The 2,578 Problems With Self-Driving Cars
California Says This Time, For Sure, It Will Issue Rules on Driverless Cars
BMW to Demonstrate Car That Can Find a Spot and Park Itself in a Garage
Autonomous vehicles could transform parking as well as driving, new research suggests. Parking lots could house more driverless cars than human-driven ones, but autonomous vehicles could also lead to nightmarish gridlock if they slowly cruise the streets waiting for their owners, instead of paying to park.
The typical vehicle spends 95 percent of its lifetime parked. The need to store parked vehicles has turned a lot of potentially valuable real estate into parking garages—for example, in the United States, roughly 6,500 square miles of land is devoted to parking, which is larger than the entire state of Connecticut.
Autonomous vehicles could, in principle, transform parking lots. When people park cars, they need space to open the doors and walk away, but driverless cars do not need such space. And parking lots devoted to autonomous vehicles would not require elevators and staircases—the owners could simply be dropped off by the machines at the entrance or some other destination and let the vehicles park themselves.
To see how much autonomous vehicles could increase the capacity of parking lots, researchers at the University of Toronto calculated the optimal size of the grids in which driverless cars park themselves. If a grid is too large, it takes longer on average for one of the autonomous vehicles to exit, since the surrounding cars have to spend time relocating as needed. If a grid is too small, it wastes space.
Image: Sina Bahrami
The researchers developed a computer model that simulated the effects of various layouts for autonomous vehicle parking lots. They found that such lots could accommodate 62 to 87 percent more cars than ones for conventional vehicles. This more efficient use of space could lead to smaller parking lots, and possibly free up urban space. “That land could be allocated to other activities—residential space, perhaps, or maybe green areas for everyone,” says study co-author Sina Bahrami, a transportation engineer at the University of Toronto.
The researchers detailed their findings in the March 2018 issue of the journal Transportation Research Part B: Methodological. They are now investigating how autonomous vehicle parking lots might become even more efficient if they know when driverless cars are scheduled to arrive and leave, Bahrami says.
However, if the owners of self-driving cars want to avoid paying for parking and decide their autonomous vehicles should cruise the streets instead, robotic gridlock could result, says transportation planner Adam Millard-Ball at the University of California, Santa Cruz.
Driverless cars that cruise instead of park could make traffic worse simply by increasing the number of vehicles on the streets. Moreover, cruising is also less costly at lower speeds. Millard-Ball estimates that even after factoring in energy, depreciation, wear and tear, and maintenance, cruising may only cost 29 to 50 cents per hour. His computer models suggest that as few as 2,000 empty self-driving cars in downtown San Francisco may slow traffic to less than 2 miles per hour.
Preventing such a future by regulating cruising is challenging, Millard-Ball says. Any law that, say, targets autonomous vehicles cruising for more than 10 minutes without a passenger may also discriminate against vehicles cruising for reasons other than avoiding parking, such as picking up parcels, he notes.
Instead, Millard-Ball suggests congestion pricing, where motorists essentially pay a fee to enter city centers and other potentially congested areas. “We’re trying to give cities the tools and incentives they need to think about policies that are going to be robust if and when autonomous vehicles arrive,” Millard-Ball says.
Millard-Ball detailed his findings in the March 2019 issue of the journal Transport Policy.