As driverless cars appear on roads across the globe, Europe is putting safety at the heart of its autonomous driving efforts. As BMW’s chief executive officer Harald Krueger has repeatedly declared: “Safety comes first.” Some may see Europe as the tortoise in the race to produce autonomous cars. But the rigorous testing and regulated safety favoured by Europe may still help it inch first over the finishing line.
Whichever focus wins — entrepreneurial or regulated — data management will be a critical capability in delivering driverless vehicles. Technology provider DXC Technology is partnering with automakers and providing data management in myriad forms. Data harvesting, storage and analytics, alongside manufacturers’ prolific and rigorous test cycles, make a winning formula, says Jim Miller, vice president and chief technology officer of DXC’s Cloud and Platform Services.
“Ingesting data into a central location for analytics and machine learning to be applied can present a significant challenge for manufacturers,” says Miller. In the first place, difficulties are presented by the sheer volume of the data. Automakers have large fleets of test vehicles that run 24×7, and the vast swathes of data generated need to be mined for patterns, which help formulate the algorithms.
The second challenge is data fusion, says Miller — dealing with multiple types of data. Significant amounts of data are in video format from cameras, in LIDAR (light detection and ranging) data and radar. Additionally, vast amounts of data are produced by sensors, including steering wheel and throttle position, speed and g-force, says Miller. “The list goes on — and the test data in different formats all provides patterns for algorithms to learn from,” he adds.
The “safety first” ethos of the auto industry means it will probably reject the general trend towards blended artificial intelligence (AI) — humans and algorithms working together — Miller believes. “A handover between driver and autonomous system sounds simple, but the reality is more complex,” he says, adding that in an emergency, it may not always be possible for a human to recognise a complex driving situation fast enough to safely take over.
He is sure about the three forces that will shape the journey of the autonomous car, though. “Technology is table stakes, but policy (including governance and environmental factors) is a key subject that crops up in any industry discussion,” he says. Car ownership is the third driver. “Autonomous vehicles (AVs) will change car ownership patterns. We’ll see ride-sharing and the use of AVs in high-occupancy lanes, for example, where risks are lower.”
Autonomous vehicles will likely be produced in all shapes and sizes, but Miller prefers to talk about future trends rather than specific models. He envisages driverless vehicles able to process the destinations of passengers, and optimise the route — and cost — of the journey. AVs are flourishing on business campuses, he notes, where he expects autonomous multi-passenger vehicles to be the first deployments.
In the United States, where the average lifespan of a car is 11 years, the turnover of the nation’s vehicle fleet will take decades. During that time, traffic on the roads will consist of a mix of traditional-type vehicles, advanced driver-assistance systems (ADAS) and driverless cars. As Miller notes, the autonomous journey will be incremental: “Automated emergency braking and collision avoidance were high-end a few years ago. Today, they’re standard.”