The Shapes of Streets to Come – How New Transport Technology will Reshape Urban Space

Kevin Krizek and I wrote “The Shapes of Streets to Come – How New Transport Technology will Reshape Urban Space” which appeared in European Financial Review (registration required) (reprinted below).

In the same issue, John Renne has a piece called “Changing Preferences for Transportation and Transit-Oriented Communities Signal a Gradual Move to a Post-Oil Based Society“(which I don’t entirely agree with)




The Shapes of Streets to Come – How New Transport Technology will Reshape Urban Space

By David M. Levinson and Kevin J. Krizek

Autonomous vehicles are coming. At their best, AVs are stimulating an impulse to drive genuine innovation. At its worst, they are a hubris that causes us to overthink the solutions to transport problems in cities.


Big changes are coming for how people will get around in cities across the globe. The most important change will hinge on the introduction of autonomous vehicles (AVs). Simultaneously, cities will witness the conversion of the vehicle fleet to being primarily electric-powered (from a grid rapidly converting to renewable energy and off-the-grid solar charging) and new ownership models like shared mobility become more common.

The overall pace of deployment of AVs and their effects will vary by the size of the city, the cost of labour, and the desire for politicians and their constituents to innovate. How all of these factors play out on the multiple stages and multiple scales (e.g. the neighbourhood, metropolitan, and national levels) will prove exciting to watch. The best part is that you not only get to observe how things will play out; you get to participate as well. We preview many of the prevailing tensions of this emerging landscape below.


Autonomous Vehicles

After decades of technological slumber in the automobile industry, self-driving cars are here. Rolling on the roads today in semi-urban environments are cars that can recognise speed limits and adjust their speed instantly. They can maintain a safe following distance from other cars, and brake when needed. They can even recognise the difference between cars, buses, and cyclists. The technology is at the cusp of being widely deployed, something that will take place over the next two decades. Significant other hurdles, however, lie ahead. These impediments include how cultures might adapt (e.g. how quickly will people surrender control, and their comfort with technology), legal regimes might change (e.g. standards, reconciling responsibility in crashes, the role of licensing), and street designs will be altered (e.g. the extent to which AVs will be apportioned separated lanes on different types of roads, and how soon human-controlled vehicles will be prohibited in places). We focus on the last hurdle here.

One of the strongest but often unrealised arguments for the advent of AVs relates to street capacity. Where today a freeway travel lane is typically 3.6 meters wide, with AVs, a standard lane might only have to be 2 meters (just wider than the width of a full- size car or SUV). This alone nearly doubles capacity. Farther into the future, lanes might be dynamically resized rather than permanently fixed in paint. Being automated, these cars can trail one another more closely as well. Instead of following at two seconds or one-and-a-half seconds, they might follow at one second or less, increasing throughput.

Today the average vehicle carries only one or two people, yet the average vehicle has seats for four or six or eight passengers. The fleet is oversized, especially in the US. Americans have a propensity to buy large cars for the few times when they may need it. The extra seats, however, sit unused most of the time. Automation, combined with mobility-as-a-service presents opportunities for new vehicle forms.


New Types and Forms of Cars

Smaller one and two-person vehicles can be the new standard, and larger cars the exception, only summoned on demand when needed. In lower density areas, travel distances remain large, but the use of AVs will allow suburbs and small towns to wring out more road capacity, provide good arguments against road expansion, and claw-back space that has been given over to parking.

A city can start to realise large benefits because it can get more capacity out of that smaller vehicle, more energy efficiency, a greater range for the battery, and so on. One of the more noticeable elements will be a transformation of the shape of the car. Vehicles will begin to sport new designs whose markets will be defined by required use. Consider an enclosed motorcycle that’s electric, quiet, safe, stabilised, and automated. It is safe because, not only is the vehicle designed well, with a roll bar and all that, and because it is driven by a computer with nearly instant reaction times, but because all the other vehicles are also automated. Small cars require less space and it is easy to see how future cars will starkly contrast with their ancestors. Meanwhile, sensors and computers replace the human-facing control functions; electrification is changing the entire motor system, so future cars will be simpler to manufacture and maintain than the internal combustion engine.


