Time for Autonomous Vehicles to Disrupt Transportation Planning
BY MARK McDOWELL
Over the coming decades autonomous vehicles (AVs) will insinuate themselves into our society and the accumulating affects will be almost as significant and ubiquitous as the introduction of the automobile itself. They will cause us to change our thinking about everything from street design, to transit, to urban design, highway design, inter-city transportation, architecture, and regional planning. Yet, though this highly disruptive technology will be introduced well within the planning horizon of organizations charged with laying out our future, there is insufficient cognition about its implications. The purpose of this article is to summarize the development status of these vehicles, imagine some of their “second order” affects, and consider implications for infrastructure development and planning.
In discussing autonomous vehicles we distinguish between semi-autonomous and (fully) Autonomous Vehicles. Semi-autonomous vehicles can operate without a driver in constrained environments and conditions, such as on an interstate or only at certain speeds. By (fully) AVs, we mean vehicles that can operate in driverless mode in essentially all environments, perhaps excluding only unusual weather events. These vehicles can read normal traffic signs and respond, identify bicyclists’ hand signals, sense pedestrians and so forth. In general, they can operate on city streets without those streets being substantially modified.
For years automakers have been enhancing their vehicles with “intelligent” features such as parking assistance or Mercedes’ well-advertised autonomous braking capability. In 2015, Cadillac will introduce an “adaptive cruise control” feature which will allow the vehicle to automatically speed up or slow down depending on the vehicle in front, and in 2017 will add a self-steering feature at speeds in excess of 70 mph, as well as the ability to communicate with other equally-equipped vehicles. It is reasonable to expect that other carmakers will follow in short order and similarly enhance their products’ capabilities. Lux Research estimates that as of today, about 3 percent of the U.S. automobile fleet is equipped with some kind of limited self-driving feature. This should rise to 92 percent by 2030.
The semi-autonomous car is just a step toward the fully autonomous vehicle. All major automakers—as well as Google, Tesla, various university engineering departments, and any number of unknown garage-based “tinkerers”—are working on either components or complete systems for AVs. Mercedes, Audi, and Google are currently testing fully autonomous vehicles in California and Nevada. Nissan claims it will have a self-driving car on the road by 2018.
There is also broad interest in equipping trucks with autonomous (frequently termed “robotic”) capabilities. All the major European truck manufacturers are testing, under highway conditions, “platooned” trucks, where the lead vehicle has a driver followed very close — too close to be safe if under the control of humans — by additional units not requiring a driver. Rio Tinto, the international mining company, is using ten self-driving, no-human-on-board, ore trucks at an Australian mine and plans to expand the fleet to one hundred fifty.
There is also great interest on the regulatory front. A number of states have established regulations for testing autonomous vehicles and more are developing them. The National Highway Traffic Safety Administration (NHTSA) has announced a potential rulemaking requiring that all automobiles be capable of communicating with each other. Recognition of the benefits of autonomous vehicles, and their inevitability, isn’t lost on other developed countries. The United Kingdom just passed regulations governing the operation of autonomous vehicles and Holland has announced the goal of making the entire country safe for autonomous vehicles—particularly trucks—within five years.
Driving interest in AV technology are three factors: safety, fuel efficiency, and better utilization of existing infrastructure. It is estimated that over 90 percent of accidents are caused by some form of driver error and cost the economy $300 billion annually. As penetration rates advance, autonomous vehicles could practically eliminate accidents while improving fuel economy by up to 25 percent due to better breaking and acceleration; and increase road and highway capacity resulting in less congestion.
But these are just what we might call “first order” affects, analogous to the initial benefits derived from horseless carriages: lower operating costs, faster transit times, and, from a societal standpoint, less urban pollution from manure. But as horseless carriages morphed into automobiles, they began to affect society in ways beyond those anticipated when the family car was just a promise. We call these “second order” affects.
In the case of autonomous vehicles some of these might include the following.
Demand for enhanced infrastructure
Benefits from autonomous vehicles grow faster as the percentage of surrounding vehicles increases. Constituents are therefore likely to demand infrastructure enhancements, such as dedicated lanes or dedicated roadways that maximize vehicle utility.
