Published Mar 23, 2020
During the hype of self-driving cars, many people said parking would no longer be needed. The rationale is pretty simple — if a car can drive itself, it can go on to pick up the next person or delivery, and it will never need to sit idle in a parking lot.
The reasoning sounds plausible, and the conclusion is wonderful. After all, why waste spaces for unused cars. Parking lots can be turned into buildings, restaurants, parks… It’ll make urban areas much better places to live.
However, the critical link that’s missing in this chain of reasoning is: Will cars always have a next someone or something to pick up?
To answer the question whether all self-driving cars will always have places to go, we need to look at urban transportation at a macro level.
For most cities, there are two traffic peaks on weekdays: One in the morning when people go to work, and one in the evening when people return home. Take a moment to think about the gap between the peak and the valley: Where are the vehicles that transport people at the peak, but not used at the valley? You probably know the answer: Mostly parked.
Parking serves as a storage for vehicles during low demand periods. Only an almost-flat traffic curve could make parking disappear, and we know that is not the demand pattern in cities.
As a result, in a future with self-driving cars, a big proportion of vehicles will still need to be parked somewhere during the day. Then the next question becomes: Where to put self-driving cars when they are not needed?
There are many places self-driving cars can go when they are not needed.
An obvious but terrible solution is to let these empty vehicles circle around city streets. This is even worse than parking. Essentially, the empty vehicles use streets for parking, except they also slow down other vehicles and waste energy.
Another option is to let self-driving cars leave urban areas and park outside cities. This has its own issues: It’ll lead to four traffic peaks rather than two: one is the normal morning peak, one when cars head out of the city after the morning peak, one when cars come back to the city to pick people up, and the last one is the normal evening peak. Another issue with this solution is related to the evening peaks. When you need to move a large amount of cars into the city to pick people up, it’ll cause traffic and delays. Therefore some cars need to come into the city earlier than others, leading to parking problems once again.
These are just two most obvious solutions, but it should be pretty clear by now that all solutions have to involve moving a large amount of empty cars around, which is actually more problematic than simply parking them near their destinations.
At the end of the day, we still need to put self-driving cars that are not in need within cities during the daytime. The number of cars that we need to consider is proportional to the gap between the traffic peak and valley. Now the question becomes: Can we reduce the gap between peak and valley with self-driving cars?
One might argue that with self-driving cars, more people can share vehicles, therefore not as many cars will be needed. So the gap between traffic peak and valley will be smaller.
However, this is a very static view of an existing system. We need to also think about its dynamics. Consider what will happen if traffic becomes better: people can afford to live farther away from cities so they can get more space and lower rent. These options will attract more people to the area. The result? Traffic will keep increasing with the net inflow, until the commute time becomes unbearable again. Actually, there is an economic term for such dynamics: induced demand.
Don’t get me wrong, this is fantastic because with similar commute time, an urban area can accommodate many more people. More people can live in areas with great opportunities, and it’ll spark more social interactions that are crucial for innovations. Just don’t expect traffic to get better in the long run.
And the result of such dynamics? More people means more transportation demand, which turns into more cars on the road during the peak time. The traffic valley will likely to increase as well, but the gap between peak and valley is not going to narrow. They’ll increase proportionally, which means a bigger gap in absolute number.
Therefore, we need to put more cars somewhere in the city during the daytime, and we will face the counter-intuitive effect that demand for parking spaces in urban areas will rise.
With that conclusion, the final question is: How will parking change in the era of self-driving cars?
When cars can drive themselves, where to put them when they are not in use becomes an important question. Ultimately, we want to strategically place them somewhere in cities so they do not cause much more traffic and are close to demands during the day. That requires a real-time system that knows about the demand (where people need cars), the supply (where are the cars), and the parking availability (where are available spaces to put the cars). Real-time data and communication between the three components is the key to make it work.
An important benefit of self-driving cars is that we might no longer need on-street parking. Once we know where available spaces are, cars can drive themselves to parking lots and garages nearby, and eventually save the precious curb spaces for better use.