Pace Voice XML Lab

1
Pace Voice XML Lab

• Dr Tappert
• Shreenath Laxman and Darshan Desai

2
Vision Statement

• Voice Browsers offer the promise of allowing
  everyone to access Web based services from any
  phone, making it practical to access the Web any
  time and any where, whether at home, on the move,
  or at work.
• Thus Application development has to be designed
  not only with web as a focus but should be able
  to span all kinds of devices.
• Also Multilingual access to data will enable
  reaching to more number of people

3

• Targeted audience
• Students
• Work towards developing multi-modal
  applications.
• Faculty
• Academic application development
  example grade
• information etc.
• Administrative staff
• Application such as on phone registration,
  course changes , class schedules etc.

4
Architecture
5
What is Voice XML?

• HTML is a markup language that facilitates data
  to be visually displayed.
• Voice XML is a mark up language that will enable
  Data to be heard

6
Comparison of the Mark Up Languages
7
Grade Information System

• http//aom2.pace.edu/vxml/html/gradeWeb.asp
• 1-800-555-VXML

8
Work Flow of Grade Information System
9
What will Pace Lab Facilitate

• It will provide with the gateway to develop voice
  application
• It will provide interface for developing
  applications template library
• It will enable novices to use the code library to
  learn and build advanced applications
• Facilitate everyone to deploy applications on
  multi platform

10
Architecture
11
Hardware Components

• Cisco router Voice Over IP
• IBM Voice Server
• Local Area Network
• PSTN
• Firewall (optional for Voice Server)

12
Architecture
13
Studio Architecture

• Registration
• Development
• Deployment
• The Studio essentially enables the user to load
  his application URL which can be referenced by
  the Voice Server.
• This URL is retrieved and loaded when the call
  is routed to the IBM Voice Server.

14
Studio Architecture
15
Voice XML Language Components

• ltform id"login"gt
• ltfield name"userid"gt
• ltgrammar type"application/x-gsl"gt
• lt!CDATA
• (Char_speakd1 Dig_speakd2 Dig_speakd3
  Dig_speakd4 Dig_speakd5) ltoption strcat(d1
  strcat(d2 strcat(d3 strcat(d4 d5))))gt
• (cancel) ltoption "cancel"gt
• gt
• lt/grammargt
• ltpromptgtPlease Enter your Login IDlt/promptgt
• lt/fieldgt
• ltfilledgt
• ltassign name"application.userid"
  expr"userid"/gt
• ltsubmit nextGOTO_PAGE" namelistGOTO_PARAM
  ETERS"/gt
• lt/filledgt
• lt/formgt

16
Voice XML Components

• Framework contains 6 modules
• Resources - resource bundles, config files
• Prompts - pre- recorded wave files, text for
  prompts
• Grammars - grammars files
• Templates - dialog designs
• Includes - sources code, common
• Common - XML DTD, DB abstract, generic codes

17
Template Structure

• ltform id"login"gt
• ltfield name"userid"gt
• USERID_GRAMMAR
• USERID_PROMPT
• lt/fieldgt
• ltfilledgt
• ltassign name"application.userid"
  expr"userid"/gt
• ltsubmit nextGOTO_PAGE" namelistGOTO_PARAM
  ETERS"/gt
• lt/filledgt
• lt/formgt

18
Template FrameWork
Example of a prompt file for US_EN

• lt?php
• USERID_PROMPT'ltpromptgtltaudiogtPlease say your
  six digit user ID.lt/audiogtlt/promptgt'
• USERPWD_PROMPT'ltpromptgtltaudiogtSay your four
  digit password.lt/audiogtlt/promptgt'
• MENU_USERID_INVALID_PROMPT'ltpromptgtltaudiogtTha
  t user ID does not exist. Login
  again.lt/audiogtlt/promptgt'
• MENU_USERPWD_INVALID_PROMPT'ltpromptgtltaudiogtTh
  at password is incorrect. Login
  again.lt/audiogtlt/promptgt'
• ?gt

19
Template Framework

• lt?php
• USERID_GRAMMAR'srcbuiltingrammar/digits?len
  gthUSERID_LENGTH"gtlt/grammargt
• USERPWD_GRAMMARsrcbuiltingrammar/digits?lengt
  hUSERPWD_LENGTH"gtlt/grammargt
• ?gt

20
Progress

• Great Support from Pace CSIS Dept
• IBM Voice Server SDK
• Cisco Router

21
Conclusion

• Work towards building Advanced Applications
• Research in the field of Multi-Modal Architecture
• Research in Speech Recognition

22
Questions?

• desaidarshan_at_hotmail.com
• shreenathl_at_usa.net
• ctappert_at_pace.edu