Mobile Email Speech System (M.E.S.S.) Plan

Home
My Life
Resume
Transcript
Documents
Galleries
Code
E-Books
Resources
Greeting Cards
My University
IEEE
Pakistan Pics

1.0 Introduction

The Mobile E-mail Speech System (M.E.S.S) is a stand alone e-mail system which can receive e-mails from an Internet Service Provider and translates the ASCII text into speech.  This system uses cellular technology to retrieve the e-mails.  The M.E.S.S. is a voice activated system.  This system has been designed to be implemented in a vehicle.  It is primarily geared for the person who is constantly on the go.   It will be a quick and effective way of accessing information while travelling, commuting,  or being away from the office.  It will be a safer alternative as opposed to using a laptop and driving at the same time.  This system can also be used for a different purpose of aiding the blind.  One of the unique characteristics of our project is the text to speech feature which converts text e-mails into speech to be heard by its users.  Thus, it can be a standalone e-mail retrieval system in a residence.

2.0 Requirements

The M.E.S.S. will be able to download e-mails from a mobile location.  The system will then parse the message to remove unwanted portions such as headers and attachments.  The M.E.S.S. will then dictate the message out to the user.  The user will interface with the system using voice commands.                

The physical layout of the M.E.S.S. will consist of two main components.  The user interface will be implanted in the dashboard of the vehicle, while the processing unit will be located in a larger area such as the trunk or glove compartment. In greater detail the requirements are:

Connection / Download Requirements

• The ability to choose from various different sites to download the e-mails. The ability to download messages from a remote site using cellular technology along with normal computer to computer communication as per normal dial-up connectivity.

• The system will download as many e-mails as possible and store them into memory for user access.

E-mail Parsing

• The ability to parse the e-mail messages to remove all unwanted headers as well as attachments in order to avoid unnecessary unintelligible speech from being outputted.

• The ability to have the message ready for the RC Systems Double-Talk Text-to-Speech converter.

User Interface Requirements

• The user will have the ability to repeat a message, skip a message or listen to a previous message.

• The ability to read the e-mail message to the user in a comprehensible voice.

• The user interface must be large enough to allow the user to easily see it.  It must also be well lit so that it can be used in all lighting conditions.

• The system must be able to transmit the output of the system through the stereo system of the vehicle.  This will ensure that unnecessary  speaker components will not be required.

• The ability to recognize four (4) different users by their voices.  This will be achieved by having the waveforms of the users saved in the memory of the system.  The system will then attempt to correlate between the inputted voice and the saved waveforms.

• To allow for hands-free operation using voice commands.  This will be achieved by using a microphone and speech recognition software.

 Desirable Features

• Although two-way e-mail communications through speech is desirable, we have decided to concentrate on just the one-way communication system for this present project.  As time permits we will attempt to make this part of the project possible.

3.0 Preliminary Design

Our M.E.S.S. system will be comprised of a number of various components.  These will include:  

1. RC Systems DoubleTalk (text-to-voice converter-Hardware)

2. Embedded-PC computer system

3. Vehicle sound system and connection module

4. User Control Interface

5. Cell phone

6. FM Transmitter

3.1 Modules

Our M.E.S.S. system will be made up of three modules:

1. A User Interface

2. A main computer board

3. An ASCII Text-to-Speech Converter

3.1.1 User Interface

This module will act as the conduit for human to machine interaction. The user will be able to interact with the system by using voice commands, and a keypad. The interface will communicate back to the user through an LCD display and speech through a vehicles speaker system.                              

3.1.2 Setting up Users

First the user interface must be set up for all of the required users. (Maximum 4) The interface setup should be configured while the car is stationary for safety reasons. A new system will ask for the following questions to get the system configured properly. 

The system will ask a new user to be created. The users will be assigned the number in the order they are entered into the system. They will be assigned a five number PIN for a backup log-in password.

Next, each user will be asked to give a sample of their voice saying a certain phrase. This sample will then be recorded in the embedded PC as a WAV file. These samples will be used for security. The users must say the corresponding phrase to log in. If a non-user tries to log-on, then the PC will compare their voice to the samples in memory and deny them access. The user can also use their five number PIN to log in.

The next step in user setup is to set up the Internet Service Provider. The system will ask the user for the number of ISP's (Maximum 3). It will then ask for the ISP's phone number and the user's username and password. Then it will proceed to get information on the next ISP if required.

Next, the voice commands recognition setup will take place. This will be done using software. This requires the users to read text into the PC so it can understand their voice. This will allow for you to speak commands to the PC to control the system.

3.1.3  Deleting or Modifying Users or ISP's

User's can go into the setup and delete themselves or one of their ISP's. The user must first be logged into the system as themselves.

3.1.4 Logging In

Once the system verifies that the person is an appropriate user, it will bring up the user's ISP list. The user will then be able to say the corresponding ISP number or press the button next to the ISP and will be automatically logged into the appropriate ISP.

