“We’ll build an entire stack of self-driving technology,” Raj Kapoor, Lyft chief strategy officer, said at a media event Thursday at the company’s San Francisco headquarters. “We want it to be open to share with the industry. ... There’s a unique opportunity in time for a company like Lyft to take a leadership role ... and get people together working on this problem, rather than isolated technology development as has been happening so far.”
Lyft has no plans to manufacture its own robot cars. Instead, it wants “to create a self-driving system we can partner with the auto industry to bring to life,” Kapoor said. Those efforts will be fueled by data collected from its vast network of ride-hailed vehicles.
“Very, very key is the scale we have: 1.2 billion network miles” a year, said Luc Vincent, vice president of engineering. Lyft can use data from its cars to make high-resolution 3-D maps, a critical technology for self-driving cars to navigate, for instance.
Vincent, who joined Lyft early this year, previously spent 12 years at Google, where he led development of Google Street View, a part of Google Maps that provides panoramic views along many streets worldwide.
Lyft’s new Palo Alto facility will open next month, and the company expects to have several hundred engineers there for self-driving cars by year’s end. It will be called Level 5 Engineering Division, after the designation for cars that can drive themselves in all circumstances. (No cars on the market today have achieved this level.)
Dozens of companies — automakers, Silicon Valley stalwarts and well-funded startups — are already hard at work on developing self-driving cars, which most experts expect to be a reality in the near future.
“Some may ask: Aren’t you late to the game? People have been working on this for years,” Vincent acknowledged. But he said the technology has evolved so quickly, with such innovations as deep learning — a form of artificial intelligence that can respond to evolving situations — that it’s easier to get started now.
“When you take a partner-centric approach, using lots of off-the-shelf equipment, lots of building blocks of software, the effort is significantly more cost-effective than a few years ago,” Kapoor said.
Lyft has already said that it will open its ride-hailing network so self-driving companies can put their robot cars into service on it in the future to pick up passengers. It’s announced four such partnerships: General Motors, which is a major investor in Lyft; Waymo, the self-driving division of Google parent Alphabet; British carmaker Jaguar Land Rover; and self-driving startup NuTonomy of Cambridge, Mass. Lyft executives plan to test NuTonomy’s self-driving cars as Lyft ride-hailed vehicles in Boston this year. NuTonomy has been testing its cars with passengers in Singapore since last fall.
Former Uber CEO Travis Kalanick famously infuriated drivers by saying the company would save money once it no longer had to pay the “dude” in the driver’s seat.
By contrast, Lyft has taken pains to say that its human drivers aren’t going away. It visualizes a hybrid network of human-driven and self-driving cars, with the system selecting whichever is best for a passenger’s requested route.
“Drivers play an important role and will continue to play an important role,” Kapoor said. “Some people prefer to have a human driver; in the short run there will be instances that self-driving cars can’t handle.”
In a blog post, Vincent went even further: “Lyft will always operate a hybrid network, with rides from both human-driven and self-driving cars,” he wrote.
Drivers might also take on new tasks, such as assisting elderly passengers, running cafes in cars or being car concierges who travel along, Kapoor said.
Long-term, robot taxis will lower ride costs, inspiring more consumers to subscribe to transportation services, rather than owning their own cars, Kapoor said, analogous to how the cheaper rides from Uber and Lyft grew the industry beyond the size of the taxi market.
Last year, Lyft co-founder and president John Zimmer predicted that private car ownership will be obsolete by 2025. But Lyft now seems to be setting expectations that the transition to self-driving could be a lengthy one, cautioning that there are numerous circumstances — rain, construction, traffic and nighttime — that could faze autonomous cars.
“It won’t happen tomorrow (with) some mechanized fleet of autonomous vehicles blanketing San Francisco,” said Taggart Matthiesen, senior director of product. “It will happen in small pockets.”
©2017 the San Francisco Chronicle. Distributed by Tribune Content Agency, LLC.
