Either way, should we expect this new wave of companies to catch up with the frontrunners? Unsurprisingly, Cruise executive vice president of engineering Mo ElShenawy wasn’t convinced. “The state-of-the-art technology that exists today is not enough to get us to where Cruise is,” he said.
Cruise is one of the most advanced driverless car companies in the world. It has been running a live robo-taxi service in San Francisco since November. Its vehicles operate in limited areas, but now anyone can use the Cruise app to hail a car and have it park on the side of the road when no one is there. “We’ve seen real reactions from customers,” ElShenawy said. “It’s super exciting.”
Cruise built a sprawling virtual factory to support its software, with hundreds of engineers working on different parts of the pipeline. ElShenawy believes the mainstream modular approach is an advantage because it allows companies to swap out new technology as it emerges.
He also dismissed the idea that Cruise’s approach won’t generalize to other cities. “We could have launched in the suburbs somewhere a few years ago, but that would have us stuck,” he said. “Our choice for a complex urban environment, like San Francisco — where we see thousands of cyclists and pedestrians as well as emergency vehicles and cars — was deliberate. It forced us to build easily scalable thing.”
But before Cruise can drive into a new city, it first has to map the streets in centimeter-level detail. Most driverless car companies use these types of high-definition 3D maps. In addition to raw sensor data acquired during the journey, they provide additional information to the vehicle, often including cues such as lane boundaries and the location of traffic lights, or whether there are curbs on a particular street.
These so-called high-definition maps are created by combining road data collected by cameras and lidars with satellite imagery. Hundreds of millions of miles of roads have been mapped in this way in the United States, Europe and Asia. But road layouts change every day, which means mapping is an endless process.
Many driverless car companies use high-definition maps created and maintained by professional companies, but Cruise makes them itself. “We can rebuild the city — all the driving conditions, the street layout, everything,” ElShenawy said.
This gives Cruise an edge over mainstream competitors, but newcomers like Wayve and Autobrains have ditched HD maps entirely. Wayve’s cars have GPS, but they otherwise learn to read the road using only sensor data. It might be harder, but it means they are not bound by a specific location.