The answers to many of the problems surrounding the introduction of autonomous vehicles may involve keeping human drivers on the road too, according to Lyft Vice President of Autonomous Vehicle Programs Nadeem Sheikh.
Sheikh spoke about applying a hybrid network approach to the deployment of self-driving vehicles at the 2018 Automated Vehicle Symposium in San Francisco, California.
Lyft introduced a small fleet of self-driving cars to its network earlier this year, giving the general public a chance to experience autonomous vehicles firsthand.
The popular ride-sharing company partnered with leading global technology company Aptiv to deploy 30 autonomous BMW 5 Series cars in Las Vegas, Nevada. Now, when travelers request rides to and from locations serviced by the self-driving fleet, they are able to opt in or out of the experience.
Lyft’s self-driving vehicles are limited to a narrow set of specific routes and conditions, and human safety drivers are present in all the cars.
Riders who have opted into Lyft’s controlled autonomous experience have reported feeling safe and finding the overall experience “cool,” according to Sheikh’s presentation.
“People were actually falling asleep in the back of the car,” Sheikh said.
Despite the general public’s interest, significant operational challenges still exist when it comes to widespread roll-out of the technology.
Offsetting high prices
The process of building autonomous vehicles comes with a hefty price tag. For ride-sharing services like Lyft, post-production upkeep costs are also something to consider.
Lyft’s human drivers currently shoulder the costs associated with vehicle maintenance, cleaning and storage. The drivers repair flat tires, find parking spots each night and throw away stray coffee cups left in their back seats.
The company will have to take on those additional costs where its self-driving fleet is concerned.
One way the company intends to do this is by building service centers in some of its most active cities. The centers will be equipped to service both self-driving vehicles and current Lyft drivers’ private vehicles.
The centers will provide maintenance, cleaning, parking and charging services, according to Shiekh.
Recognizing limited functionality
Not only are self-driving vehicles expensive to produce and maintain, they also come with a unique set of functional limitations.
Autonomous vehicles can’t go nearly as many places as human-operated vehicles.They also can’t drive at highway speeds, and dead miles can kill economics for companies like Lyft and Uber.
Shiekh said the self-driving vehicles Lyft currently deploys can only travel at speeds under 35 mph, require well-marked roads and must avoid difficult intersections and areas with bike lanes. This heavily restricts where self-driving technology can be used.
While Shiekh expects most of these limitations to be worked out in time, it is clear that autonomous vehicles are no match for human drivers in 2018.
Managing rider expectations
Another unique challenges for ride-sharing companies hoping to deploy more autonomous fleets is managing customer expectations.
Shiekh said most customers have come to expect a three to five minute ETA when using ride-sharing apps in the U.S., so it is important to make sure autonomous vehicles are close enough to the requested pick-up location to avoid losing customers to the competition.
Riders not only expect a timely pick-up, they also expect to be picked up and dropped off at the specific locations they request, not a couple blocks away, making it important for self-driving fleets to be able to navigate a variety of different types of pick-up and drop-off locations, according to Shiekh.
Difficulty navigating these expectations is compounded by the fact the riders are not willing to pay more for an autonomous ride because, on the novelty rubs off, they really only care about getting from one place to the next.
The hybrid solution
Shiekh detailed Lyft’s approach to overcoming these challenges during his presentation.
The multi-layered solution included incremental deployment, creating a human-autonomous hybrid network and eventually, persuading the public to ditch their personal vehicles in favor of a ride-sharing subscription service.
Shiekh presented a three-step deployment plan, with today’s highly limited self-driving cars making up phase one and fully autonomous vehicles making up phase three. During each phase, Shiekh predicted Lyft’s human driver force would continue to grow.
Human drivers make up the overwhelmingly majority of Lyft’s network during phase one, and they will continue to represent about half the network once phase three is accomplished, according to the presentation.
In order to continue climbing the deployment ladder, Lyft is focused on keeping the process cost effective and utilizing its human drivers.
Stay up-to-date with the latest commentary and insights on FreightTech and the impact to the markets by subscribing.