Personalization and electrification are roadways to the automotive future as carmakers work on creating unique, eco-friendly driving experiences for drivers.
In the future, "software inside the vehicle will tell the control systems how to behave based on the environment it's operating in and what the driver is asking from the automobile," said Ryan McGee, a technical electrification expert with Ford Advanced Research & Engineering in Dearborn, Mich.
And analytics is helping pave the way, McGee told us in a recent phone interview. "Analytics is something we are using to improve the way future vehicles will work."
Since McGee works on electrification, he described one potential scenario for a plugin hybrid -- part battery electric vehicle and part hybrid electric vehicle. A plugin hybrid uses a large energy battery that supports significant driving distances. As that battery depletes, the driver switches over to the gasoline engine, and the car runs like a regular hybrid until the battery is recharged.
"So with this plugin technology, one of the things as engineers we need to think about is that energy in the battery. How are we going to use it? Where along the route are we going to spend it? It's a decision that the onboard vehicle system needs to make. OK, I've got this certain kilowatt hours of energy. How do I use it?"
McGee raised the possibility of various entities -- local, state, and federal governments, for example -- defining geo-fence areas in which plugin hybrid drivers would want to limit engine operation. In such cases, the control system would have to recognize a green zone in the car's route and understand that the vehicle needs to operate emissions-free in that area if possible.
"It would save energy so when it gets to the geo-fence area, there's enough energy to drive through it electrically," he said. "Without geo-fence information like that, the car would use the energy as soon as it could." And without analytics, the control system wouldn't know how much energy to save for driving through the green zone.
Analyzing an individual's driving history would help the control system make decisions based on the way the driver handles the vehicle. Does he drive the speed limit or faster? Does he use cruise control for steady speed? Does he take a certain route through the green zone? "If we can learn about driver behavior over time, we can learn how to save energy for that green zone. That's one example of how looking at driving history can improve the way the system operates."
Predictive analytics are equally as important, McGee said. He cited Ford's experimentation with the Google Prediction API. The automaker seeded the Google Prediction API service's machine learning algorithm with a bunch of driving history data for modeling. With a model, the vehicle could then ask at the start of a trip, "Where I am going next?" McGee said.
"Let's face it, most people are fairly habitual, like me -- if it's five in the evening and I'm at work and I turn on my car, 99 percent of the time I'm going home." Using the Google Prediction API "tuned with past driving data, we could get a prediction about where the car would be going next. Once we know where the car is going to go next, we can compare that information with green zones and automatically plan how the energy gets used without the driver really being involved."
Even if the prediction is slightly off -- five miles one way or the other -- the analytics could still prove valuable. "I may end up saving the right amount of energy anyways if it's close enough."
Ford has lots more to learn and discover on this journey, especially about how to collect, model, and analyze data in "a smart way," McGee said. But it doesn't plan on being taken for a ride.
If your car could analyze your driving behavior, what would it have to say about you? Share on the message board below.