Paul Godsmark sent me his presentation on autonomous vehicles (Download Autonomes ITE NAS 03 Oct 2012) a couple of weeks ago and he also posted on the ITE Community bulletin board. Our industry seems to be ignoring this tidal wave at our own peril. There's a great article/podcast on Freakonomics about atuonomous vehicles (I listen to all of their podcasts – I'm starting to think I should have went into economics). The Freakonomics article also includes links to thoughtful articles on the topic from The Economist (see this, and this).
An article in Popular Mechanics in the late 80's about driver-less cars piqued my curiosity as a high school student and got me looking broadly into the transportation field when I went to college. I believe the concept goes back to at least the 60's. So I think a lot of folks have filed the concept of autonomous vehicles away with ideas such as colonizing the moon (that idea didn't propel Newt Gingrich to the White House).
Burying the head in the sand really is a poor position by our industry though. As the Freakonomics and The Economist articles point out, autonomous vehicles are coming through an evolutionary process not a revolutionary process. It is very reasonable to assume we will have autonomous vehicles within a decade. Only one or two more dots need to be connected to get there.
How our profession (and ITE) responds to this will be very interesting. Either we jump in and offer our expertise or we will be relegated to the sidelines as mechanical engineers and computer programmers end up taking over the systems we work on. As everyone know, Google just does stuff. They don't wait for a blessing from some large committee or a large government grant.
So the thought experiment becomes, how do traffic engineers and transportation planners add value to a system of autonomous vehicles? A few quick thoughts:
1. Freeway operations – Think of the ramp metering systems we work on to control our freeway systems. Now expand that thinking to being able to divert vehicles to underutilized routes. But also consider the big brother implications – forcing a minority of vehicles out of their way for the betterment of the system.
2. Traffic signals – Adaptive traffic control systems such as the InSync system could do away with the need for our profession to do signal timing/re-timing. Add in positioning data of all of the cars on the system. A common complaint of traffic data used in signal timing is that turning movement counts are limited by the supply given by the existing signal timing. Now in saturated conditions the computers could see the real demand coming a mile away.
3. Safety – Minnesota has done wonderful things with it's Towards Zero Deaths campaign that crosses many state agencies. On the balance, it is feasible autonomous vehicles would lower crashes by an order of magnitude. Statistically maybe even driving crashes close to zero. No more worrying about distracted, tired, or drunk driving. Of course autonomous vehicles would face the same problems as planes or trains – a failure would likely have fatal results.
4. Tolling – It's every economists solution for solving congestion. One can envision dozens of ways tolling could be implemented and we could abandon the gas tax.
5. Road design – Max Donath from the U of MN spoke at the last NCITE section meeting about applying technology to transportation. He's a pioneer in developing ITS solutions including an in vehicle display for snow plow drivers that gives a projection of lane lines and obstacles even in white out conditions. Many of the things he's working on are part of the evolutionary process towards autonomous vehicles. I imagine there are many things we could do with our road design, signing, and marking that would provide better data for the computers.
6. Capacity – With computers driving, some day could we convert two 12 foot lanes into three 8 foot lanes? Reversible lanes are a rarity because of the complexity involved with switching them. With computers that will follow the rules, could the concept be expanded on arterials with heavy commuter patterns?
7. Efficiency – We extol TDM strategies to try to work on the demand side of the supply/demand equation. The hour commute would be a lot more bearable if it felt like you could work in a chauffeured vehicle. Also, I'd personally find the road trip to the Grand Canyon more enjoyable if the car would drive through the night while all of us slept.
8. Traffic Data – Our brains can't understand the large quantity of data we could be collecting from a computerized, autonomous vehicle system. No more setting out tube counters or cameras. Mountains of ADT data and trip generation data could be collected. NAVTEQ and its competitors are pretty close to this destination already. Being a transportation planner in the future may require a computer science degree specializing in databases. The forecasting/capacity analyses currently done in impact studies and corridor studies will look very archaic.
Transportation planners and traffic engineers should be thinking these items through (and many others – I assume my list is only scratching the surface). I'd be awfully nervous if I was a 25 year old, freshly minted PE who is specializing in traffic signal timing.