Testable Prediction: Fleet Electrification and Autonomy, each with > 50% share of new vehicles by about 2030.

Another article about my presentation at the Commercial Vehicle Outlook conference: James Jaillet writes in CCJ News “Evolving use of U.S. highways, increasing truck freight efficiency highlight CV Outlook panel on freight infrastructure

… Smarter trucks, smarter roads and efficiency and financial gains from both were central themes of the panel’s discussion.

Levinson said for the first time since the automobile began its 100-year boom in the early 20th century, miles traveled per capita has stagnated, with demographic changes (retiring baby boomers, for instance) and online commerce being two of the key drivers. Traffic congestion, however, may not be impacted, because total miles traveled in the U.S. is still on the rise due to population increases and other long-term trends.

As the world’s relationship with the automobile changes, Levinson said, so does the potential for alternative fuels and electric-powered vehicles.

“Full battery electric vehicles and plug-in vehicles are coming into the market now,” he said. “If you turn this type of data through an s-curve (a model used to predict uptake of new technologies), you come up with around 2030 for when half of car sales will be electric.”

That uptake will be later for trucks, he said, though he still sees electric trucks on the horizon.

Similarly, Levinson said, around 2030 half of cars sold [ED: I think I said miles traveled] will be some form of autonomous. “Technology will drive regulators rather than the regulators driving technology,” he said.

And as automated lane-keeping systems become more prevalent in heavy trucks and passenger cars, narrower lanes and narrower vehicles could follow, Levinson said. “We can rethink how we design cars,” he said. “With lane keeping systems, we can have narrower lanes and shared lanes. Two cars driving in tandem in a lane,” he said, could reduce congestion and improve overall efficiency of U.S. transportation. …

Is Bicycling Contagious? Effects of Bike Share Stations and Activity on System Membership and General Population Cycling

Recent working paper

Nice Ride Membership Map
Nice Ride Membership Map

This paper presents new evidence about the role of bike share systems in travel behavior using a diffusion of innovation framework. We hypothesize that bike share systems have a contagion or spillover effect on (𝐻1) propensity to start using the system and (𝐻2) propensity to bicycle among the general population. We test the first hypothesis by modeling membership growth as a function of both system expansion and the existing membership base. We test the second hypothesis by using bike share activity levels near one’s home in a model of household-level bicycle participation and trip frequency.  Our study shows mixed results. Bike share membership growth appears to be driven, in a small part, by a contagion effect of existing bike share members nearby. However, we did not identify a significant relationship between proximity to bike share and cycling participation or frequency among the general population. The findings hold  implications for marketing, infrastructure investments, and future research about bike share innovation diffusion and spillover effects.

KEYWORDS: Bike Share; Diffusion of Innovation; Travel Behavior

Maximum Homerdrive, On Deployment of Driverless Cars and Trucks

Dorothy Cox at The Trucker wrote about my panel at the really interesting  CVO Conference yesterday in Dallas “Industry must change, but tech advancements carry their own challenges, panelist says“. The conference can be followed on Twitter at the  #CVOutlook hashtag .

… And of course the conversation got around to autonomous trucks, the bywords of the day.

By the year 2030, maybe half of all miles driven will be by driverless trucks, predicted David Levinson, chair of transportation at the University of Minnesota. [Ed: NOT ACTUALLY WHAT I SAID, I PUT THE TIMEFRAME FOR CARS IN MID 2030s, BUT OK, I WAS TALKING FAST]

But he painted a somewhat different picture than other speakers in discussing the subject. He said a driver might be engaged in supervising multiple drone trucks on set routes from a control center, intervening if something goes wrong, and told attendees that in the future, goods pickup and delivery could be bid on, leading to supply chain networking and consolidated home delivery.

Where does that leave trucks?

Evaluating the “Safety In Numbers” Effect With Estimated Pedestrian Activity

Recent working paper:

Pedestrian risk vs. PM pedestrian flow
Pedestrian risk vs. PM pedestrian flow

 

Pedestrian and bicyclist collision risk assessment offers a powerful and informative tool in urban planning applications, and can greatly serve to inform proper placement of improvements and treatment projects. However, sufficiently detailed data regarding pedestrian and bicycle activity are not readily available for many urban areas, and thus the activity levels and collision risk levels must be estimated. This study builds upon other current work by Murphy et al. (1) regarding pedestrian and bicycle activity estimation based on centrality and accessibility metrics, and extends the analysis techniques to estimation of pedestrian collision risk. The Safety In Numbers phenomenon, which refers to the observable effect that pedestrians become safer when there are more pedestrians present in a given area, i.e. that the individual per-pedestrian risk of a collision decreases with additional pedestrians, is a readily observed phenomenon that has been studied previously. The effect is investigated and observed in acquired traffic data, as well as estimated data, in Minneapolis, Minnesota.

