George,

My old colleague and mentor, Dr. Marty Wachs over at Rand, a couple of years back looked at the supposed congestion reducing impacts of automated vehicles a few years back and predicted not much overall traffic congestion reduction, with predicted growth factored in. But, automated cars will help to reduce the impact somewhat, over as bad as it "could be".

Reducing accidents will do a lot to reduce the variability of travel times, due to less speed differential accident jamming on our freeways during congested periods. Such accidents, even stopping only one lane of traffic, can reduce vehicle though-put or flow rate (number of cars moving past a given point in given time) for the entire segment by up to 70%. Under normal conditions free-flow speeds is about 1800 vehicles per lane/per hour, but with an accident blocking one lane can cause the flow rate to fall by up to 70% across "all" lanes on the impacted segment. These kind of accidents are more likely as traffic density rises (peak periods), and contributes especially to the variability we experience in congestion delay from day to day.

On the other hand, background loading of the freeway, until it reaches just stop and go from excess facility demand will cause break down of speeds to under 25, can reduce through put by up to 50% everyday as we have documented with PeMS counts every day (without accidents) in the Sepulveda Pass on the 405, or the four level in Downtown LA, during the pm rush hour. Since capacity is actually reduced from the reduction in flow rate, caused by slowing from excess facility demand (in realspeak: more cars coming on, than are getting off) and the freeway remains at a reduced flow rate and speed, until the excess demand falls, allowing speed and flow rate to recover.

What automated cars can do is connect people seamlessly to interregional and urban rail transit services, since your "Jonny Cab" will be waiting for you in lines at the station, where you will get off, wave your phone at it, get in, tell it where you want to go, while you text to your hearts content, while the computer does the driving. And, it's finally here, now, as I write this.........

Uber, without the already underpaid Uber driver, since the driver is around 80% of the cost of a taxi, bus, truck, or dial-a-ride on the road today..... Cheap mobility for everyone, but at a potential enormous job loss.... in other words who will there be to pay for the ride? Robotic consumers are in the works next for next year Consumer Electronics Show, maybe?

And, we can't seem get PTC done right on choo-choo's....WTF?

As Rosanna Rosannadanna used to say, "it's always something........."