Tuesday, August 1, 2017

Sustainability and the Promise of Driverless Cars

Bottom line. Sustainability is why I'm interested in driverless cars.

And It's why we should all be interested.

We  are, after all, facing existential challenges .  Central to them all is the fact that 5 billion people (give or take) are shortly to join us in the developed world. While a wonderful thing in general, with countless fringe benefits and unseen potentials, it is almost certainly unsustainable with the transportation tools at hand.

And as much as I'd like to sleep at the wheel, or no longer have to look for parking, I wouldn't care about driverless cars if they didn't have the potential to tackle real-world problems.  If this technology doesn't increase our ability to deal with the very real challenges facing us today then I have no  interest in it at all.

Can driverless vehicles, and the driverless system they enable, make a difference? Not just an incremental one, not just a marginal improvement,  but change fundamentally the way we move ourselves, and our things, across the face of the earth?

Here's why I think they can.

To a greater extent than trains, or horses, perhaps more than every form of transportation since walking, every aspect of modern life is accessible to the automobile. Because of this a fundamental change in this one piece of technology has the potential to change society significantly and at every level. and what more fundamental change could be made than to replace the guidance, communication, and control system with the most powerful modern technology, the computer?

One of the things that astonishes me about the conversation centered on driverless cars is how limited is in scope. Especially as this appears to be a potentially significant change to a system which is integral to almost every human endeavor.  The title of this blog was chosen to reflect the narrow scope of the conversation surrounding a technological change in an earlier era, one that saw the internal combustion engine as merely the replacement of one method of propulsion with another, somewhat less reliable one. This is the effect that the replacement of horses with the internal combustion engine had. And hardly is the ink dry on the change this minor adjustment has wrought than another change is upon us.

It is, however, a question that will likely be in the minds of those individuals that attempt to adapt this tool to deal with the very real problems we will face in the coming years. Those that focus exclusively on the technological, legal, and social challenges should not lose sight of the larger goal: a sustainable, perhaps even sustaining, transportation system.

Saturday, May 20, 2017

WSF? How Driverless cars will transform the Washington State Ferry System.

Driverless vehicles have the potential to reestablish the Washington State Ferry System to its historical role as a critical regional transportation network - halving trip times, doubling or tripling the capacity of the existing system - all without significant changes to boats or docks. Ferry facilities will take up less than 20% of the real estate they presently require when driverless vehicles become the norm.

All of which means that the Washington State Ferry system is an ideal example of the enormous disruption driverless vehicles are set unleash on regional transportation infrastructures.

The System Today 

Each region has its own set of unique transportation challenges, but few are as daunting as those facing the managers of Seattle/Tacoma's regional transportation network. Most serious are the numerous choke-points created by hills, canals, lakes, and, finally, by Puget Sound itself.  Moving north to south is difficult enough on the region's two (highly congested) primary freeways. But moving east to west, especially during rush hour in the Seattle urban area, is almost impossible.


These large inland bodies of water, the results of a great glacial excavation project of previous millennia, were once the primary regional transportation network. While this inland marine waterway has made Seattle/Tacoma a primary ports for trade with Asia, it also constitutes an effective limit on Western Washington regional development. 

All this in spite of the fact that The Washington State Ferry system is the largest in the US (3rd worldwide), and moves almost a million cars every year. But for all that size it serves relatively small communities on the Western side of the sound, and supports little of the economic activity between them. It seems counterintuitive that driverless cars should have any significant effect on the function of a marine highway but in this case, like many others, drivers vehicles have an outsized potential making existing infrastructure operate with an entirely unprecedented dynamism.

To understand why this is so it's first necessary to understand how the existing infrastructure operates. The ferries are fast (typical transit time is not much more than 30 minutes) and large (up to 2500 passengers and 200 cars. Overall system capacity is impressive with 22 boats carrying up to 30,000 passengers every day.  Loading times for foot passengers are quite short and anyone arriving at the terminal 10 minutes before departure he would be assured of a spot on board. But any use of transit at either end adds a significant level of uncertainty. A driver, however, would have to budget at least half an hour more at the front end ( if all goes smoothly), 20 minutes more at the backend (for unloading), and significantly longer during rush hours, holidays, and game days. 

Attending Thanksgiving celebrations on the other side of Puget Sound (with family on Bainbridge Island) will include a nearly three hour return journey (2 hours in the car line, boarding, boat ride, disembarkation, ride home) to cover the distance of little more than 15 miles. My brother-in-laws daily commute is more typical - 10 minutes bicycle ride the front end, 30 minutes on the boat, and another 10 minutes bike ride on the Seattle side - a bit more than the national average but not by much.

Not that driverless vehicles will not make the boats go faster.  But the typical ferry route involves more time at the terminal then underway. And the terminals themselves represent almost as significant an economic investment than the ferries they serve. It is an examination of the potential to utilize driverless vehicles in the terminal-ferry interface is that gives us insight into the transformative potential of this new technology.

