Advancing Automated and Connected Vehicles: Policy and Planning Strategies for State and Local Transportation Agencies (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

7 Introduction Vehicles that are increasingly automated and connected have the potential to transform the country’s transportation paradigm by providing significant safety and efficiency ben- efits to many modes of travel, and they could have significant environmental and land use impacts. Just as passive safety sys- tems (e.g., airbags and seat belts) and entertainment features (e.g., radios and in-vehicle displays) became ubiquitous and revolutionized travel in the mid- to late 20th century, auto- mated vehicle (AV) and connected vehicle (CV) technologies are set to revolutionize travel in the early to mid-21st century. While these two technologies are undergoing rapid evolution and development, trends are arising such that the public can begin to understand how and when they will find their way into vehicles and onto public roads. The benefits of AVs are potentially vast and include improved safety, mobility, environmental and land use considerations, productivity, and convenience. CV technology similarly has the potential to dramatically improve the safety and effi- ciency of travel for many drivers. AVs and CVs also may have drawbacks and pose risks. These technologies will solve some problems but could also create new ones. For example, cyber- security vulnerabilities associated with CVs could compro- mise safety. Congestion could increase with the proliferation of AVs as driving becomes less onerous and individuals who previously did not drive have more opportunities for travel. This research report presents and develops a rationale for policy and planning strategies at the state, regional, or local levels that, if implemented, could help nudge private-sector choices toward outcomes that would benefit society. Why would these governments want to do this? State, regional, and local governments seek to ensure the safe and efficient opera- tion of public roadways and to foster equity across users of the system; overseeing AV and CV technologies is a natural exten- sion of this longstanding mission. The policy and planning strategies were developed using the best information currently available. To place the research and its recommendations in context, a review of the current state of AV and CV technolo- gies and regulatory activity is provided in this chapter. State of the Technologies AVs For purposes of this work, an AV is one that takes full con- trol of all aspects of the dynamic driving task for at least some of the time. This study focused on the role of higher level AVs in mitigating or exacerbating the societal effects of driving or in creating new effects. The higher levels of vehicle auto mation are designated Levels 3, 4, and 5, according to the SAE Inter- national (2014) taxonomy (see Table 1). Recently released fed- eral policy guidance adopts the SAE International definitions to encourage consistency in describing automated functions. This study, while focused on the higher levels of automa- tion, recognized that many of the required components of the driving task that are performed by humans, however imper- fectly, are hard to replicate using technology. Sensors may suf- fer from insufficient range or resolution, and from occlusions and blind spots, and may perform differently depending on the time of day (or night). Computers need to mimic human intuition, learn from mistakes, and understand situational context. This will include responding to atypical scenarios and conditions, including degraded (or absent) lane mark- ings, inclement weather, temporary construction zones or work zones, emergency vehicles, and a host of other challeng- ing situations. Despite these challenges, AV technology is dif- fusing through the vehicle fleet. Level 1 automation technology is available in the form of adaptive cruise control systems. Tesla integrated and deployed its autopilot technology in its Model S via a software over- the-air update. This system combines automated steering/ lane-keeping with an adaptive cruise control capability, as well as automated lane change (triggered by manually activating a turn signal). Tesla has billed the system as Level 2; however, there is some dispute as to whether it actually meets the defi- nition of Level 3 because the system monitors some aspects of the driving environment. Google’s initial prototype AVs qualified as Level 3 since the driver still needed to monitor the driving environment and take control in certain conditions. C h a p t e r 1

