Speech Subsystem Concepts

Interpreter

Takes in spoken text and performs appropriate actions / API calls

Grammar

Set of valid words and phrases that the recognition engine is expecting at a particular point of time

Scenario: API Requirements

Mechanic Bob has a workorder with stickies. A view of the model is displayed on his PEDD.

Bob says, “show IETM” in order to change the display from the model to the IETM.

speech.interpreter.GlobalInterpreter.interpret()

ui.UserInterfaceManager.displayUserInterface

(int userInterfaceID,

int displayStatus,

int panelID)

Scenario (cont.)

Bob says, “activate IETM” in order to switch grammars to the IETM grammar.

speech.interpreter.GlobalInterpreter.interpret()

speech.grammar.CentralGrammarHandler.

setActiveGrammarHandler(IETMGrammarHandler)

Bob says, “next step” and the speech system speaks the 1^st step to him.

speech.CentralInterpreter.ActiveInterpreter. interpret()

speech.IETMInterpreter.interpret()

speech.Synthesis.speakStep()

Scenario (cont.)

Bob performs the step spoken to him and continues saying “next step” until he is done with the repair.

Bob says, “activate workorder” in order to switch the active grammar to the workorder grammar.

speech.interpreter.CentralInterpreter. GlobalInterpreter.interpret()

speech.grammar.CentralGrammarHandler.

setActiveGrammarHandler(WorkorderGrammarHandler)

Scenario (cont.)

Bob says “submit” in order to automatically set the status to awaiting approval and submit the completed workorder to the database.

speech.CentralInterpreter.ActiveInterpreter.interpret()

speech.WorkorderInterpreter.interpret()

maintenance.submitWorkorder()

Scenario: Sequence Diagram

speech Package

speech.interpreter Package

speech.grammar Package

Javadoc

Javadoc:
http://www.contrib.andrew.cmu.edu/~rig/javadoc/index.html

Hardware/Software Allocation

The Speech subsystem will use IBM ViaVoice with JSAPI extensions as its speech engine.

ViaVoice needs a maximum 260 MB in disk space.

ViaVoice needs a maximum 48 MB RAM (including OS)

Services provided by Speech

void speech.Initialization.initialize()

Used by the UI subsystem to start the speech subsystem

The speech subsystem will be running as long as the STARS executable is running on the PEDD

Services needed by Speech

From Maintenance subsystem:

submit a workorder to the DB

void submitWorkorder()

advance the current sticky

void nextSticky()

set the current sticky back one

void previousSticky()

set the status of a workorder

void setStatus(int status)

Services needed by Speech (cont.)

From UI subsystem:

Switch displayed screen from model to IETM and vice versa

boolean displayUserInterface(int userInterfaceID, int displayStatus, int panelID)

Status of the Speech Subsystem

Requirements Analysis done

System Design done

Object Design done

Working on implementation

Looking into the API for WebManual IETM viewer

Formalizing all grammars: global, workorder, model, IETM

Future Schedule

Implementation

11/22 to 12/5

Unit Testing

12/1 to 12/5

Integration Testing

12/4 to 12/9

Dry Run

12/7

Client Acceptance Test

12/9

Open Issues

WebManual API à our creation of a dynamic IETM grammar depends on this


	Interpreter
		Takes in spoken text and performs appropriate actions / API calls
	Grammar
		Set of valid words and phrases that the recognition engine is expecting at a particular point of time


Mechanic Bob has a workorder with stickies. A view of the model is displayed on his PEDD.
Bob says, “show IETM” in order to change the display from the model to the IETM.
	speech.interpreter.GlobalInterpreter.interpret()
	ui.UserInterfaceManager.displayUserInterface
		(int userInterfaceID,
		int displayStatus,
		int panelID)


	Bob says, “activate IETM” in order to switch grammars to the IETM grammar.
		speech.interpreter.GlobalInterpreter.interpret()
		speech.grammar.CentralGrammarHandler.
		setActiveGrammarHandler(IETMGrammarHandler)
	Bob says, “next step” and the speech system speaks the 1^st step to him.
		speech.CentralInterpreter.ActiveInterpreter. interpret()
		speech.IETMInterpreter.interpret()
		speech.Synthesis.speakStep()


	Bob performs the step spoken to him and continues saying “next step” until he is done with the repair.
	Bob says, “activate workorder” in order to switch the active grammar to the workorder grammar.
		speech.interpreter.CentralInterpreter. GlobalInterpreter.interpret()
		speech.grammar.CentralGrammarHandler.
		setActiveGrammarHandler(WorkorderGrammarHandler)


	Bob says “submit” in order to automatically set the status to awaiting approval and submit the completed workorder to the database.
		speech.CentralInterpreter.ActiveInterpreter.interpret()
		speech.WorkorderInterpreter.interpret()
		maintenance.submitWorkorder()


	Javadoc: http://www.contrib.andrew.cmu.edu/~rig/javadoc/index.html


	The Speech subsystem will use IBM ViaVoice with JSAPI extensions as its speech engine.
	ViaVoice needs a maximum 260 MB in disk space.
	ViaVoice needs a maximum 48 MB RAM (including OS)


	void speech.Initialization.initialize()
		Used by the UI subsystem to start the speech subsystem
		The speech subsystem will be running as long as the STARS executable is running on the PEDD


	From Maintenance subsystem:
	submit a workorder to the DB
		void submitWorkorder()
	advance the current sticky
		void nextSticky()
	set the current sticky back one
		void previousSticky()
	set the status of a workorder
		void setStatus(int status)


	From UI subsystem:
	Switch displayed screen from model to IETM and vice versa
		boolean displayUserInterface(int userInterfaceID, int displayStatus, int panelID)


	Requirements Analysis done
	System Design done
	Object Design done
	Working on implementation
		Looking into the API for WebManual IETM viewer
		Formalizing all grammars: global, workorder, model, IETM


	Implementation
		11/22 to 12/5
	Unit Testing
		12/1 to 12/5
	Integration Testing
		12/4 to 12/9
	Dry Run
		12/7
	Client Acceptance Test
		12/9


	WebManual API à our creation of a dynamic IETM grammar depends on this