JLR engineers use the America's Cup race to gather performance intel

In the late 1980s, Formula One cars began being fitted with sensors to measure fuel use, engine characteristics, braking performance and other metrics. Now manufacturers of road cars want massive amounts of data as they develop self-driving vehicles.

At the Ford display last month at SAE International's World Congress Experience in Detroit, two Ford vehicles that use big data -- a racing version of the GT supercar and a self-driving Fusion sedan -- were parked within feet of each other.

Both vehicles were loaded with sensors; the Fusion's trunk was filled with electronics equipment used to collect and transmit data.

Ford wants to have a fleet of Level 4 autonomous cars on the road by 2021. Dave Schmidt, a Ford data project manager working on the company's self-driving cars, says the connection today between the GT and the Fusion is how the useful data are used to improve performance, as is the case with the Land Rover R1.

"A key part of autonomy is being able to analyze data, to optimize the performance of the vehicle, to optimize the service the vehicle is trying to deliver," says Schmidt. "You could drive down a long highway in the middle of Wyoming and never experience an event. There's probably nothing relevant in that data, but you've collected a tremendous amount of it."

So how to extract the important information? Indeed, Schmidt says engineers are struggling with how to offload huge amounts of data from autonomous cars. F1 cars use Wi-Fi to do data dumps. And Schmidt says Wi-Fi looks like the best option now because using cellular phone technology is slower and more expensive.

Dealing with big data is a problem Delphi is also working to solve. The supplier is developing automotive electrical architectures geared to moving data 10 times faster than today, a target Delphi says is required. Last month, it invested in three companies that specialize in managing data.

By 2025, IHS Markit estimates that 630 million of the expected 1.6 billion light vehicles on the roads will be connected, producing a staggering amount of data. An autonomous Audi SQ5 that Delphi engineers rode in from San Francisco to New York in 2015 generated nearly three terabytes of data; Delphi engineers spent months parsing it.

Munoz says the experience gained working with massive amounts of data will teach engineers how to sift through and find data that can be used to improve vehicle performance.

JLR last year rolled out a suite of self-driving technologies that enable a Land Rover vehicle to drive itself in some situations off road. The Surface Identification and Over the Horizon Warning systems, for example, rely on big data to enable the vehicle to be aware of its surroundings so that critical safety decisions can be made.

"If there are any learnings from R1, it is that we have this huge data set," Munoz says. "How can we turn big data into significant small data and then work with that small data to get answers relatively quickly?"