Isra-Mart srl : Ford and Google tie up to predict car trips and optimise plug-in hybrids

www.isra-mart.com

Isra-Mart news:

Ford is working on a technology that learns where and when a motorist drives, customising the car's performance to suit particular journeys and optimising plug-in hybrid vehicles to make them more efficient.

The project uses Google's Prediction API, a machine-learning system that takes historical data and makes predictions based on the information.

The Prediction API can be used to analyse everything from document and email classification, through to spam detection and language identification. However, Ford wants to use it to analyse drivers' trip data, enabling the car to 'guess' where the driver might be going.

A scenario described by Ryan McGee, a technical expert at Ford Research and Innovation, is a car asking its driver whether they were about to drive to work, for example.

If the driver answered positively, the car would check to see whether there is an 'electric vehicle-only driving zone' on the way to work, where it makes more sense to use the electric motor in the vehicle.

The car would then use the petrol component of the engine until it reached the EV-only driving zone, and would then automatically switch to the electric motor.

"This isn't vapourware - we are actually working on it," McGee said, adding that the system could eventually adapt to changes in users' driving habits over time, understanding how their patterns changed.

"Ford has been connecting drivers to their vehicles for some time now using its Sync system," he continued. "The next stage is to connect the vehicle to the cloud."

A cloud connection would be necessary to harness the computational power needed to handle the prediction processing.

Ford has demonstrated this technology with a series of test drivers, who opted to provide their data. However, McGee did not outline how privacy would be protected should the system scale commercially.