Short-term Goals
The goal of the Autotender project is to create a machine that can automatically create mixed alcoholic drinks. The Autotender will be preset with hundreds of mixed drinks in different variations (there's more than one way to make a Bloody Mary). Everything will be input and selected by a touchscreen graphical user interface. It will also be possible for the end-user to create custom drinks on the fly by simply entering the ratios on the screen. I.E 1 part vodka, 2 parts orange juice, etc... and then dispensed into selected drink size.
 Long-term Goals
The Autotender will eventually be RFID enabled allowing users to save their drink preferences and dispense drinks simply by scanning their RFID tag as the graphical user interface naturally becomes less intuitive as the machine is used more. This would allow a user to simply place their glass under the dispenser, scan their tag, and it dispense their last drink, favorite drink, or whatever preset option the user has selected.
This would also allow for tracking and record-keeping allowing for such things as lists of who drank the most, calculating a users BAC (if user has entered the other necessary parameters and only drank from this machine), etc...
The Autotender is currently in the design stage. Nothing prototyped/tested as of yet.
 Liquid Dispensing
To keep costs low, but still maintain decent accuracy the prevailing method for dispensing the different liquids is to use a servo controlled bulb pump. The bulb will be clear to allow for automatic priming and sensing of proper liquid dispensing so that end-users are not shorted by a half-filled bulb due to source liquid running out. The sensing will be done by 2 simple LED and photo transistor pairs. 1 at bottom of bulb, and 1 at top. The top pair is for detecting a full bulb so that the correct amount of liquid is dispensed, then bottom pair is used to determine if there is liquid present in the bulb at all (indicating that we might have ran out of vodka). This can be used to detect even clear liquids as the light from LED will be refracted when passing through the liquid changing the level of light sensed at other end.
I actually thought of another idea that may not be as cheap, but will be more accurate and less complicated. The new idea is to have the source liquids stored at top of machine, each one is fed down through a tube that connects to a small solenoid valve (http://www.ascovalve.com/Common/PDFFiles/Product/8256_NSFR2.pdf) all of these valves then connect to the holding tank which would now contain a liquid level sensor (http://www.sparkfun.com/datasheets/Sensors/Flex/eTape%20Datasheet%206573P-8.pdf). The sensor then measures the amount of alcohol dispensed based on volume of tank & height of liquid. While the liquid is being dispensed the rate of flow will be calculated so that the valve is shut off BEFORE nominal liquid is dispensed to prevent dispensing more than needed.
Liquids will be dispensed to a central mixing tank containing several sets of rotating blades. These will simply be connected to a DC motor and run when it's time to make a drink. The blades will turn on before any drink is dispensed so that the drink is being mixed will adding ingredients. Nothing real fancy here. These blades are non-metallic as any possible spark source inside of a sealed container full of combustible liquids can't possibly end well.
Still juggling with ideas on how to handle ice. Not set on any specific method at the moment. Ideas include blending it with the liquid itself (I've never actually had a drink made with crushed ice?), or simply dispensing it as cubes into the cup, or perhaps both methods as a user selection? There's also the simpler method of just not doing ice at all and require a separate ice machine, no sense reinventing the wheel right? And in fact for initial prototyping purposes ice creation/dispensing will not be included as I'm more concerned with innovation rather than recreation.
 Final Drink Output
The final drink will be output through a simple solenoid valve at bottom of mixing tank, gravity fed down into cup area.
 Graphical User Interface
The original idea was to used an Ipad, but after reading about the pandaboard, it would be a lot cheaper and easier to develop using the pandaboard combined with an off the shelf LCD touchscreen monitor.
Feel free to provide any feedback or ideas you have on the Autotender, I am open to suggestions. This is an open project and full schematics and circuit diagrams will be disclosed so you can build your own.