Adoption of autonomous vehicles: Addressing the details

By Mary Ebeling

Although the mass media has been quick to tout the practicalities of switching to autonomous vehicles, many complexities are also apparent. As SSTI and others have noted, a transition to autonomous vehicles may be closer than we think. However, at the recent TRB meeting the number of papers highlighting the opportunities and challenges associated with these vehicles made a clear statement about the enormity of the change DOTs, MPOs, and policy makers anticipate.

Much has been made of the expected benefits resulting from autonomous vehicles, including more efficient use of existing roadway capacity, increased safety through the reduction or near elimination of crashes, more efficient freight delivery, and air quality improvements. However, in addition to recognizing the benefits of autonomous vehicles, new work offers a reality check on this optimism and points out some of the potential challenges. These include identifying necessary revisions to state and municipal policies and codes on vehicle operation; identifying the extent of necessary infrastructure upgrades for streets and highways; identifying needed changes to insurance coverage; accounting for older cars without compatible technology, as well as non-motorized transportation modes; incorporating higher level of detail into local and national street maps to include all roads, signals, bike lanes, and pedestrian crosswalks; and other issues yet to be identified.

In a recent paper by Michael Sivak and Brandon Schoettle of the University of Michigan, the authors lay out other considerations that raise questions about whether autonomous vehicles will smoothly transition into mainstream usage. Several of these include:

  • The decisions autonomous vehicles make will be only as good as the vehicle’s programming. Whether that programming can match the predictive skills of a driver with decades of experience is unclear.
  • There will be a long transition period when autonomous cars will share the road with traditional, human-operated vehicles. There is reasonable concern that these two very different types of operators may make choices that come into conflict.
  • Road users that are not using autonomous technology may not act as predicted by programming. In urban areas, pedestrians and bicyclists must be taken into account. And in rural areas less common vehicles, such as large, slow-moving farm equipment or Amish buggies may be encountered.
  • Construction zone or crash incident areas may feature traffic being re-routed in real time and therefore conditions not updated in GPS or road databases.
  • Some vehicle mechanical failures may no longer lead to crashes, however tire or brake problems will likely still be difficult events to fully anticipate.

As the discussion on how we handle autonomous vehicles moves into a detailed consideration of what implementation means, it becomes increasingly clear that the path toward autonomous vehicles is not a simple one, and many legal, cultural, and infrastructure hurdles still exist.

Mary Ebeling is a Transportation Policy Analyst at SSTI.