8Several original equipment manufacturers (OEMs) have indicated that they failed to find the value in Level 3 auto- mation compared to the cost to develop (Visnic 2016), and human factors studies have shown that the requirement to provide the human driver sufficient time to retake control is hard to achieve (or expect)(Merat et al. 2014). As such, many OEMs are now focusing their efforts on development of Level 4 automation technologies, with a few exceptions (Audi is reported to offer Level 3 capabilities on its A8 in 2018) (Blackburn 2016). Many of the OEMs and technol- ogy companies are actively testing their prototype vehicles on public roads, including Google in Mountain View, Cali- fornia, and Austin, Texas; Uber in Pittsburgh; and Nissan in Sunnyvale, California. There have been a number of safety incidents involving these prototype systems, although many have been caused by other human-driven vehicles. In Febru- ary 2016, Google’s AV was involved in a minor crash with a transit bus, the first in which the firm’s AV was at fault (Davies 2016). Tesla has also garnered much attention after the first fatality occurring while its autopilot feature was engaged, although reports indicate that the driver had set the speed higher than the posted speed limit and that the driver may not have been monitoring the driving environment at the time of the crash (Shepardson 2016). Many OEMs have made bold claims about when Level 4 technology will be available in new models, beginning with Volvo’s claims of readiness in 2017 (albeit in a limited deploy- ment) and Tesla’s claims of readiness in 2018 (Volvo 2016; Korosec 2015). Others have followed suit, estimating readi- ness in the 2020–2021 time frame, including Nissan, BMW, Ford, and Toyota (which claims it will have Level 4 AVs avail- able in time for the 2020 Tokyo Olympic Games). The time frame for bringing Level 5 automation technology to market is hard to project; however, several industry analysts estimate this technology will be available on public roads in the late 2020s (Cellan-Jones 2015; Ulanoff 2016). CVs CV technology generally refers to a combination of equip- ment (e.g., DSRC onboard units and roadside equipment) and V2V and V2I applications (e.g., forward collision warning, intersection collision avoidance, emergency vehicle alert, sig- nal priority, etc.). Several manufacturers are actively develop- ing and testing DSRC devices and CV applications, including Kapsch, Savari, Cohda Wireless, DENSO, and Arada Systems. Other companies are developing vehicle-to-everything (V2X) equipment that uses other forms of wireless communications, including cellular, Wi-Fi, and Bluetooth® (Qualcomm, Savari, etc.); however, USDOT and others are committed to DSRC being the primary mechanism for vehicle safety applications. General Motors was the first domestic OEM to commit to inte- grating DSRC-based V2X technology into its newer vehicles, initially planned for its 2017 Cadillac CTS model. Delphi will supply the V2X equipment, which was developed by Cohda and NXP (Yoshida 2014). Level Name Description Human drivermonitors the driving environment 0 No automation The full-time performance by the human driver of all aspects of the dynamic driving task, even when enhanced by warning or intervention systems. 1 Driver assistance The drivingmode-specific execution by a driver assistance system of either steering or acceleration/decelerationusing information about the driving environment and with the expectation that the human driver will perform all remaining aspects of the dynamic driving task. 2 Partial automation The drivingmode-specific execution by one or more driver assistance systems of both steering and acceleration/decelerationusing information about the driving environment and with the expectation that the human driver will perform all remaining aspects of the dynamic driving task. Automateddriving systemmonitors the driving environment 3 Conditional automation The drivingmode-specific performance by an automated driving system of all aspects of the dynamic driving task with the expectation that the human driver will respond appropriately to a request to intervene. 4 High automation The drivingmode-specific performance by an automated driving system of all aspects of the dynamic driving task, even if a human driver does not respond appropriately to a request to intervene. 5 Full automation The full-time performance by an automated driving system of all aspects of the dynamic driving task under all roadway and environmental conditions that can be managed by a human driver. Table 1. Levels of driving automation (SAE International 2014).

