My entry is a cloud based app that implements an extensible calculus that describes observed and predicted driver and car behavior. Input is live vehicle data from osisoft, including gear position, speed, acceleration, headlight and windshield wiper usage patterns. Output is commercially viable information based on a predictive model of driver and engine: driver identity, personality, current mood, geographical locations: entering, leaving, next, and expected near future mechanical problems. I demonstrated three practical uses: gas station, fast food, supermarket.
For the gas station I used an iobridge board to mix additives to personalized gasoline.
For the fast food and super market UI, the geofencing, I made a simple iPad app based on an esri template.
For the GPS voice I used Siri on the iPad. "Free icecream in half a mile on the right"
The Mac OS X apps and shell scripts are helpers, used while debugging. dhmapp.sf is the source code for the entry.
Code, screenshots etc: dhm.com/m2m/