Connected autonomous vehicles enjoy the benefits of needing less reaction time, and thus less spacing between cars than human drivers, especially when they do not need to evade or handle human drivers. Under the industry standard first proposed by Intel, connected vehicles could only need one second for reaction time, but they could need an extra second of reaction time where human drivers and connected cars are mixed.

 

Connected-only freeways could benefit from even bigger efficiency gains by preventing too much traffic from entering at a single point----or too many vehicles entering multiple points which could arrive at a freeway exit, for instance, at the same point and moment----allowing freeways to reliably move at top speeds. This can be accomplished with selective congestion before entering an autonomous freeway----and in the long-run, with more effective measures, like variable pricing for peak traffic. In the future, this could be combined with a new form of autonomous mass transit to transform freeways into moving more people into downtowns than a subway system ever could.

​​

The early and most primitive connected traffic systems are widely expected to use freeway lanes more than twice as efficiently as freeway lanes used by human drivers. With a control algorithm, traffic can be re-directed so that the only traffic allowed on a typical central city freeway is traffic which is either arriving or departing the central business district. A 2016 GDOT traffic study showed nearly a third of Atlanta's downtown connector freeway was thru-traffic, which passes through the central business district rather than arriving or departing in it. A prioritization algorithm could allow most central city freeways, like Atlanta's downtown connector, to carry three times as much traffic into their downtowns as what they currently carry. A new system based around new infrastructure tailored to autonomous managed traffic could bring many times more.​

One of the greatest opportunities connected vehicles could bring to traffic flows is their reduction to reaction time. While humans generally rely on two seconds of reaction time, connected-only traffic could require just one second. But what happens if technology improves and, over time, the standard reaction time drops?

 

Cities without long-ranging plans may not have designed flexible systems capable of handling these changes. This is why Autonomous Symmetry has designed Atlanta’s infrastructure model for autonomous car reaction times all the way down to below half-a-second.

​

Based on history, I do not expect most cities to adequately plan for connected traffic. Most cities are likely to be caught off-guard every step of the way. For the cities which do not prepare, their central business districts could grow more painfully, their physical forms less desirably, and their talent pools less reliably.