9 The federal government has played a significant role in supporting the research, development, and piloting of CV technology. The USDOT Connected Vehicle Safety Pilot Program sought to demonstrate that DSRC-based CV tech- nology was ready for large-scale deployments. Executed in Ann Arbor, Michigan, this program equipped vehicles with vehicle awareness devices, aftermarket safety devices, and retrofit safety devices, and deployed DSRC infrastructure to assess the functional performance of V2V and V2I safety applications (Bezzina and Sayer 2015). USDOT is also currently sponsoring three additional CV pilot deployments in New York, Florida, and Wyoming. The pilot program in New York provides an opportunity to study and evaluate the use of CV technology in a dense urban environment with significant pedestrian and cyclist traffic, in addition to vehicular traffic. The pilot will install in-vehicle equipment on up to 10,000 city and fleet vehi- cles to test V2V applications, such as inter section movement assist and forward collision warning, and will install roadside infrastructure in Manhattan and Brooklyn to test V2I appli- cations, such as pedestrian in signalized intersection and red light violation warning (Galgano et al. 2016). The pilot pro- gram in Wyoming is focusing on applying CV technology along freight-intensive corridors that experience signifi- cant weather-related incidents and delays. DSRC onboard equipment will be installed in a combination of maintenance vehicles, emergency vehicles, and private trucks, and infra- structure will be installed along Interstate 80 to communi- cate road conditions, variable speed limit zones, and detour information (Gopalakrishna et al. 2015). The pilot program in Tampa will evaluate CV technology deployed in a suburban- to-urban corridor that includes managed lanes that experience significant congestion and delays while bringing thousands of vehicles to and from a dense urban center with high pedes- trian traffic. V2V safety applications such as forward collision warning and intersection movement assist will be evaluated, as well as V2I applications such as curve speed warning and transit signal priority (Waggoner et al. 2016). Significant research and standardization have gone into the development of CV technology, specifically related to DSRC. SAE and the Institute of Electrical and Electronics Engineers (IEEE) have been actively working on standards documents for DSRC (SAE J2735, IEEE 1609.2/3/4) and V2V performance (SAE J2945/1). The various DSRC manufactur- ers will be required to certify that their equipment conforms to these standards to ensure interoperability of vehicles from different OEMs using different hardware. USDOT has orga- nized several CV “PlugFests” throughout the country where CV vendors were able to test their devices’ performance, inter- operability with other equipment, and conformance to afore- mentioned standards (Abuelhiga 2013). Regulation, Legislation, and Standards AVs NHTSA (2016b) released the official Federal Automated Vehicle Policy in September 2016, issued “as guidance rather than in a rulemaking in order to speed the delivery of an ini- tial regulatory framework and best practices to guide manu- facturers and other entities in the safe design, development, testing, and deployment of HAVs.” It focuses on HAVs (SAE Levels 3–5) and includes guidance for AV performance, a model for state policy, references to NHTSA’s existing regula- tory tools, and a discussion of potential new regulatory tools that could help the government facilitate the development of AVs. This policy applies to all organizations developing and testing AVs, including OEMs, suppliers, technology firms, and other research and development organizations. The guidance includes a 15-point safety assessment that encourages AV developers to give serious consideration to designing a robust system as it applies to things such as operating domains, minimum risk fallback conditions, data recording and shar- ing, post-crash behaviors, cybersecurity, and ethical consid- erations. The model state policy reaffirms that states retain their responsibilities for licensing and registering vehicles, defining and enforcing traffic laws, and regulating insurance and liability requirements and policies. The framework envi- sions that each state’s AV-related policies and regulations will be administered by a single lead agency and associated tech- nology committee, with stakeholder consultation included. This agency would be tasked with defining and coordinat- ing processes for registering and licensing AVs and their test drivers, issuing test vehicle permits, handling applications for testing on public roadways, and involving jurisdictional law enforcement. The policy also outlines NHTSA’s existing regu- latory tools, which include interpretation letters, exemptions, rulemaking, and enforcement. It also outlines potential new tools and authorities that NHTSA could use, such as safety assurance, where manufacturers are required to provide pre-market testing data and results; post-sale regulation of software changes, where any changes to AV software after the sale are regulated; functional and system safety requirements, where the 15-point safety assessment is made mandatory; and enhanced data collection, which would require enhanced data recorders and reporting requirements. As of June 2017, 18 states—Alabama, Arkansas, California, Colorado, Florida, Georgia, Louisiana, Michigan, New York, Nevada, North Dakota, Pennsylvania, South Carolina, Tennessee, Texas, Utah, Virginia, and Vermont— and Washington, D.C., have passed legislation related to autonomous vehicles. Since 2012, at least 41 states and D.C. have considered legislation related to autonomous vehicles

