Can your car anticipate your driving patterns and optimize its performance to use fuel more efficiently, or extend your range if you are driving an electric vehicle (EV)? Can it see into the future? Ford is trying to develop the software and systems to enable it to do so.
Google has been developing technology for cars to drive themselves, but the Ford project has more limited aims. If your car could gather information as you drive, and gradually build a model of your driving habits, plus download information about your driving environment, it could fine-tune its operation to save energy.
Ryan McGee, technical expert on vehicle controls architecture and algorithm design at the Ford, says, “We have this massive amount of data. The question is what to do with it.”
A recent piece from Greentech Media outlines the concept:
“Code-named Green Zone, the software tries to anticipate where you plan to drive. Say it’s 8 a.m. on Tuesday. Your car knows that this is the second day in a five-day sequence in which you drive 23.5 miles to the same destination. The software crunches data about your driving habits, the topography of the drive, any details about traffic and time-to-destination, and information about how the car performs. It then tries to maximize the power the car draws from the battery pack and minimize the work performed by the gas engine.”
This scenario is for an EV with a backup gas engine, like the Chevy Volt. But the same concept could be applied to hybrid vehicles like the Toyota Prius, plug-in hybrids, full battery-electric vehicles like the Nissan Leaf, or even just internal combustion gasoline or diesel vehicles.
The article says the car’s systems would connect to cloud-based data resources to manage all the necessary data. In addition to data from the car and driver themselves, such a system would obviously also incorporate data such as the local current and forecasted weather, the local topography and traffic along the anticipated route, the availability of charging points at the likely destination, and so on. Each of these is a complex data model of its own.
Your car would know (probabilistically from historical data, via models, or via real-time data collection along the route) what the traffic was like ahead. In hybrid vehicles the different systems (electric motors, gasoline engine) operate optimally under different conditions. The car could plan its use of these resources to most efficiently deal with different speeds, idling times in stop-and-go traffic or at lights, anticipated episodes of acceleration or braking, and so on. It wouldn’t have to wait for you to press the gas or the brake to know what was going on.
Greentech says “The probabilistic principles underlying the experiment are similar to predictive algorithms exploited by search engines. In fact, Ford uses Google’s predictive APIs.”
We are all familiar with the computers that have become important parts of automobiles to operate their many high-tech systems. But now we should get ready for cars that have whole IT systems, and communicate moment-by-moment with vast data structures in the cloud.