Researchers program plug-in hybrid cars to learn fuel efficiency

(from Reuters video)

CALIFORNIA, United States (AFP) — Researchers at the University of California, Riverside have created a computer program for plug-in hybrid electric vehicles (PHEVs) that learns how to save fuel, improving fuel efficiency by more than 30 percent.

PHEVs can use either gasoline or electric power. Lab tests have shown the most efficient way for them to run is to switch between power from the battery and gas throughout the trip, rather than running exclusively on the battery then switching to gas when it runs out.

The key to optimizing fuel efficiency, then, lies in making the vehicle learn when to use which energy source, according to Matt Barth, who led the research.

“If you look at plug in hybrid technology today, a lot of it is just simply based on sensors that are on the vehicle itself. So it has an onboard energy management algorithm that uses sensors that kind of figures out what’s the best way to use fuel at different times. And so, what we’re doing and proposing is not only look at the sensor data on the vehicle, but look at data also outside. So traffic data, traffic signal data and also looking at patterns and how people drive. And using all those new data sets, we’re trying to improve those plug in hybrid algorithms to get a little bit better efficiency than what we’re getting today,” Barth said.

Barth hopes the new technology could be available on new cars within three to five years.

“The data exists. The communication capability exists. It’s really just putting more smarts on the vehicle that can take advantage of that information,” he said.

The number of electric vehicles — including PHEVs, battery-electric and fuel cell electric passenger light-duty vehicles — on roads worldwide rose to a record high of 2 million in 2016, but the vehicles are still only a small fraction of new cars sales.

Barth says the fuel-efficiency algorithm could provide a competitive advantage to automakers in the PHEV market.

“I think all manufacturers would be interested in using that outside information to improve their vehicles,” Barth said.