10 (National Conference of State Legislatures 2016). Some of the adopted legislation relates primarily to terminology and tax- onomy, such as a bill introduced in Louisiana. The adopted legislation in a number of states and districts including Cali- fornia, Nevada, and the District of Columbia authorizes opera- tion and testing of AVs on public roads. In North Dakota and Florida, the legislation calls for studies and pilots of AV tech- nology. Several enacted bills, including those in California and the District of Columbia, require a human driver to be sitting in the driver’s seat while the AV is in operation. The Tennessee bill prohibits local governments from banning vehicles equipped with AV technology. CVs In 1999, the Federal Communications Commission (FCC) set aside 75 MHz of spectrum in the 5.9 GHz band for intel- ligent transportation system (ITS) vehicle safety and mobility applications, and it has been the basis of DSRC research and development ever since. While USDOT and seemingly much of the auto industry remain committed to the use of 5.9 GHz DSRC, technology and telecommunications companies have been recommending that the band be shared between DSRC and Wi-Fi enabled devices. Given the time frame since the original allotment was made and the tremendous advances and changes in technology, FCC recently issued a public notice that invited organizations to submit plans for testing equip- ment within the band and committed to completing that test- ing by early 2017 (Alleven 2016). This shift has caused concern among OEMs and others that altering the plan for 5.9 GHz DSRC could negate years of research and development and could delay a broad deployment of CV safety applications. Even given this potential policy shift, USDOT has contin- ued to move forward with its plans for CVs. In early 2014, NHTSA indicated its intention to move forward with the regulatory process regarding CV technology, specifically V2V communications capability. In August of that year, an advance notice of proposed rulemaking (ANPRM) initiated that pro- cess, which proposed creating a new Federal Motor Vehicle Safety Standard that would require all light-duty vehicles to be equipped with V2V communication capability (NHTSA 2014). This notice followed a significant amount of research funded by USDOT to study the feasibility of using this tech- nology to improve safety for many drivers (Harding et al. 2014). NHTSA developed a notice of proposed rule making (NPRM) based on comments and a readiness report of the ANPRM and issued it in December 2016. While this pro- posed rulemaking signals a positive step toward CV tech- nology beginning to be present on public roads, it does not include any requirements on specific safety or mobility applications that must be running on the equipment. Addi- tionally, CV infrastructure falls outside the scope of the pro- posed rulemaking, meaning CVs will only be able to take advantage of V2I safety and mobility applications if a state or local government or transportation organization has made the commitment to invest and deploy roadside equipment and applications. Relevant Stakeholders This research analyzed potential policy and planning strate- gies that could be used to encourage public- and private-sector interests in advancing the technologies. The private-sector actors that were the focus of this research are producers and consumers of AV and CV technologies: • Producers include automobile manufacturers, technology firms, and Tier 1 suppliers (i.e., Tier 1 companies are direct suppliers of parts to automobile manufacturers). USDOT and university research institutions are also producers of CV technology but not relevant actors for this research because of the focus on private-sector actors. • Consumers include private individuals and private-sector fleet owner/operators. In addition, the research was concerned with the state and local transportation agency perspective in two ways: • Determining the impacts that AVs and CVs might have on these agencies. • Identifying actions that state and local agencies could take to realize societal benefits of the technologies. Per the latter bullet, the study identified 18 policy and plan- ning strategies for consideration by state and local agencies. Report Organization Following this introduction, the report is organized into the following sections: • Chapter 2: Examines potential impacts of AVs and CVs and describes outcomes most beneficial for society. • Chapter 3: Summarizes the role of state and local policy and planning in nudging the private sector to make choices to benefit society. • Chapter 4: Describes and assesses each of the 18 policy and planning strategies (associated assessment tables are pro- vided in the appendix). • Chapter 5: Provides conclusions and recommendations from the research.