In the freight sector, “platooned” autonomous (and semi-autonomous) trucks, in addition to safety and greater fuel efficiency, promise substantial economic benefits to their owners. Platooned vehicles will dramatically reduce driver costs, which today represent around 35 percent of per-mile operating costs. Passing a platoon of trucks, which could be hundreds of feet long, however, will to face resistance from the driving public. The question then arises whether, in some corridors, economic benefits would be sufficient to make construction of separated, autonomous-trucks-only roadways viable. Such roadways could also cut down on congestion which costs the trucking industry over $9 billion annually. Given the fact that semi-autonomous trucks are already in use, it would seem that state transportation departments ought to be thinking about whether such roadways are warranted, and even initiate early stage planning.
Less congestion or need for greater capacity?
A key question in thinking about autonomous vehicles is what percentage of the public will continue to own private cars. With sufficient AVs on the road, mobility will never be more than a few minutes away through a smart phone app that optimizes personal preferences and tradeoffs over available vehicles. Many people will of course continue to own cars, but with the enhanced mobility of driverless cars we can imagine that many will not. There could be many more models of vehicle ownership/sharing than current technology affords. Some people will own and drive exclusively, others might own a vehicle for commuting, for example, but make it available the rest of the day through an Uber-like network, others will rely on chartering, joint chartering, ad hoc ride-share, taxi-like services, and so forth.
The consequences for planning of how and how many people make these decisions are enormous. It is estimated that for every 5 percent decrease in automobile ownership, vehicle miles traveled (VMT) will increase 10 percent. Offsetting that is the estimate that AVs should dramatically increase street and highway capacity. Should planners not be asking what the net affect would be on which corridors?
Affects on urban form
It is estimated that one shared AV could replace 11 automobiles. The implications for parking requirements and thus urban and suburban form are obvious. Or, if the majority of customers arrive in driverless vehicles, then do we need on-street parking? Do curbs become just a nuisance?
Affects on suburban form: no end to sprawl
With increased mobility producing higher VMT, there is no reason to believe sprawl will cease. The economics of suburban real estate combined with lower mobility costs due to not owning a car, should encourage continued robust suburban growth.
Affect on transit: More service for less money
AVs will scramble our current notions of transit, particularly in low density, suburban settings. Today a 40-passenger bus costs about $300,000. Many fewer than 10 AV’s, on average, would be needed to equal the capacity of one bus – and probably at significantly lower cost. It would therefore make prima fascia sense in limited demand settings to transition from buses to AVs and offer door-to-door service, faster transit time, and greater comfort than a bus. Additionally since AVs are perfectly suited to taking small numbers of people to diverse destinations, a transit agency untethered from buses and rails should see a huge growth in demand from former automobile owners. Except that the private sector may provide the service first.
Affect on paratransit
The driverless vehicle would greatly enhance mobility for the disabled, especially if concepts such as the one below are developed further.
Affect on Inter-City Transportation
AVs have implications for intercity transportation as well. Imagine vehicles that will pickup passengers in front of their house, head to the appropriate interstate (AVs only of course) and insert themselves into a passing platoon of vehicles travelling at say 115 mph. Given such capabilities, developing AV-exclusive roadways might be more cost effective than building “higher” speed (maximum 115 mph) rail. Certainly they would be more convenient. Why would anyone get in an AV just to go to the train station when they could get to their destination faster and possibly less expensively using an AV all the way? Perhaps it would be wiser for state DOTs think about exclusive throughways for autonomous vehicles rather than plan rail projects that are 10 or 20 years out.
Policy makers are not anticipating, and certainly not planning for, the secondary affects of AVs. For example, metropolitan planning organizations (MPOs) are required by the federal government to prepare a long-range transportation plan (LRTP) that looks forward 20 to 30 years. This is within the time frame when a substantial portion of the automobile fleet is likely to consist of fully autonomous vehicles. I doubt the current LRTP of any MPO takes this into account. As a consequence, we are designing infrastructure for yesterday instead of tomorrow. Since it can take 15 years between inception and completion of a federally funded project it is entirely possible—given the rapid pace of AV technology development—that by the time a major project designed today is completed, it will be inappropriate, if not obsolete.
