I’ve already written about this before. Namely, rideshares like Uber/Lyft increase congestion in cities – not alleviate it. Mostly because these services put people that would have likely taken public transport/walked/biked into a car instead. It doesn’t reduce driving either since they’re just swapping one car on the road for someone else’s. Since then, there have been numerous other studies that have confirmed this effect.

A new study by EHT – Zurich’s Institute for Transport Planning and Systems seeks to find out if that equation changes if autonomous taxi services were introduced. Using an agent-based simulation on the city of Zurich, they tested a number of scenarios (service models, owner models, etc), examined costs, and how disruptive the shift would be. The simulation uses an agent-based system in which individuals make decisions based on time/cost/etc, instead of overarching rules. This has previously produced really accurate studies.

Turns out, the impact is not as much as people expect, and the fleet to do it actually must stay relatively small when paired with existing, good quality mass transit systems to be viable. This doesn’t shock me in European cities with great public transit, but I wonder how it would play out in American cities without that infrastructure. My gut tells me it likely would be the same as Uber/Lyft study results.

Either way, the study is definitely worth a read:
https://www.ethz.ch/en/news-and-events/eth-news/news/2019/06/driverless-congestion.html