Autonomous vehicles, by definition, do not rely on any sensors in the roadway or communications from traffic signals to be able to function.  In fact, all of the technology is on or in the vehicle, so, seemingly, no changes to our infrastructure are needed.  That is (maybe surprisingly) incorrect though.  Here is a list of some of the infrastructure impacts that I believe local governments might consider addressing in the next few years:

• Update traffic signs and markings to ensure they’re meeting the national standard – vehicles will be highly reliant on highly visible and clear signage and markings.
• Eliminate/reduce parking in the urban core and potentially add more parking at remote locations
• Add more drop-off/pick-up locations in front of office buildings, storefronts, transit stations, etc (instead of parking)
• Reduce lane width to increase the capacity of roadways, which will potentially provide added space for bike lanes, and/or improve walk-ability.
• Increase speed limits
• Adjust traffic signals and timing
• Develop new predictive models for pavement maintenance

Some of these activities should happen proactively (e.g., updating traffic signs and markings), but others may happen retroactively (e.g., reducing lane width and increasing speed limits). That being said, most local and regional government planners do not have these kind of infrastructure impacts (not to mention traffic and transit impacts) in their long-range transportation plans.  Those plans address the next 10-20 years and most do not acknowledge the impacts of autonomous vehicles… what will it take for this to change?!