Innovations in Related Modes

The complexity of how and when – not if – to accommodate AVs will be further complicated because other modes of transport are re-inventing themselves as well. Different types of mobility-as-a-service are coming on line. These include new forms of taxis and transit services that are both smaller and bigger than a standard bus. Taxis will be more extensively used because the vehicles are smaller and driverless, and so cheaper. In urban areas, there will be more frequent transit services in selected corridors, which will be less expensive to provide as labour is automated away. Elsewhere, today’s infrequent bus and commuter rail services will be replaced by mobility-as-a-service type of options; instead of having a bus that comes once an hour, people will be using taxis – often single passenger taxis, maybe shared ride taxis. While the exact market configuration (who owns, who rents, who shares a ride, who rides alone) will be sorted out over time, it is clearer to see how, like today in Manhattan, people who live in dense cities won’t be owning cars, but instead will subscribe to a service, buy the services on demand, or find it provided by the public as a “free” utility, like the elevator in an office building.


Role of Walking and Bicycling

Amidst this uncertainty, bicycling and walking will thrive for shorter and medium distance trips. These might be trips where people yearn for physical activity or just want to be outside. Their use will continue to be constrained by weather and hills, though e-bikes, with electrical pedal assistance, will mitigate some of that. This is one of the reasons we will likely see an increase in the attention devoted to physically powered movement for next few years. It is green and energy efficient. It makes many people feel better. Most importantly, bicycling and walking are modes that are relatively known and proven in selected markets.

All of this is to say that traditional modes, bicycling and walking will continue to exist and begin to thrive even in the US. This owes to increased population densities in central cities (and trip distances therefore decreasing), increased safety because AVs are less likely to kill them than human driven vehicles are, and a growing inclination to more fully connect with others and their environment.


Infrastructure Needs and Who Gets What Space

How will street space be appropriated in a manner that will allow multiple modes to harmoniously co-exist? Answers to this questions will play out differently between and within cities. Fundamental geometric limits ultimately dictate the usefulness of these improvements. Where the intensity of development is higher, several modes will compete for the same space. Different modes can safely mix in the same shared space at slow speeds, as is now found in historic sections of many European cities. Further away from these cores in lower density areas, where space allows, the infrastructure provision for modes will be more segregated.

Typically, local municipalities operate the local streets and state or regional agencies maintain the connecting the backbone. On the backbone, we envision special (managed) lanes for automated cars for a period of time, just as today we have express, HOV, or high occupancy/toll lanes. It may even be the same lanes will benefit all users, as separated lanes will allow reaping the benefits of closer following distances than possible with mixed (human plus automated) traffic. As all vehicles become automated, all lanes will be managed.


Shapes of Streets and Shapes of Cities

The ability for cities to dynamically reconfigure lanes and repurpose streets will be the central challenge. The speed and manner in which cities and regions respond to the onset of these big changes will vary. Some places will move quicker than others. Politics and openness to innovation will be important. But ideas are light baggage, and successful policies will be quickly copied and emulated.

Shapes and patterns of development of communities determine how most people get around. The size and nature of buildings and roads are important. The prevailing perspective is that cities have evolved under different technological and political regimes; they therefore embody the DNA of their continents. Granting exceptions, North American cities have a DNA that is distinct from their Asian, European, or Latin American counterparts. Cities in Europe are smaller in size, they were mostly formulated in an era prior to automobile, and their networks for movement are more multi-modal. It’s sometimes easier to get things done without a car, automated or not. Hundreds of European cities already have an extensive track record severely limiting automobile traffic in historic districts. The Italians call it the ZTL (Zona a Traffico Limitato). We expect European cities will be keener to take back even more street space from the new AV, and use it for landscaping, linear parks, cycle tracks, and high-frequency transit services. Places in other parts of the world will follow suit.


