AI-based range prediction more accurate than manufacturer systems

Real-world trials of an AI-based range prediction system have delivered dramatically higher accuracy than manufacturer systems, says Spark EV Technology.

Spark EV Technology has collaborated with BP to provide real-world trials which validated its advanced range prediction system for electric vehicles.

The project involved BP employees based at five locations drive 10,000km in two types of EVs equipped with Spark’s journey prediction system. Drivers entered their proposed journey into Spark’s phone app to record data and make comparisons with the onboard range display. Personalised advice was then generated regarding the true range – based on live data, driving style, routes-types, previous trips and charge point locations.

Data generated from the trial highlighted that the vehicles’ displayed range was underestimated in some instances by between 21 and 45 miles.

Justin Ott, CEO, Spark Technology, said: “At Spark our mission is to make electric vehicles mainstream by removing barriers to adoption around range and time anxiety.”

“Through our high profile collaboration with BP we have taken the first steps to demonstrate the strength of our unique technology and how it can drive greater uptake of EVs through more accurate range predictions, based on real-world conditions and driving styles.”