3.1.5 Retrieving Mail

After the user is logged in, as many new messages as our PC can handle will be down loaded. The graphical interface will then display all the message's information on the screen. The message at the very top will be highlighted. To have it read out the user will say the following command "Read". If the user wants to select a different message they can scroll down to the other messages by using the voice command "Scroll Up" or "Scroll Down" The system will only copy the e-mail from the ISP. So when the system is turned off all of the messages will be lost on the PC. And the e-mails will also remain on the ISP server.

3.2 The Embedded-PC/modem Unit

This module will create, delete, modify and store user information, recognize users through voice, connect via modem and cell-phone to an ISP as specified and download e-mail. It will then parse the message to remove unwanted headers and attachments.  The e-mails will then be converted into ASCII text and are sent to the text-to-speech converter. The functions above will be performed using a C programmed interface. It will also allow for speech recognition commands to download and hear the messages. Speech recognition commands will be performed using software.

3.3 The RC Systems DoubleTalk ASCII Text-to-Speech Processor

This module will convert the ASCII formatted text into speech and parlay the message to the user, through the vehicle's sound system. The program will execute so that the e-mail's text is sent to the Text to Speech processor. The speech will be transmitted to the sound system using a FM transmitter.

3.4 Cellular Connection

The cellular connection will be achieved using a Motorola DPC 650 cell phone, which will be connected to a modem on the laptop. There will be a cellular phone adapter that will connect to the modem on the laptop.  There will also be a modem in the server.  The connection that is used will be a dial-up connection.

3.5 FM Transmitter

To connect the M.E.S.S. to the vehicle audio system we will implement a frequency modulation transmitter.  This device will transmit the output from the DoubleTalk processor to the FM receiver in the vehicle.  The FM receiver will be tuned to the appropriate frequency in order to receive the transmission.  The transmission will be fairly low in power to avoid being received by other vehicles.

4.0 Project Implementation

Roles:

Project Manager:      Jim Cheung

Project Designer:      Ahsan Upal  (Research/Testing/Text)

                                     Jim Cheung  (Software)

Dimitrios Fourlas (Server Setup)

Kanaris Paraskevopoulos  (Software)

Kristen Sparrowhawk  (Software/Hardware)

Implementation of Project: All members

4.1 How we will work on project

Our group meets tri-weekly and discusses the project and how it is going.  We divide all of the work and assign it to individual team members to work on or research.  By dividing up the workloads we hope to get more accomplished and have a better-rounded, technically sophisticated project.

4.2 Basic Assumptions

Supplies: We own several PC components (motherboards, hard drives, ram etc.) which we can use in creating the two test systems for this project.  One machine will be the mobile PC, which will be a Pentium IBM laptop with the text to speech converter interfaced to it, as well as the cell phone.  This is the unit that we are modeling as being the unit in the vehicle.  The other system will be the e-mail server.   We expect to build the M.E.S.S. system with these components as well as others that we will purchase such as cables and the like.  We will be implementing most of the project in the form of software, which will involve scripts to interface to the cell phone and the text-to-speech converter.  To implement the vehicular unit as a final product, we are making the assumption that purchasing equipment such as a dash display and keypad interface, and the embedded PC would cost up to one thousand dollars ($1000).  We have come to this estimate from the research that we have done on the hardware that we would need to implement for this system.  Obviously we have not completed the design so this is strictly an estimate that we will try to maintain.

4.3 Equipment

Our team plans on using the test equipment that is available at the University of Regina Engineering department for the majority of our testing and troubleshooting.  As well, we have a variety of textbooks, and help books for the coding of the software and development software.  The engineering labs offer the equipment that we need for all of our testing. 

4.4 Advisors

Our main project advisor is Dr Stephen O’Leary.  He is assisting us on both the technical and non-technical aspects of our project.  We also have access to talking to any of the other professors, associate professors, and lab instructors of the Engineering Department of the University of Regina.  We feel that between all of these sources we have an excellent supply of technical support that we may need if we run into problems or have any questions.

Prototype

We will be creating a prototype of the project as a test setup with the two PC’s and their associated hardware for one user and one ISP.  In principle, if we can create a working project in this test setting, then we believe implementing it using an embedded-PC and special display and keypad in an actual vehicle would be relatively easy.  Our main project restriction is the financial aspect, therefore we will have complete design schematics for the actual implementation in a vehicle, however our prototype will be lab based. 

Methods of Communication

Our team does the majority of our communicating via telephone and e-mail, especially since most of us are employed as well.  We also use e-mail for the purpose of sending documents.  We keep minutes of our meetings and at the end of every meeting that our team holds we schedule the next meeting.  We keep in constant communication so that we can keep our project on schedule and running smoothly.

All major design decisions for this system will be made as a group.  We have decided that if, during the design of the system, we cannot come to a unanimous decision on any particular aspect a vote will be taken to decide what direction we will take.

