Doug Newcomb

How Data Can Help Cities Hone Deployment Of Self-Driving Cars

If the vision of the future according to many automakers and companies like Uber and Google is accurate, we’ll see self-driving cars buzzing around cities soon. Ford, for example, plans to have fully autonomous vehicles available for ride-sharing in urban areas in just four years, and Uber has already been pushing the boundaries of the technology in several cities (and its legality in San Francisco).

But a recent study by the connected mobility company Inrix using StreetLight InSight, a transportation analytics platform from StreetLight Data, found that some of the core urban areas that ostensibly would be the best targets for deployment of autonomous vehicles (AVs) may be off the mark. Avery Ash, autonomous vehicle market strategist for Inrix, said that the study is designed to help cities “leverage big data to have a better understanding of mobility as they plan for the impact of autonomous vehicles.”

Ash added that the study started with two questions to help understand car travel in cities. “The first was what are the ideal cities in terms of travel habits to benefit from highly autonomous vehicles deployed in shared-use fleets,” he said. “The second, at an individual city level, was where in the city makes the most sense for the individual mobility goals of that city.”

To answer the first question, Inrix analyzed anonymized data on vehicle trips that began and ended within a 25-mile radius of the downtown area of 50 U.S. cities. It then compared this with vehicle trips that were outbound, inbound or just passing through to establish a percentage of intra-city travel.

These downtown-area trips were also analyzed by their distance, and each city was given a score based on the percentage of intra-city trips and their distance. Ash said the emphasis on shorter trips is because shared-use AVs will most likely be electric vehicles.

“For the top 50 list, we looked at two key factors based on the expectations that autonomous vehicles deployed in a shared-use, on-demand fleet model are really designed to prioritize electric vehicle deployment,” he said.

“In this scenario, where you’re deploying electric vehicles, it comes with some limitations in its current form, namely range and charging infrastructure,” he added. “So the ideal application for these electrified AVs is shorter use.”

Of course, New York will have different priorities for AVs than, say, Nashville. That’s why Inrix wanted to answer a second question geared towards the individual mobility goals of cities.

To get a more granular view of a city’s mobility issues and optimize corridors for AV deployment, Inrix added more layers to the trip data from the 50 cities. To do this it aggregated anonymous data from millions of connected cars, parking availability history, and U.S. Census demographic data to analyze three cities: New York, San Francisco and Austin.

Inrix also overlaid three demographic factors to better map AV deployment: household income, the percentage of residents who are below 17 or above 65 years of age, and what Inrix called “commute mode share.”

Income data was included to help cities ensure that AV technology benefits a broad spectrum of the population, while the census age data was added since “those are populations less likely to have licenses and would most benefit from the added mobility options that would come with highly autonomous vehicles,” Ash noted. Commute mode share reflects areas where “you have a higher incidence of people driving,” added Bob Pishue, an Inrix transportation analyst and contributor to the study.

Let’s pick Austin, since it’s SXSW week and the Texas capitol is one of four cities where Google’s Waymo self-driving car project is testing on public roads. Austin also has major issues with traffic and a goal to provide equal access to mobility options for a wide range of citizens.

The two maps below show data on travel patterns, parking and demographics, with darker areas showing sections of the city likely to benefit most from AVs.

Maps of Austin show data on travel patterns, parking and demographics, with darker areas indicating sections of the city likely to benefit most from autonomous vehicles.

Maps of Austin show data on travel patterns, parking and demographics, with darker areas indicating sections of the city likely to benefit most from autonomous vehicles.

The map on the left is a larger view of Austin and shows a higher density around downtown, and another area north of downtown where Inrix found a high concentration of origins/destinations per zone. The map on the right adds trip and demographics data and show several areas around downtown Austin that are ripe for AV deployment.

The BGC1 area just north of downtown has a trip density twice the study-area average, while BGC2 in the downtown core has a trip density over three times the study-area average. But BGC2 has a lower percentage of people under 17 years of age and over 65 than BGC1 and a lower percentage of households making less than $40,000 per year, which means that AVs may provide greater advantages to people in BGC1 compared to BGC2

Zooming in, the map on the right of central Austin shows that the core downtown area (BC4 and BC5) has an overall higher trip density than the adjacent areas of BC1 through BC3. But BC4 and BC5 have fewer people younger than 17 and older than 65 and households making below $40,000 a year as well as a higher commute mode share. While the BC4 (the busy 6th Street entertainment district) and BC5 (the popular recreational area along the river) would seem well suited for AV deployment, BC1 through BC3 in Austin may make more sense.

“We see this as being an ounce of prevention is worth a pound of cure,” said Ash of the study and its findings. “It is much more preferable for a city to put thought into autonomous vehicle deployment and make sure they’re using this technology as a tool to address their mobility challenges,” he added, “as opposed to dealing with potential repercussions down the line.”


This article was written by Doug Newcomb from Forbes and was legally licensed through the NewsCred publisher network. Please direct all licensing questions to