Hacker News Reader

Comment by joshuahedlund | Hacker News Reader

Why is "reducing air pollution" a potential application of driverless cars? The other benefits allow people to driver faster and safer on the same amount of roads, but I don't see how that would encourage less driving that would lead to less air pollution.

Faster driving = less gas mileage, and if driverless cars dramatically lowers the opportunity cost of driving people may be encouraged to put even more miles on their car. Wouldn't that all increase air pollution? Don't get me wrong, I think the other benefits are enormous, but am I missing something obvious?

Edit: Thanks for the great answers below.

View on HN

Self-driving cars could form convoys to utilize slipstream. You could reduce the safety distance, because the reaction time compared to a human driver is massively reduced. Additionally, the driving style of the average driver leaves a lot of room for improvement.

How big are those slipstream effects compared to speed scaling? Drag power goes as the third power of speed; take average urban speeds from say 30 mph to 300 mph, and the drag goes up a factor of 1,000x -- to on the order of a megawatt.

Relatedly, the same thing that gives you slipstream (lower following distances at a given speed) increases car density on the road at any speed, which increases road capacity, which decreases congestion and idling. My guess is that this effect, not the aerodynamics, is the biggest benefit.

Also note that self-driven cars can park themselves and shut off the engine to avoid congestion instead of simply sitting in a slow-moving queue at a "metering light".

Less time sitting in traffic due to more efficient road usage could help maybe?

Also if self driving cars became the norm, I could see a lot of automobile usage becoming similar to a "taxi" situation where the nearest car picks you up, and you could opt for an economical plan where the car picks up passengers going the same way.

At this point it's just speculation because nobody knows what kind of novel usage patterns and macro effects we'll see if driverless automobiles become the norm. We just know it will be a huge net positive in the long run. If only for the millions of truck drivers who would be unemployed and hopefully quickly re-employed doing something else useful.

Its quite true that driverless cars per se would have just the effects you mention. But if driverless cars mean that suddenly nobody owns a car and just rents a car from the automated taxi fleet when they need a ride we'd probably see most of them be lightweight single occupancy vehicles, meaning less pollution overall.

Every time I see someone explain why they need a car they always mention big grocery trips and often taking kids somewhere or other. It doesn't seem like driverless cars will do away with those, and if they're really so necessary, most cars available probably won't be lightweight and single occupancy.

I'm sure that there will still be plenty of multi-passenger cars.

But to your point, most people who drive trucks actually use the bed maybe a dozen times a year, and many people who have high occupancy vehicles take far more trips with only themselves or 1 additional passenger than they do fully loaded.

Single occupancy vehicles aren't practical for most people because they need a larger car some of the time, but if you don't own it you're free to request whatever vehicle you need right then.

Going to work by yourself--lightweight single occupancy. Picking up the kids--minivan.

Right, I just realized I forgot another important effect if we go the autonomous taxi route. Currently we have somewhere around 5 parking spaces for every car. I bet we could at least cut that in half with autonomous cars, making it reasonable to develop more densely, meaning shorter driving distances, and making trains a more attractive option for some trips.

> Currently we have somewhere around 5 parking spaces for every car.

Not in the places with lots of population density.

For example, SF may have more cars than parking spaces. It certainly doesn't have 5x as many.

> I bet we could at least cut that in half with autonomous cars, making

almost no difference.

For example, I have two parking spaces. One on the street in front of my house and the other in my backyard. The latter isn't going away and the former won't make any difference.

Most people choose public transportation to avoid the agony of commuter traffic (mine is ~2hrs per day). Not only is it time consuming but you can't do anything else during that time other than get bored. If I had a self driving car, I could work during the commute, making the incentive for taking the train much lower.

It would take much analysis and user studies to see if effects like this that encourage people to drive more or car sharing effects will dominate. But reducing air pollution definitely is not a foregone conclusion.

One more: the taxi fleet model (fewer cars) makes it easier to replace cars with newer models, which tend to include more environmentally friendly technology.

A lot of fuel is spent on accelerating the vehicle, much more than to sustain the speed. That is why you spend more gas driving in the city with jams and traffic lights. If every car was driven by a computer, they could in theory form a network and do for example, the perfect meshing on the intersections without stopping, only accelerating or deccelerating a little. Computers could also drive big traffic slowly but steadily without lots of starts and stops.

Reply on news.ycombinator.com