Autonomous Vehicles: To Park Or Not to Park?
June 6, 2019 | In The Press
One of the more astonishing statistics about Apple’s new headquarters (registration required), ironically named Apple Park, is that it features 11,000 parking places for 14,000 employees. The campus has more square footage dedicated to parking (about 3.5 million) than offices (3.25 million).
As the CEO and co-founder of a fleet-management platform, I know one looming question regarding the advent of autonomous vehicles and services for cities is what will become of parking. There may be as many as 2 billion parking spaces (paywall) in the United States. Will they disappear or overflow?
One prediction is that parking as we know it will vanish. If shared autonomous vehicles become the primary means of personal transit, we’ll need fewer vehicles — and places to park them.
On the other hand, the researcher behind a February 2019 report (paywall) estimated that operating a driverless car costs just 50 cents per hour, while INRIX — a connected car services company — reported that the average hourly cost to park in the city is $6.00. With numbers like that, it could be much more cost-effective for vehicles to cruise around empty or to return home or to depots rather than to park — in either case adding to congestion.
Some urban designers are preparing for a parking-free world. The Los Angeles Times noted in 2017 that one of the largest apartment builders in the area is designing parking garages that can serve other purposes in the future, like revenue-generating office space, assuming that the demand for parking will decline. Existing parking structures could be repurposed for more people-centered uses. The parking spaces in the country make up a substantial land mass. Public parking lots could be converted to open space; private ones could be redeveloped for more dynamic uses.
Can we have it both ways — increased traffic with less parking and more open space? Not likely. The prediction of increased traffic assumes widespread individual ownership of autonomous vehicles. The parking projections assume widespread sharing. Dr. Kara Kockelman of the University of Texas said, “If everyone did it, you could get rid of 7 out of 8 cars on the road, so you’d need an eighth of the spots.”
A study Kockelman coauthored modeled traffic in the Austin area and found that one shared autonomous vehicle could do the work of nine conventional vehicles “while still maintaining a reasonable level of service.” That’s likely a net time saver as the 2011 IBM Global Parking Survey found that, globally, people spend an average of 20 minutes searching for parking.
And cities may also have to change their policies on parking, perhaps even reversing the requirements to call for limited auto parking spaces or paid parking to encourage alternative modes of transit. In Norway, for example, the Oslo Solar high rise offers free parking for bicycles and charging stations for electric vehicles, but no spaces for conventional cars.
Shared autonomous vehicles will need to go somewhere when not in use and to be cleaned and recharged. Facilities to service and store the vehicles face their own parking challenges. A study by University of Toronto researchers ( via IEEE) noted that if vehicles are self-parked too tightly, chaos could ensue when, for example, a vehicle at the center of the lot needs to get back to work. Fleets will have reshuffle their positions if they don’t have enough buffer space around each one. That buffer space, though, reduces the number of vehicles that can be serviced.
For vehicle sharing — dropping off and picking up multiple riders — to be more convenient than solo vehicle journeys, the rides and the fleets that deliver them will need to be well-orchestrated. (While this is something my company helps with, companies can also accomplish this themselves.) For example, fleets will need to pool, pick up and drop off riders by optimizing their trips. They’ll need to plan when to return to the charging depot and coordinate parking so it does not block vehicles ready to return to service. Other ways to foster shared transit are to create “fast-track” carpool lanes in cities as well as on highways and to create “red carpet” bus-only lanes to give buses an edge over private autos. An alternative approach I’ve seen many call for is to make single occupancy vehicle travel less convenient than shared transit by prohibiting or charging for entry in city centers by car with congestion pricing (paywall).
New modes of transit alone don’t always work. Witness peer-to-peer ride-hailing services, or transportation network companies (TNCs), like Uber and Lyft. According to a 2018 report, they have increased traffic on city streets by 180%, as they “put 2.8 new TNC vehicle miles on the road for each mile of personal driving removed.” With autonomous vehicles, traffic will only be as efficient as the services that the fleets are able to provide. Models like Kockelman’s show that the potential is enormous if service providers and cities can get the services right.