• Learning Language (LUA) new environment (OPPR)
• Possibility of iCat being damage

21
Risk of project

• Addressing risks areas
• Being part of iCat community group
• Start learning LUA programming as early as
  possible
• Ensuring that iCat is kept in one place
• Reduce chances of breaking it

22
Project plan
REQUIREMENTS
ANALYSIS DESIGN
IMPLEMENTATION
EVALUATION
14/02 TO 23/02
24/03 TO 14/05
15/05 TO 03/09
04/09 TO 01/09
23
Progress report
24
Conclusion

• It is a very challenging work
• Still the beginning, more to come

25
Questions?