AVs and the Future of Cities

At their best, AVs are stimulating an impulse to drive genuine innovation that will make lives richer and more connected, faster and safer, and more productive. At its worst, they are a hubris that causes us to overthink and over-engineer the solutions to transport problems in cities. Either way, AVs are coming. It is just a matter of when and how.

Political and legal systems in cities will be forced to play catch up with technological systems. The cities that do will win the 21st century. The others are building unnecessary capacity justified by extrapolating the exhausted trends of the past. The aim of transport should be ensuring people and goods can reach their destination safely and efficiently. This requires focusing on what will improve access tomorrow, not what might have increased access yesterday.


Featured image courtesy of the author Kevin J. Krizek

About the Authors

David Levinson and Kevin J. Krizek are the authors of The End of Traffic and the Future of Transport, available on Amazon Kindle and Apple iBooks.

levinson-webDavid M. Levinson is Professor in the Department of Civil, Environmental, and Geo-Engineering at the University of Minnesota and will soon be joining the University of Sydney. He holds the RP Braun/CTS Chair in Transportation.

Krisek-webKevin J. Krizek is Professor and Director of the Environmental Design Program at the University of Colorado-Boulder. He also serves as the visiting professor of “Cycling in Changing Urban Regions” at Radboud University in the Netherlands. Krizek was a 2013 fellow of the Leopold Leadership Program and was awarded a 2014 US-Italy Fulbright Scholarship.Technology will Reshape Urban Space

Does poor road condition increase crashes?

In a region well known for its severe weather, maintaining pavements to meet high standards remains a challenge. Changes in weather states (such as the freeze-thaw cycle) leads to distresses in the pavement materials. There exist claims that poor pavement quality reduces the ability of roads to drain and reduces the ability of vehicles to resist skidding, and is thus associated with more crashes. In order to improve road safety, several pavement maintenance treatments are carried out, such as “rout and seal cracks” and “hot-mix patching” for improving pavement roughness and distress (Tighe et al., 2000). Others have found that crash rate depends on the pavement type and pavement condition. Crash rate of tined pavement sites is larger than the rate of ground pavement sites. When the pavement condition is wet or icy, crashes are more likely than under dry conditions (Drakopoulos et al., 1998). , When the pavement condition is poor, severe crashes are more likely, but when the pavement condition is very poor, severe crashes are less likely to occur than poor pavement conditions (Li et al., 2013). In accident rate estimation models, the results indicate that most important independent variable is “AADT”, and “geometric design” (lane width and access control) and “pavement condition” (friction, serviceability index, and pavement type) are also important variables (Karlaftis and Golias, 2002). Our research proposes to statistically test such claims of a relationship between incident number and road quality, while controlling for traffic data (AADT and percent truck), segment length, crash conditions (date, road characteristics, and road surface), and pavement type. To investigate the relationship, we combine data from various sources. We then conduct a statistical analysis for ascertaining the effects of good road quality on incident number and severity. This paper describes the data, methods, hypotheses, and results in turn.

The Synergistic Effects of Transit Oriented Development and Transit Hubs on Accessibility in the San Francisco Bay Area

Recent working paper: AccessInBayArea

This paper is a case study of the accessibility impacts of transit projects and nearby development on transit accessibility in a region, which already has significant levels of accessibility via transit. The project under consideration is the San Francisco Transbay Transit Center and the associated planned development. Findings indicate that both portions of the project increase accessibility via transit in the region. However, the contribution from the planned development is far greater. Furthermore, the increase in accessibility from the project as a whole is greater than the sum of the contributions of the individual portions of the project. This indicates that in areas where there is already transit service, the development of land near the transit service can have a greater impact on accessibility levels than the improvement of connections between transit services.

