Smart Cities


Cities can prepare for autonomous vehicles now, say Bern Grush and John Niles

There is no shortage of mass media excitement about the arrival of autonomous vehicles (AVs). A lot of it is hype. The majority is frivolous.

There are many projections of how these vehicles will be used: “Send them circling the block to save parking fees while you shop!” or “AV taxi rides will be so cheap and convenient that no one will own a car!” are two that have been oft-repeated. The common ethos is that these innovations have properties that we will be eager to consume as they arrive. This is marketing and we are falling for it. Accordingly, AVs are projected to make dramatic impacts on our cities and roads regardless of whether they constantly circle the block or if no households own them but ride services do. A January 2019 research report from consulting firm McKinsey & Company, How China will help fuel the revolution in autonomous vehicles projected that by 2040 66 percent of the total automotive passenger kilometers travelled (PKT) in China would be in AVs and that 55 percent of passenger kilometers (PKT) would be in AV service fleets, reflecting that the overwhelming majority (83 percent) of AVs would be non-household vehicles.

We are not questioning the feasibility of McKinsey’s dramatic projection for China. In our textbook The End of Driving: Transportation Systems and Public Policy Planning for Autonomous Vehicles (Elsevier, 2018), we project for North America that robotaxi service fleets will serve 50 percent PKT in the period 2032-2063, a broad range of future years reflecting important unknowns.

“It is currently unlikely that any city would be ready for such a deep penetration of AVs without significant effort and that seems only remotely likely from today’s perspective”

More importantly, however, we assert it is currently unlikely that any city would be ready for such a deep penetration without significant effort — effort that seems only remotely likely from today’s perspective. Without knowing the final capabilities or arrival time of pervasive vehicle automation or whether these future fleets might be majority private or majority shared, can we anticipate, define, and prepare for the beneficial changes we want from AVs for our cities and regions? Can we identify the tasks and metrics needed to intentionally channel AV technology toward enhancing urban liveability rather than simply consume what will be on offer? Can we define targets for 2030? For 2040? For any significant planning horizon? We suggest that each urban planning entity consider the local AV potential and some related metrics. Starting with high-level descriptions, tasks could be based on some of the liveability goals often associated with what is believed to be achievable with fewer private automobiles within a broad planning domain. The task descriptions one city or region might set would differ from those of another. Current social preferences, ownership levels, weather and climate, planning budgets, and the like would influence how these tasks are expressed and what would be the desired target levels of their metrics. Certainly cities such as Beijing, Copenhagen, or Houston would express and measure their goals very differently. Their planning authorities can think through these local contexts.

Municipal, regional, and in some cases state governments would influence the reach and boldness of such forward thinking — there will be regulations and budgets to consider. And that is fine — this starts as an exercise before it needs significant budgets.

“It is currently unlikely that any city would be ready for such a deep penetration of AVs without significant effort and that seems only remotely likely from today’s perspective”

Here are four sample tasks for consideration, representing an expansion of concepts that we lay out in our book.

1. Specify the characteristics of attractive communities where household vehicles become options rather than necessities.

Many cities are working on transit-oriented developments, aimed at reducing car use within a half mile (800 meters) radius of significant transit stations. But the connecting street grid within these centers typically allows private vehicle entry, even if parking and speeds are limited. Furthermore, such TODs represent an insignificant portion of the land in urban regions, and private vehicle use is generally overwhelmingly important in the full urban geography of cities, from NYC-NJ down to micropolitan regions. With few exceptions, registered vehicle populations and VKT continue to rise after the pause for the Great Recession. How could using future common carrier passenger fleets of AVs influence the spread of car-free communities within more cities, to extend over larger areas, and to become a regional or national norm rather than an exception as they are today?

How could your city, your transit system, and your streets to be set up for this to become an everyday choice for an overwhelming majority of trip-takers in an urban AV future? When someone says (or believes): “no-one will own a car”, consideration of how to reorganize kerbs and sidewalks, fleets and energy distribution systems needs to be undertaken. Reducing the requirement for parking spaces for multi-unit residential properties is a start, as is facilitating active modes like electric bicycles. Highly consequential is the logical fact that on the way to the fully automated “no-one-owns-a-car”-world a region must pass through a very difficult world of “only-half-the-cars-are-automated”.

All of these issues can be considered now, regardless of uncertainty about how fast AVs will arrive and how pervasive they will become. In fact, this task should be undertaken even if AVs never arrive. It could be started now. It would be a shame to be unprepared for this if indeed AVs did arrive.

A metric for this task might be 50 percent reduction in regional, private household vehicle use by 2040. Use might be measured as in VKT/VMT, PKT/PMT, or per capita ownership. You choose.

2. Plan for the organization of large contiguous service areas for robotic shuttles, taxis, and delivery systems.

Do this to increase reach and access, as well as connectivity with existing fixed-route transit while lowering the relative utility of owning a personal vehicle — automated or not — for those living in this service area.

This cannot happen automatically or by default. It will not happen under the aegis of the automotive manufacturers or fleet operators; there is too much infrastructure involved. And the commercial providers — both car manufacturers and fleet operators — earn greater dividends if they can solve inefficiencies at travelers’ expense and with tax-payer subsidies, as our current automobile and transit systems operate today.

