4 ms·
I don't understand the "2 trips for one passenger". The car will pick up the closest human that needs a ride and the operator will optimize the trips globally b
by kjksf 7y ago
I don't understand the "2 trips for one passenger". The car will pick up the closest human that needs a ride and the operator will optimize the trips globally based on where cars are, where humans are and where humans what to go. It's about as optimal as it gets.
Yes, cheaper trips will lead to more trips and therefore more congestion.
But there are also advantages of having a massive car network.
There are actually plenty of roads for most of the time. It's only during rush hour that you have congestion (LA being an exception).
Rush hour traffic happens due to people commuting to work, usually one person in a car.
The density of rush hour traffic makes it easy to optimize.
Without changing cars, we can put 4 people in a car, which is 2-4x better than the common commute of 1 person in a car.
People don't want to share cars?
Implement rush hour pricing scheme that heavily penalizes single occupancy (say, $4 for shared ride, $16 if you want a car to yourself).
The next step would be to make mini-buses for rush hour traffic: at lest 12-16 person. Once you know the traffic patterns, you can optimize by building cars / busses with the capacity that fits the pattern.
What makes it even easier is that rush hour traffic is stable: people go to work and back roughly in the same numbers, which means you can do long-term predictions and investments.
- Johnny555 7y agoI don't understand the "2 trips for one passenger". The car will pick up the closest human that needs a ride and the operator will optimize the trips globally based on where cars are, where humans are and where humans what to go. It's about as optimal as it gets. Unless commute patterns change dramatically, in the morning most people will be going downtown to jobs, there will be a few that live in downtown and go to jobs elsewhere, but by and large, all of the cars will be bringing in commuters from outside of down and dropping them off in downtown centers -- there will be no other human looking for a ride from that car. Without changing cars, we can put 4 people in a car, which is 2-4x better than the common commute of 1 person in a car. We can do that today, but we don't. It's not clear why a self-driving car would change that. How often do people take uber-pool instead of having a car to themselves? (I don't know the answer to that question). People don't want to share cars? Maybe they won't want to, but as autonomous cars become more popular, only the affluent will make that choice - a car is expensive, especially when it sits in a garage at home and work (that you pay for) unused for 22 out of every 24 hours.
- drivingmenuts 7y agoSeems like a reasonable solution would be to create vehicles that give the riders the illusion of privacy. Open cab vehicles, such as modern cars and buses, are terrible at at providing privacy for more than one person. If, instead, you have vehicles with modular, single-occupant “pods”, then it goes a little way toward solving the problem of sharing space with unknown others.