The Impact of Analysis Boundaries in Accessibility Evaluations: A Case Study

Recent working paper:

A case study transit project, the Harris County Transit Re-Imagined Bus Network, is utilized to evaluate the impact of analysis boundaries on accessibility analysis. Results from the case study indicate that choice of analysis boundaries can have a significant impact on the value of absolute accessibility measures. In general the trend shows that the tighter the analysis boundary is to the network, the higher the value of the absolute accessibility measure. However, relative accessibility measures such as percent difference between scenarios are consistent regardless of analysis boundary size. This indicates that the choice of analysis boundary is of only moderate importance for scenario comparisons within the same analysis boundary. However, when comparing between different regions or in areas within different boundaries, the choice of those analysis boundaries could have a significant impact on all results. Furthermore, care should be taken to indicate the analysis boundary used whenever an absolute accessibility measure is presented.

The City is flatter: Changing patterns of job and labor access

Recently published:

Standard deviational ellipse: Accessibility to 750,000 jobs or more within 20 minutes in the AM peak: 1995-2005
Standard deviational ellipse: Accessibility to 750,000 jobs or more within 20 minutes in the AM peak: 1995-2005


This study measures accessibility by automobile for the Minneapolis – Saint Paul (Twin Cities) region from 1995 to 2005. In contrast to most previous analyses of accessibility, this study uses travel time estimates derived, to the extent possible, from actual observations of network performance by time of day. A set of cumulative opportunity measures are computed with transport analysis zones (TAZs) as the unit of analysis for 1995 and 2005. Analysis of the changes in accessibility by location over the period of study reveals that, for the majority of locations in the region, accessibility increased over this period, though the increases were not uniform. A “flattening” or convergence of levels of accessibility across locations was observed over time, with faster-growing suburban locations gaining the most in terms of employment accessibility. An effort to decompose the causes of changes in accessibility into components related to transport network structure and land use (opportunity location) reveals that both causes make a contribution to increasing accessibility, though the effects of changes to the transportation network tend to be more location-specific. Overall, the results of the study demonstrate the feasibility and relevance of using accessibility as a key performance measure to describe the regional transport system.


The Resurfacing and Restriping of Franklin Avenue SE – A Review |

We don’t do enough street reviews on I aim to help rectify that gap.
The neighborhood of Prospect Park in Minneapolis has recently experienced the trauma of a road resurfacing. Franklin Avenue SE, which forms the East-West spine through the neighborhood (map), usually connecting old Minneapolis west of the River with St. Paul east has now reopened. (The Franklin Avenue Mississippi River Bridge remains under construction. This will warrant a separate review by someone).  Now recovering, we can evaluate what has changed. 

What was one of the worst roads in Minneapolis in terms of ride quality is now significantly better. My perception of Minneapolis roads, warped by the road I use most often, has gone from a grade of F to B. The road is smooth, far smoother than before, with the patchwork of patches now a continuous solid surface. It still feels a bit lumpy to drive on, but clearly it is an improvement. Maybe it’s my suspension, but I think laying asphalt in hot weather is less than ideal conditions.

Most of the pedestrian crossings along the resurfaced section were rebuilt to comply with modern ADA standards, EXCEPT,  the worst one at Franklin SE and Arthur Ave SE. Was it forgotten, or was it so bad (because the grade was so steep) that there was no point in fixing it anyway. Or were Minnesotans just cheap?

This curb cut was not improved.

The road was mostly restriped for bike lanes. Some on-street parking was removed. It is now continuous with the section of Franklin Avenue in St. Paul (which runs from the city line to Eustis street, before it just gives up at the Mn 280 interchange). Props to St. Paul for getting a bike lane done before Minneapolis, this is rare.

There is a nice set of green striping and bike box at the intersection of Franklin Ave/27th Ave/West River Road. Which is the best you can do if you insist upon signaling rather than a roundabout.

The green painted road

I did say “mostly restriped.” The section in front of St. Francis Cabrini Church was striped going uphill (eastbound), but not striped southbound (westbound). Stripes resume just west of Thornton Street. I can speculate about city-church politics, and the need for parking. Let me just suggest the bike lane is a pointed gun*, shooting bicyclists quickly down the road, to mix with traffic for a one block section before they return to the dignity of their own lane. Perhaps the idea is that this bike lane gap will shame someone into backing down or putting in some money and fixing it properly. If I were a betting man however, I would bet in five years the gap is still there.