While the affects of autonomous vehicles are unclear, what is clear is that they will begin affecting public demand for infrastructure and attendant funding well within common planning horizons. Granted, a good deal of analysis needs to go into the impacts of AVs before an LRTP can prudently incorporate them. But the relevant work needs to get underway and commonly used assumptions about future settlement patterns, corridor traffic density, available transportation options and the continued viability of various transportation modes needs to be challenged. The next transportation reauthorization bill should encourage state and local planners to take this on-rushing technology into account and offer the hope that a thoughtfully constructed future is not delayed by our inability to imagine it.
The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of The Eno Center for Transportation.
 The National Highway Traffic Safety Administration (NHTSA) has outlined five categories of vehicles on a scale of increasing autonomous functionality. But for infrastructure planning purposes the simple distinction proposed is sufficient.
 Jones, Willie, “Self-driving cars Will Be $87 Billion Market by 2030”, IEEE Spectrum. Retrieved from https://spectrum.ieee.org/cars-that-think/transportation/self-driving/self-driving-cars-market-2030 on August 2, 2014
 Ross, Phillip E., “Nissan’s Ghosn Now Says Robocar Sales Could Start in 2018”, IEEE Spectrum, retrieved from https://spectrum.ieee.org/cars-that-think/transportation/self-driving/nisssans-ghosn-now-says-robocar-sales-could-start-in-2018 on July 22, 2014
 Maverick, Tim, “New Robo-Trucks Poised to Revolutionize the Mining Industry”, WSD, Retrieved from https://www.wallstreetdaily.com/2014/05/21/mining-robo-trucks/ on July 17, 2014
 Gabriel Weiner and Bryant Walker Smith, Automated Driving: Legislative and Regulatory Action, Retrieved from cyberlaw.stanford.edu/wiki/index.php/Automated_Driving:_Legislative_and_Regulatory_Action on November 18, 2014
 Jeffery, Terrence P.,” DOT Proposes Mandating Cars Broadcast Location, Direction and Speed”, Cybercast News Service. Retrieved from https://www.cnsnews.com/commentary/terence-p-jeffrey/dot-proposes-mandating-cars-broadcast-location-direction-and-speed on September 18, 2014
 Jones, Willie, “The UK OKs Self-Driving Cars on its Roads”, IEEE Spectrum, Retrieved from https://spectrum.ieee.org/cars-that-think/transportation/self-driving/the-uk-grants-licenses-to-selfdriving-cars on Aug. 11, 2014
 Ross, Phillip E., “Dutch Trucks Will Drive Themselves”, IEEE Spectrum, Retrieved from https://spectrum.ieee.org/cars-that-think/transportation/self-driving/dutch-trucks-will-drive-themselves on July 17, 2014
 Fagnant, Daniel J. and Kara M. Knockelman, PhD., P.E., “Preparing a Nation for Autonomous Vehicles”, Eno Center for Transportation, 2013, pp. 3 – 8
 op. cit., p 8
 Torrey, W. Ford IV and Dan Murray, “An Analysis of the Operational Costs of Trucking: 2014 Update”, p 13, America Transportation Research Institute, Arlington, VA
 Fender, Katherine J. and David A Pierce, “An Analysis of the Operational Costs of Trucking 2012 Update”, America Transportation Research Institute, Arlington, VA, p 3
 Jafffe, Eric, “City Lab”, “Imagine: A World Where Nobody Owns Their Own Car”, Retrieved from https://www.citylab.com/commute/2014/02/imagine-world-where-nobody-owns-their-own-car/8387/ on Feb 16, 2014
 Loc. cit.
 MacKechnie, Christopher, “How Much Does A Bus Cost to Purchase and Operate?”, Retrieved from https://publictransport.about.com/od/Transit_Vehicles/a/How-Much-Does-A-Bus-Cost-To-Purchase-And-Operate.htm on October 13, 2014