Loading walk on passengers is easy - an 8 foot wide gangway gets literally hundreds on board in minutes.  Cars load much more slowly however. When all goes well  the two lanes that feed the open ended ferry will allow it typical ferry to load 200 cars in under 20 minutes (@ 18/min). this rate is largely influenced by the fact that the process must be done at a speed which accommodates drivers of all skill and experience levels. The mix of vehicles is heterogeneous, and there is very little opportunity to organize them during the loading process. Unloading is faster (subject to traffic congestion at the destination), but not significantly so. 

The System in a world of Driverless Cars

Loading autonomous vehicles would be different in at least three significant ways; 

-The process would be more highly organized with vehicles moving in a highly predictable matter including higher speed and smaller spacing between vehicles. The final load configuration would be more uniform and significantly more efficient. Even without other advantages this would result in at least 25% time savings at the terminal.

-Vehicle loading and unloading could be synchronized. There are (at least) two lanes available and loading and unloading would be simultaneous. Time-saving here could be assumed to be significant as well, perhaps as much as another 25%.

-Passengers loaded directly onto the ferry without use of intervening modes. During the boarding/unloading process vehicles could board the ferry, discharge passengers, and leave the ferry, without significantly affecting turnaround times.

The halving of turnaround times at the terminal would result in at least a 25% reduction in trip times and increase of the number of trips at the ferry could take on a given day by about 25% to 35%. Most importantly the overall system capacity would see corresponding increases. Other potential improvements to ferry system infrastructure deriving from the development of autonomous vehicles include;
Reduced vehicle size and weight associated with driverless cars increase ferry loading.
Just in time arrival of vehicles to the ferry reduce the need for large holding areas.
Greater reliability leads to increased commercial use of ferries

The most significant effect in the long run, however, would derive from the effect of driverless based system on the passenger experience. The average car and transit user would see their overal commute time of well over an hour ( 30 min pre-board, 30 min trip,15 min off-load) decrease by at least 50%. Greatly increased reliability and shorter trips will attract greater numbers of users to the system, reduce the costs per user (or lower existing tax supported subsidies) and lead to an increase in regional economic activity. In short, the marine highway transformed into a Freeway, turning one of the major constraints on regional growth into a catalyst for sustainable development. 

Looking into the future often suggests a return to historical patterns. Before the advent of the automobile and an efficient highway system it was the Puget Sound's "Mosquito Fleet" of steamships that tied the growing towns of the region together. It should not be surprising then that this unique inland waterway will continue to play a significant, even expanded role during this next phase of development. Although it's certainly too early to begin working out the details of how to harness driverless technologies, the time is coming when a preliminary study should be done. System managers might be surprised at what they find. 

Saturday, January 28, 2017

While Level 5 automation is a distant dream.

Level 4 is already on the Road

An old story - a new technology, a shiny new toy, and the promise of a boundless future. We want this new technology to break the bonds that limit our present world, to be complete in and of itself. But history suggests that without a proper road the future will never arrive. The driverless car is only half the story. The other half is the story of the road.

For a great discussion of the difference between levels 4 and 5 of automation in driverless vehicles it's hard to beat Gill Pratt's speech at CES.

                 "at a level V automation the car can drive fully autonomously under any
                 traffic or weather  conditions, At any place at any time. No human required."

Level IV 4 on the other hand limits autonomous operation to well map areas under nominal driving conditions. Pratt's speech makes it clear that level 5 is a long way off, though it remains the declared goal of every carmaker interested in driverless vehicles.

It makes sense that carmakers would look to their product as the solution to all the challenges facing automation. Carmakers build cars, laborers build roads. History suggests, however, that no emerging technology ever fully solved the challenges it was designed to overcome. The world has always had to be rebuilt to take advantage of new innovations, and reflects their limits as well.

Imagine the challenge facing the first automobiles. A world adapted to the needs of horse-drawn transport - vehicles that were simple, rugged, and not expected travel much faster than ten miles an hour. Signage? Traffic control? Not when you are being pulled by a semi autonomous creature with his own collision avoidance system. And the first cars looked very much like the vehicles they were replacing and their name - horses carriage - reflected that.

with roads like these...........

It wasn't until the 1930s that infrastructure was built to take full advantage of the potential of the automobile. Broad, straight freeways allowed flexible mobility on a regional scale that would've been unimaginable 30 years before. The investment was an enormous, but the return was (for better and for worse) the modern world.

The difference  between the world of 1910 at 2010 is that the infrastructure that exists today is much more closely matched to the infrastructure needed for this new technology. Initially, at least, driverless vehicles will be almost identical to existing cars in size, weight, and speed. The roads are large, reasonably well marked, and increasingly well mapped. The challenge is more political and regulatory than technical. All that is required is the carving out of a lane here, a dedicated pathway there, and a system for connecting and coordinating traffic signals. Challenging enough in truth, but not on the scale of building an interstate highway system or a transcontinental railway.

Forget level 5, lets just build a road we can use

Those first few bits of infrastructure dedicated to driverless the vehicles will be like a camel's nose under the tent - a beginning that will be reasonably easy to expand upon. Before we achieve the goal of creating a car that can drive itself at any place, and at anytime, the world will become much more accommodating to autonomous transport. The discussion of level 4, or level 5 automation, will become an historical footnote.