Bike lane in front of Francis Cabrini is striped eastbound, not westbound. A confident bicyclist takes the road anyway.

The lane ends at the railroad tracks (just east of the church). If these abandoned tracks were to become a maintained trail towards the University of Minnesota, this might be an acceptable place to end it. As it stands now, the tracks are merely attracting weeds and broken glass.

The Bridge over I-94 on Franklin Avenue is striped, though there is a gap between the bike lane and sidewalk on the northern (westbound) side of the bridge. This keeps the bike lane straight. It denies the lane an opportunity for a protected section.

To be clear, the bike lane is striped, in a few places there is a small buffer where there is extra pavement. There are no barrier protections, not even plastic bollards. I am sure I don’t feel comfortable with my kids riding on this street, even with the lane, because of the lack of protection. It is probably fine for spandex wearing bicyclists however. Concrete curbs would help, but of course reduce flexibility for drivers who might need to use the bike lane to swerve for some reason (even if slowly), as well as road maintenance workers or wide trucks. One of the proposals early on would have had both bike lanes on one side of the road, so they could be protected, but wide enough to run a pickup truck on for snow clearance. This didn’t happen.

Another bicyclist with kid still chooses to use the sidewalk rather than the new bike lane.

The one property which had no driveway access before construction was given a driveway, and a brick car landing was constructed on their property. Apparently city staff opposed this, because code? I am glad to see this actually happen. I am happy my neighbors cared enough, and had enough free time, to turn out. Yet, something so obvious should not require 15 people to show up at a meeting.


This property now has a driveway and a place to park off-street


* The pointed gun metaphor is something I learned as a planner, when the DOT would widen a road upstream of a bottleneck, inducing more demand, and pointing even more cars at the bottleneck, it was like loading a gun. The road is a gun, the cars are the bullets in this metaphor. The bullets (drivers) have some agency, though they are fairly well predictable in the aggregate. This is not the kind of thing staff would say in public of course. The intent of loading the gun was to force the issue of widening the bottleneck, which was a bottleneck because of some physical or social constraint, like a neighborhood using the street as their shopping district.



Cross-posted on

Engineering for Society: Learn – Discover – Transform

The Department of Civil, Enviornmental, and Geo- Engineering at the University of Minnesota has posted its new Strategic Plan: Engineering for Society: Learn – Discover – Transform.

Though it is the product of the department as a whole, I did participate in its drafting and framing.  Happy reading.

Intraurban Accessibility and Employment Density

Recent working paper:

This study investigates the relationship between urban accessibility and firm agglomeration, as reflected in patterns of urban employment densities. We use measures of accessibility derived from the regional highway network, combined with small-scale (Census block-level) data on employment from the Longitudinal Employer-Household Dynamics (LEHD) data set to generate proxies for different sources of agglomeration, specifically urbanization and localization economies. These variables are employed in a set of employment density regressions for 20 two-digit NAICS code sectors to identify the propensity of each sector to agglomerate in response to varying levels of accessibility. The density regressions are applied to sample data from the Minneapolis-St. Paul, Minnesota (Twin Cities) metropolitan region for the years 2000 and 2010. We find that in general urbanization effects tend to overshadow those of localization effects. Moreover, these effects tend to vary by sector, with many service-based sectors showing a stronger propensity to agglomerate than manufacturing and several “basic” sectors like agriculture, mining and utilities.

The Safest Path: Analyzing the Effects of Crash Costs on Route Choice and Accessibility

Recent working paper:TripCountSafest

The “safest path” is proposed to optimize the on-road safety of individuals and minimize the cost of crashes. In this study, the framework of a link-based crash cost analysis is built and applied to assess the crash cost of each link segment on the road network of the Minneapolis – St. Paul area based on Safety Performance Functions from the perspective of travelers. The safest path is then found for all OD pairs to compare flow patterns and accessibility distributions with those based on the traditional shortest travel time path. While, the safest path does not coincide with the shortest path, the accessibility distributions have similar patterns.