A meaningful metric might be 50 percent of urban passenger travel and goods mobility in zero-emission driverless taxis, shuttles, and delivery vans by 2040.

“MaaS Global looks for cities where there is an open mindset and the willingness of the local people to use public transportation and other modes of mobility rather than simply relying on the use of one’s own car”

3. Consider how to define and organize pervasive AV service fleets focused on choice, variety, reliability, acceptable comfort, and behavioral nudges for large car-free communities.

Do this to increase the probability that a majority of travelers in a region can find satisfactory trip services to a degree that owning a personal vehicle becomes increasingly less attractive. This would be the case if trip services were desirable, prompt, clean, available everywhere, and went anywhere within a comprehensive service area.

We suggest this be a specify-and-regulate approach rather than an own-and-operate play. There is no reason for government transit agencies to contemplate purchasing large fleets of robotic passenger vehicles.

A potential metric could be that the portion of regional families that own personal vehicles by 2040 would be half of the 2020 figure. This is a dramatic change. Waiting for more AV-innovation signals takes time away from planning that should be underway now.

4. Take the steps to organize and regulate kerbs and sidewalks for massive and mixed fleets of automated and manual ride-hail and robo delivery systems.

Do this because the focus of frustration will be at busy kerbs where vehicles jostle for space and compete with other modes such as transit, bikes, pedestrians, and goods delivery in addition to each other.

Kerbs and sidewalks are already a focus of trip hassles due to competition for parking, contention among modes, as well as payment and enforcement systems. This is certain to get worse, but local authorities can begin to address it now, as illustrated by some jurisdictions establishing the mandatory use of designated kerbside stopping areas for ride services. In some cases public transit bus stops have been authorized for use by transportation network companies like Lyft and Uber making brief stops, like transit buses do. Naturally, this will be a development and learning process, but waiting yet another decade to address shortcomings at kerbs and on sidewalks would be a mistake.

A possible metric could be 40 percent of regional streets, kerbs, facilities, neighbourhoods, and populations served efficiently and effectively by automated, on-demand vehicles for passengers and goods by 2040.

“Waiting for car manufacturers and ride-hail operators to decide the future of urban AV deployment will not create the cities that urban planners hope for”

All of this can start now — planning processes aiming to achieve stated quantitative metrics will be demanding and take time. Some aspects can be executed now as they are already overdue. There is enormous value to be gained immediately and most of this value could be retained even if automated vehicles never arrive. Community leaders need to stop focusing on technical novelty, AI-promises and conundrums, vehicle consumption, corporate competitions, and markets. Leaders need to focus on cities, regions, people, communities, congestion mitigation, and use cases. We need to think harder about the pending effort and effects of shifting 50 percent of private vehicle use to automated vehicle use. A 50 percent shift from household vehicle ownership to on-demand AV use offers the potential for immense urban land recovery. In 2017 there were 288 million motor vehicles in the US & Canada. According to an FHWA forecast, the current annual growth rate of vehicle registrations is 1.1 percent implying that by 2040, there would be 371 million motor vehicles in US and Canada assuming nothing else changes. Assume 85 percent of these motor vehicles (316 million) are household vehicles, and all other variables such as VKT per person, daily trip counts, trip distances, average occupancies, etc. remain constant. To simplify further, momentarily ignore mode shifts to or from active transportation or transit.

A 50 percent shift from vehicle-ownership to relying on mobility services implies dividing 316 million household vehicle equivalents in 2040 into two halves: • 158 million will be a private, household fleet delivering 50 percent of trips, similar to now. • 20 million will be an AV service fleet delivering the other 50 percent. (158/8 = 20 million. The service fleet is assumed to average eight trips per vehicle-day). Assuming the average, private motor vehicle uses five parking spots, 138 million fewer vehicles (158 million minus 20 million) frees up at least 690 million parking spots x 32m2 or 22,080 km2 of the most valuable real-estate in North America. That’s slightly larger than the landmass of Massachusetts. Waiting for car manufacturers and ride-hail operators to decide the future of urban AV deployment will not create the cities that urban planners hope for, and often work very hard to make happen. While significant penetration of AVs — private or shared — is likely a decade or two away, deferring directional, optimization, and liveability strategies will rob cities of flexibility, influence, and degrees of freedom within a decade. If you believe AVs are coming eventually, the time to start getting ready is now, even if you believe human drivers will remain dominant for many decades. The steps outlined here are important support for the alternative to SOV, of expanding mobility-as-a-service protagonists such as Uber and Lyft.

Bern Grush is the Chief Innovation Officer at Harmonize Mobility, Inc. He is an author, innovator, and entrepreneur in the fields of autonomous vehicles, parking reform, and road pricing. Bern is 25% Human Factors Engineer, 75% Systems Design Engineer and 100% dedicated to making sure we do not squander the opportunity to do good things for our cities and our planet when deploying autonomous vehicle technology.

John Niles is the Western Regional Director for Harmonize Mobility, Inc. and a Research Associate at Mineta Transportation Institute. He is currently engaged in developing practical tactics and strategy for melding transit-oriented development, commuter parking, automated vehicle deployment, HOV incentives, and traditional public transit into regional action plans that support sustainable mobility for economic growth and quality of life.


This article was first published in an alternative format by Meeting of the Minds and is reproduced with kind permission from that organisation and in particular Gordon Feller. Find out more at

Share now...