Voice control and interaction only

**BrownieMan** · 07-19-2008, 11:53 AM

Right now I'm doing the research and some coding to develop an app that strictly uses sound, so no visual display.

I like this idea for a few reasons.

1) It's much cheaper since you don't have to buy a display
2) You don't have to worry about theft as much, since there is no display to entice thieves
3) It's pretty impressive
4) You never have to worry about getting tickets (I've heard that it might be illegal to have a display in the driver's live of sight).
5) It's much safer than looking down and touching a display

I'm going to be writing this thing in VB and is only in the design and prototype phase.

Realistic ideas so far:
1) Audio/media application
2) Car Diagnostics via OBD-II
3) GPS (I'm not that great that I can code a mapper from scratch, does anyone know an open source VB(.NET) mapper that I can use?)
3a) memorizng speed traps and other other things (wifi?)
4) Phone dialing and manipulation (likewise, I think I'll need something open source to "steal" from)
5) PDA functions
5a) contacts (ie. "What is John Smith's address?")
5b) schedule (ie. "What are my plans for today?")
6) note taking

Does anyone have any comments or things they would like to see, or hear?

**karrhacker** · 07-19-2008, 01:46 PM

I hate to be negative but I tried this a few years back, (not sure it the software has changed any or not) but it really doesn't work well.. With the combination of road noise, interior noise and wind noise (unless you car is like a coffin inside) it makes it almost impossible for the computer to recognize what you are saying. Then there is the HUGE amount of data the you have to load into memory for commands such as "PLAY Shakira-Oral Fixation, Volume 2 How Do You Do" ..

As for it being illegal for the displays to be in the line of sight of the driver, well how would we use the GPS systems? or Radio functions, as many car manufacturers are starting to use large touch screens instead of the old 2" radios.

But good luck with everything, and let use know how you make out..

**BrownieMan** · 07-19-2008, 02:36 PM

Originally Posted by karrhacker

I hate to be negative but I tried this a few years back, (not sure it the software has changed any or not) but it really doesn't work well.. With the combination of road noise, interior noise and wind noise (unless you car is like a coffin inside) it makes it almost impossible for the computer to recognize what you are saying. Then there is the HUGE amount of data the you have to load into memory for commands such as "PLAY Shakira-Oral Fixation, Volume 2 How Do You Do" ..

As for it being illegal for the displays to be in the line of sight of the driver, well how would we use the GPS systems? or Radio functions, as many car manufacturers are starting to use large touch screens instead of the old 2" radios.

But good luck with everything, and let use know how you make out..

well, I already have voice recognition working very well. On my desktop at home it says Dragon Naturally Speaking (DNS) uses 3% of my processor (AMD 3200?), so that's probably at most 100mhz processing power on an intel or similar processor. It also uses about 50MB of RAM. This is on idle right now though, I'll have to check later when I'm actually using DNS.

I have the program just ype whatever I'm saying into a text field. After each time I stop talking (don't ask me how DNS does it exactly, a timer or lull in sound maybe?) the field resets itself. After that it's just a matter of parsing the words. If int eh sentence it sees the word PLAY, it knows to switch to audio mode and try and search for a file called whatever was after the word play. DNS is the only thing that really requires a lot of processing power, everything else is pretty trivial.

As for the legality thing, I think you will see more enforcement of it or regulation of GPS's and touchscreens in cars when they are in EVERY car like cell phones are.

**karrhacker** · 07-19-2008, 02:56 PM

Have you tried using the MS Voice SDK? Very simple to use.. I personally have never used DNS....

**BrownieMan** · 07-20-2008, 07:36 PM

Originally Posted by karrhacker

Have you tried using the MS Voice SDK? Very simple to use.. I personally have never used DNS....

hmm, no. I'll have to look into it.

**Excellrec** · 09-30-2008, 12:57 PM

I'm curious if anything more has been done with this. I have a showcar that I'm building that I aim to implement voice control and voice response into. I would like to be able to control multiple programs with voice control then also get feedback about situations on sensors through voice response. I know it's very ambitious but so everything about this car. Anyone have any ideas about how to go about this? Microsoft voice sdk?

**BrownieMan** · 09-30-2008, 05:30 PM

Originally Posted by Excellrec

I'm curious if anything more has been done with this. I have a showcar that I'm building that I aim to implement voice control and voice response into. I would like to be able to control multiple programs with voice control then also get feedback about situations on sensors through voice response. I know it's very ambitious but so everything about this car. Anyone have any ideas about how to go about this? Microsoft voice sdk?

I've kinda put my project on hold for now. The biggest problems I've run into is interfacing with said sensors and programs. I'm only a recent graduate of Computer Science,s o I have a lot to learn.

Interfacing (controlling?) with popular programs might or might not be possible. Our best bet is to use open source applications that can be interfaced with via popular programming languages (Java, VB, .NET). If anyone can list or point me in the right direction of some programs that would seem to fit, please do so.

As for sensors... That is far more difficult. I know almost nothing about drivers, which is what is needed to interface the hardware (sensors) with the computer. I did come across something that does primitive stuff on here, basically just turning on and off and sensing when there is a current.

Please, I encourage anyone and everyone to comment. That's how things get done.

**Excellrec** · 09-30-2008, 06:16 PM

well in my situation I was looking at using fusion brain http://www.fusioncontrolcentre.com/FusionStore/catalog/index.php

to control the sensors and then the information from the sensors would be in the fusion brain program. So perhaps it would be easier to program with sensor information from the fusion brain program? I have no idea myself, i've only dabbled in c+ a small amount.

**2k1Toaster** · 09-30-2008, 07:07 PM

As a plus, the Fusion Brain's software (MDX) is all open source C# and has voice parsing control so you can look at the Microsoft Speech SDK in use there. It works really well.

**DVLRCR** · 10-03-2008, 01:57 AM

It is fairly easy to interface with the sensors. The Fusion Brain is one way, another is using a simple micro controller to control and read the sensors and connecting to it using the RS232 port on your computer (or a USB to RS232 thing)

Thats what i was planning on doing anyway to control my AC