Preparing Cities and Towns for Self-Driving Vehicles
As time advances and new technology, adoption and traffic scenarios unfold, cities face many uncertainties when it comes to transportation and mobility.
As the availability and flexibility of self-driving vehicles such as this Mercedes-Benz F015 begins to increase, flights to nearby regions, intra-regional bus use, and inter-regional goods shipments by train could fall – and this changing transportation landscape will impact cities substantially. (Photo: mercedes-benz.com)
This is a guest post by Dr. Kara Maria Kockelman.
A substantial shift in transportation options is emerging, with various technologies advancing to market for connected and highly automated vehicles (C/AVs). Our nation’s $1,000 per capita annual crash costs may plummet – and our days of driving may be numbered – as technologists and manufacturers work hard to publicly deliver communicating vehicles in the near term and fully-self-driving technologies in the longer term.
Connected vehicles communicate a basic safety message involving their position, speed, and direction to help avoid crashes, improve traffic signal timing plans, and receive valuable feedback from roadside devices that wish to alert vehicles or their drivers to downstream issues (like tight, icy bends in the road). These dedicated short-range communications (DSRC) will probably be required on all new passenger vehicles sold in the United States in the 2020 model year, costing less than $100 per vehicle. Adding such radio-signal technology to existing vehicles will also be rather easy, but will require that drivers react to the resulting audible alerts, rather than rely on embedded automated systems that exist on many newer vehicles (like emergency braking and lane-keeping assistance).
Self-driving vehicles (SDVs) will transform personal and freight transport patterns by providing travel alternatives that compete favorably with existing mainstays. While our days of driving may be numbered, our road-use demands, or vehicle-miles travelled (VMT), will rise. With the flexibility of SDVs for rent or purchase, flights to nearby regions, intra-regional bus use, and inter-regional goods shipments by train will fall. Shared SDVs (or “shared AVs”) will allow most Americans to access such vehicles at lower cost – perhaps as low as one dollar per mile within a decade in some cities in the United States. By reducing driver labor costs, and perhaps access and use costs, car and truck travel will rise.
What all of this means is that uncertainties for cities, states, and transportation system managers are high, with many possible technology, adoption and traffic scenarios unfolding as time advances.
Many wonder if Americans are ready to embrace connected and self-driving vehicles, but our fleet’s evolution has already begun – and our research results suggest that 70 to 87 percent of our light-duty vehicle fleet will be fully self-driving by 2045. This is without any proactive policies, though (like requiring only self-driving mode in a city’s downtown or along certain freeway lanes). Falling technology prices and increases in households’ willingness to pay for such technologies, much like we have seen with smartphones and other devices, should help Americans embrace connected and self-driving vehicles. Of course, various levels of automation already exist on select vehicle makes and models, including electronic stability control or ESC (required since model year 2012), adaptive cruise control (ACC), and automatic emergency braking (AEB).
The National Highway Traffic Safety Administration (NHTSA) defines Level 1 automation as function-specific automation (like ESC or AEB), Level 2 as combined function automation (like adaptive cruise control plus lane-keeping assistance, as seen recently on Tesla’s Model S car), Level 3 as limited self-driving automation (so a licensed driver must be available within seconds to take over operation of the vehicle, if needed), and Level 4 as fully self-driving automation (where entire trips can be completed without a driver).
How much are Americans willing to pay right now for these kinds of technologies? Our survey results suggest an average willingness to pay $67 for DSRC connectivity and $5,857 for Level 4 automation, when all adults are included, or $111 and $14,196 when zero-value respondents are removed from the averages. As expected, readiness to invest in such advances rises as one’s friends and neighbors acquire the technologies. Moreover, almost 60 percent of all individuals indicated wanting to use self-driving features soon, for at least some of their current trips, but affordability and equipment failure are top concerns.