Accessibility and Centrality Based Estimation of Urban Pedestrian Activity

Recent working paper

Estimated evening peak period pedestrian activity
Estimated evening peak period pedestrian activity

Non-motorized transportation, particularly including walking and bicycling, are increasingly becoming important modes in modern cities, for reasons including individual and societal wellness, avoiding negative environmental impacts of other modes, and resource availability. Institutions governing development and management of urban areas are increasingly keen to include walking and bicycling in urban planning and engineering; however, proper placement of improvements and treatments depends on the availability of good usage data. This study attempts to predict pedestrian activity at 1123 intersections in the Midwestern, US city of Minneapolis, Minnesota, using scalable and transferable predictive variables such as economic accessibility by sector, betweenness network centrality, and automobile traffic levels. Accessibility to jobs by walking and transit, automobile traffic, and accessibility to certain economic job categories (Education, Finance) were found to be significant predictors of increased pedestrian traffic, while accessibility to other economic job categories (Management, Utilities) were found to be significant predictors of decreased pedestrian traffic. Betweenness centrality was not found to be a significant predictor of pedestrian traffic, however the specific calculation methodology can be further tailored to reflect real-world pedestrian use-cases in urban areas. Accessibility-based analysis may provide city planners and engineers with an additional tool to predict pedestrian and bicycle traffic where counts may be difficult to obtain, or otherwise unavailable.

Cohort Effects and Their Influence on Car Ownership

Recent working paper

Proportion of population with a driver's license by age and cohort (1990, 2000, 2010).
Proportion of population with a driver’s license by age and cohort (1990, 2000, 2010).

Recent trends in the United States suggest a movement toward saturation of vehicle ownership. This paper examines this trend through an analysis of car ownership in the Minneapolis- St. Paul, Minnesota (USA) metropolitan region. Data from pooled cross-sectional household surveys are used to calibrate a model of car ownership that includes birth cohort effects to capture unobserved variations in preference toward car ownership across generations. Declines in household size and worker status have significant impacts in limiting the growth of car ownership, but they are also coupled by an apparent softening of preferences toward ownership among young adults.

Keywords: car ownership; cohort; generational effect; aging; income; saturation; United States

Accessibility and Transit Performance

Recent working paper

Residual plot of the parsimonious regression model
Residual plot of the parsimonious regression model

This study disentangles the impact of financial and physical dimensions of transit service operators on net transit accessibility for 46 of the 50 largest metropolitan areas in the United States. To investigate this interaction along with the production efficiency of transit agencies, two types of analysis are used: a set of linear and quadratic regressions and a data envelopment analysis. We find that vehicle revenue kilometers and operational expenses play a pivotal role in enhancing the accessibility to jobs by transit. The bivariate linear regression models indicate a 1% increase in operational expenses and vehicle revenue kilometers increase the number of jobs that can be reached within 30 minutes by 0.96 and 0.95%, respectively. The results of the quadratic functional form, also, show transit services may have both increasing and decreasing accessibility returns to scale depending on system size, and the results are sensitive to the model used. Overall, the highest system efficiency (access produced per input) is found in the New York, Washington, and Milwaukee metropolitan areas, while Riverside, Detroit, and Austin perform with the lowest efficiency.

Keywords: Public transit; Accessibility; Envelope of output; Returns to scale; Metropolitan area

Physical Activity in School Travel: A Cross-Nested Logit Approach

Recent working paper

The tree decision for a two-level cross-nested logit model
The tree decision for a two-level cross-nested logit model

This paper considers school access by both active (walk, bike), quasi-active (walk to transit) and non-active modes (car) in a two-level cross-nested logit framework. A sample of 3,272 middle and high school students was collected in Tehran. The results of the cross-nested logit model suggest that for people who choose walking, increasing a 1 percent in home-to-school distance reduces the probability of walking by 3.51 percent. While, this reduction is equal to 2.82 and 2.27 percent as per the multinomial and nested logit models, respectively. This is a direct consequence of the model specification that results in underestimating the effect of distance by 1.24 percent. It is also worth mentioning that, a one percent increase in home-to-school distance diminishes the probability of taking public transit by 1.04 among public transit users, while increases the probability of shifting to public transit from walking by 1.39 percent. Further, a one percent increase of the distance to public transport, decreases the probability of students’ physical activity, approximately, 0.04 percent.

Keywords: Public Transit; Active Mode of Travel; School Trips; Tehran

Intra-household Bargaining for School Trip Accompaniment of Children: A Group Decision Approach with Altruism

Recent working paper

The share of travel mode in each escorting group
The share of travel mode in each escorting group

This paper tests a group decision-making model with altruism to examine the school travel behavior of schoolchildren aged between 6 and 18 years in the Minneapolis-St. Paul metropolitan area. The school trip information of 1,737 two-parent families with a schoolchild is extracted from Travel Behavior Inventory data collected by the Metropolitan Council between the Fall 2010 and Spring 2012. The proposed model has four distinctive characteristics compared with traditional developed models in the field of school travel behavior including: (1) considering the schoolchild explicitly in the model, (2) allowing for bargaining or negotiation within households, (3) quantifying the intra-household interaction among family members, and (4) determining the decision weight function for household members. This framework also covers a household with three members, namely, a father, a mother, and a schoolchild, while unlike other studies is not limited to dual-worker families. To test the hypotheses, we developed two models with and without the group-decision approach. Further, the models are separately developed for different age groups, namely schoolchildren aged between 6-12 and 12-18 years. This study considered at a wide range of variables such as work status of parents, age and gender of students, mode of travel, and distance to school. The findings of this study demonstrate that the elasticities of two modeling approaches are different not only in the value, but in the sign in some cases. In more than 90 percent of the cases, further, the unitary household model overestimates the results. More precisely, the elasticities of unitary household model are as large as 2 times more than that of the group-decision model in 25 percent of cases. This is a direct consequence of model misspecification that misleads both long-term and short-term policies where the intra-household bargaining and interaction is overlooked in travel behavior models.