If New York's taxis used efficient carpooling and ride-sharing, then the city would only need 3,000 of them, way less than the more than 13,000 on the roads today. Those are the findings of a new MIT study, which looks at ways to improve transport in the city by using self-driving taxis.
Imagine for a moment that New York's cabs were replaced by autonomous cars. These networked, self-driving vehicles would be able to route themselves to carry two to four passengers at a time, reducing the number of vehicles needed, without significantly reducing travel time, and stopping a lot of traffic by removing so many cars from the roads.
Using algorithms developed by MIT Computer Science and Artificial Intelligence Laboratory's (CSAIL) Professor Daniela Rus, the 3,000 cars could serve 98% of taxi demand and create a 75% reduction in taxi traffic. If you switched to 10-seater minibusses instead, you could knock that number down to just 2,000 vehicles, with an average wait time of under three minutes. "A system like this could allow drivers to work shorter shifts, while also creating less traffic, cleaner air, and shorter, less stressful commutes," Rus told MIT's Adam Conner-Simons
The MIT algorithm was developed using the New York Taxicab Public Data Set, with data from around three million real taxi rides. Using this data, Rus's team could experiment with changes to "fleet size, capacity, waiting time, travel delay, and operational costs." And because it runs on real data, it is relevant to actual city traffic needs. The problem is getting 3,000 cars to run on the same algorithm, of course.
It would be worth it though. The study says that 1% of the U.S. GDP is wasted on traffic congestion every year. That's around $1.21 billion. People spend 5.5 billion hours a year sitting in traffic, and vehicles burn 2.9 billion gallons of gas while idling in traffic. By switching to on-demand ride-sharing services which use self-driving, self-routing cars, we can clear traffic off the roads, freeing up space for these smart cabs. There's a snowball effect here, too, as clearer roads mean faster journeys, which attracts more people to give up their cars and use ride-sharing instead.
That's the theory, anyway. In practice, we're far slower to adopt anything new. But perhaps millennials will save us. Younger people just aren't buying cars, but they still need to get around, and are already invested in using online services for everything. Maybe we just need to wait for all the old car-owners to die off to free up the roads for better options.