The value of safer journeys, with lower operator burden and possibly avoided parking costs, is substantial. We estimate SDVs’ social benefits to be roughly $3,000 per year per vehicle initially, rising to nearly $5,000. Moreover, shared AV fleets suggest an important opportunity to improve fuel economy (thanks to smaller vehicles), reduce emissions (thanks to keeping engines warm and avoiding cold starts), reduce parking needs (as households reduce their own vehicle holdings), lower travel costs (for those who drive their vehicles less than 5,000 miles per year), connect to major existing transit lines (as a first-mile, last-mile service), and improve mobility and access for those without vehicles or unable to drive (including tens of millions of elderly Americans who regularly avoid night-time driving and other endeavors).
All told, AVs may save the United States’ economy roughly $430 billion annually. However, we must plan for the added travel demands and VMT that our networks will be asked to support. Congestion-based tolls and vehicle-type tolls can help avoid the downsides of easier travel. Congestion pricing is designed to keep traffic moving, and is especially valuable at regional bottlenecks, like bridges. Credit-based congestion pricing (CBCP) comes with travel credits or budgets for each traveler, so that everyone in a region owns a share of the network and will pay out of pocket only after exceeding their monthly budget. Our research shows how CBCP policies can improve the welfare of most travelers, while better reflecting the true costs that each of us is imposing on others (those behind us in the traffic stream) when we enter a congested roadway. In addition to avoiding congestion, we want to avoid bigger, heavier, higher-emitting, and less fuel-efficient vehicles. With the GPS systems on board these smart vehicles, we have the opportunity to price based on the external costs that different vehicles place on our infrastructure, our lungs, and our climate. Experts and citizen groups can help policymakers determine such tolls to keep things equitable and realistic as we transition our communities toward transportation systems that function much better than many do today.
As we move toward 100 percent C/AV use, traffic signal systems can be replaced by roadside monitors that send instructions to individual vehicles and mini-platoons of vehicles to make maximal use of an intersection while still offering pedestrian and bicyclist phases. And non-motorized travelers (as well as motorcycle riders) can add connected technologies to their phones, backpacks and bikes, to help smart vehicles anticipate their presence, and avoid the thousands of pedestrian and cyclists deaths this country experiences every year. While motorcycles will be very difficult to automate, a rider’s helmet face or glasses can be designed map drive paths to keep motorcyclists in synch with mini-platoons.
Our simulations also suggest that every shared AV can replace about 5 to 10 privately held vehicles, which provides many benefits. With a fleet of shared smart vehicles, we also can promoted dynamic ride-sharing, where neighbors, colleagues, and complete strangers opt to share their vehicles, saving on travel costs, congestion, and emissions. Such systems may be critical in most cities to counteract the added VMT that comes with easier motorized travel. Our current park-and-ride lots and major destination hubs (like universities, hospitals, convention centers, shopping centers and central business districts) also make great places to cluster trip-makers’ origins and destinations, and self-driving mini-buses can make our transit systems more demand responsive, frequent, affordable and valuable to all travelers.
Together, connected and automated technologies offer tremendous opportunities for improved access, mobility, and safety across the country. But it is important that we promote the best features of these new technologies, helping to avoid more serious congestion and ensure more sustainable systems.
Please click here to learn more about these technologies and associated research results.
About the Author: Dr. Kara Maria Kockelman is the E.P. Schoch Professor of Civil, Architectural and Environmental Engineering at the University of Texas at Austin. She is a registered professional engineer and holds a PhD, MS, and BS in civil engineering, a Masters of City Planning, and a minor in economics from the University of California at Berkeley. Dr. Kockelman is a recognized expert on the subjects of automated vehicles, shared self-driving electric vehicles, travel demand forecasting, urban planning, land use modeling, traveler behavior and crash forecasting, vehicle ownership and use decisions, traffic patterns under congestion pricing and managed lanes, transport emissions and economics, and benefit-cost analysis of transport investments and policies. She is well known for her work holistically characterizing the benefits and costs of different transportation investments, policies and practices.