5.0 Resumes/Responsibilities

All members of our team are presently in their 4^th and final year of their academic careers taking the Electronics Systems Engineering program at the University of Regina.

Ahsan Upal

-         researching the voice activation part of the project and selecting the operating systems for both the    

           mobile and the server

-         Involved in the PC hardware configurations for both the mobile and the e-mail server

-         email parsing and data transferring aspect of the project

Jim Cheung

-       project leader

-         main software designers for the M.E.S.S. system.

Dimitrios Fourlas

-         involved in the PC hardware configurations for both the mobile and the e-mail server. 

-         integrating the text to speech converter and selecting the operating systems for both the mobile and  

        the server.

-         FM Transmitter design and integration into the overall system.

Kanaris Paraskevopoulos

-         involved in the software design and implementation of the M.E.S.S. system.

-         FM Transmitter design and building.

Kristen Sparrowhawk

-      involved in the hardware design selecting various components for the actual project that would be implemented in a vehicle such as display, keypad, and the embedded PC module. 

-      designing the power system that would interface with the vehicle’s current and voltage outputs and meeting the systems requirements for the varied power specifications of each unit that would go into the vehicle.

6.0           Test Plan

Testing the M.E.S.S.  is done modularly.  The four function blocks requiring testing are the cell phone interconnection, the text to speech converter connection to the laptop, Hardware/Software Processing Unit.  They can be tested for proper functionality as separate units.  A cell phone with a modem will be the Transmission Unit.  The Hardware/Software Processing Unit will include an Embedded-PC Computer System, the ASCII to speech converter, a LCD screen, and a FM transmitter. The Speech Unit will consist of a car stereo, a microphone and speakers. The project team’s judgment will serve as the deciding factor in choosing one alternative over the other based on the test results achieved.  Testing will span six stages.

Stage 1: Cellular Phone Interconnection Testing

In the first stage, we will use the laptop to run a File Transfer Protocol (FTP) to an ISP server to download some text files.  Running a trouble free session will validate the two-way transmission of ASCII data and show that we can receive ASCII text.

Stage 2: Text Processing Testing

Once the e-mail is received and saved on the laptop, we will convert it to a text file, filtering out the attachments and headers to get  the text body contents (text) only.  This will all be done in software, and thus, it can be tested on the laptop. This testing will simply require our program to take an unfiltered e-mail message (with all the standard email headers), parse unwanted headers and attachments and output a filtered message containing only plain text.

Stage 3: RC Systems DoubleTalk Speech Unit Testing

In this stage, the DoubleTalk ASCII to speech unit will convert the plain text to voice.  The test e-mails will be comprised of commonly sent email messages (i.e. emails with attachments, forwarded emails, chain letters, etc.) for voice conversion.  Checks will be done to test the unit’s ability to pronounce the common English language words (including the commonly used slang), for clarity and understandable speech speed.  Voice from the text to speech converter must be comprehensible under noisy traffic conditions (volume control of stereo system).

Stage 4:  Voice Recognition Testing

 The M.E.S.S. will need to be programmed to recognize a single voice for the test setup with the future implementation allowing up to four distinct voices.  We will perform tests by repeatedly speaking into the microphone to ensure that the system will be able to distinguish between the single/four different users and guard against the non-users.  Also, there will be tests to ensure that the system will still recognize the user’s voice even if the sound of their voice changes by focusing on the persons voice inflections, speed and intonation.

Further tests will be carried out for voice command recognition sensitivity under varying traffic noise conditions.  We will simulate a typical traffic environment and adjust the sensitivity of the embedded microphone to be able to pick up the voice command over environmental noises.

Stage 5: Hardware Testing

The LCD controlled by the embedded processor provides a text based user selection display (for test purposes the laptop display).  Visual tests will be performed to make sure that the choices made by the user are reflected on the LCD display. The FM transmission will also be tested for operation, range and noise levels.

Stage 6: Final MESS Testing

Testing of the final M.E.S.S. will be done in phases.  If each individual module is successful, then the modules are slowly integrated together.  The embedded PC and LCD modules will be tested together to ensure that the LCD can display information from the embedded PC.  The voice recognition unit will then be added to ensure that the LCD and microphone can interface with user commands.  Next the ASCII to text unit and vehicle stereo system need to be tested to ensure that they are compatible.   Finally, the modules will be added together along with the cell phone and modem.

The final testing will be done once all the modules are assembled together.  Each function will be tested for compliance with our planned functional settings.  The system will be checked for any component or functional degradation incurred because of the environmental factors (i.e. heat, or vibrations due to mechanical parts moving in the vehicle).  The microphone sensitivity, speaker volume, voice clarity and comprehension will also go through a final testing in this in vehicle test.  If any problems are found, each module will be re-tested and tuned for compliance. This process will be repeated until all the operations reach the level of satisfaction initially agreed upon by all the team members and documented in this project report.

7.0 Schedule

The following is a schedule of our team’s efforts for the ensuing months: