Low-Speed Automated Vehicles (LSAVs) in Public Transportation (2021)

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.

38 Literature Review Alessandrini, A., R. Alfonsi, P. D. Site, D. Stam. Users’ Preferences Towards Automated Road Public Transportation: Results from European Surveys. Transportation Research Procedia 3 (2014): 139–144. July 2014. This academic paper includes results and econometric analysis from a stated preference survey on riding collective automated road transport systems versus minibuses. The researchers surveyed users in 12 European cities and found a slightly higher preference for automated options among males rather than females, and that people with higher education levels are more likely to prefer automated options. They also found that users care more about travel time than technology type when choosing transport options. Automotive Information Sharing and Analysis Center. “Automotive Cybersecurity Best Practices.” 2017. The Automotive Information Sharing and Analysis Center (Auto-ISAC) is a community of international original equipment manufacturers, suppliers, and commercial vehicle companies that share information, analysis, and resources relating to automotive cyber- security. This website gives general information about Auto-ISAC objectives and best practices in the area of automotive cybersecurity. Autonomous Vehicle Working Group. Draft Report of the Massachusetts Autonomous Vehicles Working Group v4.0. September 12, 2018. This draft report from the governor- mandated Massachusetts AV Working Group includes a connected and automated vehicle (AV) technology overview, governance considerations, and policy considerations. It mostly focuses on non-LSAV AVs but does contain a chart with a few examples of where LSAVs have been legislated or regulated and notes the examples of deployment in Minnesota; Las Vegas, Nevada; and Gainesville, Florida. The document also has a list of working group members, national and international AV initiative examples, and lists of resources. Bansal, P., K. M. Kockelman, and A. Singh. “Assessing Public Opinions of and Interest in New Vehicle Technologies: An Austin Perspective.” Transportation Research Part C: Emerging Technologies 67: 1–14. 2016. This academic paper on public opinion in Austin, Texas, of general AV technologies aims to help transportation professionals develop smarter and more sustainable travel options that the public will adopt. The study looks at the impact of demographics, built- environment variables, and travel characteristics on the willingness to pay for adding SAE Inter- national Level 3, 4, and 5 technologies to personal vehicles. It does not focus on shared vehicles or LSAVs but is one of the few objective, quantitative studies available to assess public opinion. Results show a willingness to pay $7,253 to add Level 4 automation to a personal vehicle. California Department of Motor Vehicles. Title 13, Division 1, Chapter 1. Title 13, Division 1, Chapter 1 of the California Department of Motor Vehicles regulations includes Article 13: Testing of Autonomous Vehicles. The article includes definitions relating to AV testing, requirements A P P E N D I X A Literature Review and Survey

Literature Review and Survey 39 for a manufacturer’s testing permit, requirements for a certificate of self-insurance, proof of financial responsibility, and other requirements. Dong, X., M. DiScenna, and E. Guerra. “Transit User Perceptions of Driverless Buses.” Transportation: 1–16. 2017. This academic study includes results from a stated preference survey of transit bus riders in Philadelphia about their willingness to ride driverless buses. The researchers use mixed-logit modeling to examine the relationship between passenger types and willingness to ride driverless buses with or without an on-board safety attendant. Overall, younger riders and males are more willing to ride a driverless bus. Of the entire sample, two-thirds expressed willingness to ride with an attendant on board but only 13 percent expressed willingness without anyone on board. EasyMile. Site Assessment Report. Client: City of Arlington. Site: Convention Center to AT&T Stadium. September 2017. This document is a slide set of the site assessment by the vehicle manufacturer (EasyMile) for the off-road path LSAV testing in Arlington, Texas. The document consists primarily of maps, drawings, and charts and an overview, requirements, and analysis of the chosen site and path for the vehicle in the City of Arlington pilot project. Eno Center for Transportation. Beyond Speculation 2.0: Automated Vehicles and Public Policy. January 2018. This report from a transportation policy think tank covers the state of the practice in and recommendations for AV policies at the federal level. The report does not focus on LSAVs but includes general recommendations for policies related to automated transportation options, including needs relating to operational, infrastructure, and security. FTA. Strategic Transit Automation Research Plan. FTA Report no. 0116. January 2018. This 2018 FTA plan for the future of transit incorporation of automated technologies outlines FTA’s research plan for automated technologies. The plan specifically notes Level 4 automated shuttle applications, including the use cases of a circulator bus service and a feeder bus service. It also specifies the use of mobility-on-demand services for use cases that shuttles could also contrib- ute to, such as first mile/last mile, ADA paratransit, and on-demand shared rides. Transit agencies could use the FTA plan to help align their plans and goals with national plans and goals. Fraade-Blanar, Laura, Marjory S. Blumenthal, James M. Anderson, Nidhi Kalra. Measuring Automated Vehicle Safety. Rand Corporation. 2018. This Rand report on AVs and safety out- lines a framework for safety for automated vehicles. It discusses how safety can and should be measured for AV technologies. The report does not focus on low-speed vehicles or shuttles but impresses upon the fact that LSAVs have very specific and limited operational design domains and use cases. FHWA. Traffic Calming ePrimer—Module 2. February 2017. FHWA has a publicly avail- able eight-0module primer online as a resource for traffic-calming practices and information. This is module two of these FHWA materials on safety, focusing on traffic-calming to reduce speeds. Module two is basic information of definitions, reasoning behind traffic-calming, and traffic-calming in context of other transportation issues. This resource backs up the importance of using low-speed vehicles in certain situations. Government Accountability Office. Automated Vehicles: Comprehensive Plan Could Help DOT Address Challenges. GAO-18-132. This study addresses the need for national policy supporting the development and planning for automated vehicles, including SAE levels 1 through 5 technologies and both passenger and freight movement. The GAO study recom- mendation is for the U.S. DOT to develop a comprehensive plan for AVs, which they complied with in the federal automated vehicle policy, Automated Vehicles 3.0: Preparing for the Future of Transportation, preceding and supporting documents. The document does not specify shuttles and low-speed vehicles or their common use cases.

40 Low-Speed Automated Vehicles (LSAVs) in Public Transportation Governors Highway Safety Association. Preparing for Automated Vehicles: Traffic Safety Issues for States. August 2018. This report outlines GHSA recommendations to states for plan- ning for AVs, specifically vehicles with level 3–5 capabilities. The report includes recommenda- tions to incorporate consideration for automated driving systems in the areas of management, traffic laws, and AV testing and deployment. It also includes examples of state policies and relevant resources. The report does not focus on or discuss LSAV applications. Intelligent Transportation Systems Joint Program Office, U.S. Department of Transporta- tion. “ITS Research 2015–2019: Connected Vehicles.” 2018. ITS JPO at U.S. DOT awarded three pilot projects a collective $45 million in 2015 to test connected vehicle dedicated short-range communications technologies. There are reports and updates for each project, as well as general information about applications of connected transportation technologies. This resource is just about connected vehicles, not automated vehicles, and includes research on both vehicle-to- vehicle communication and vehicle-to-infrastructure communication. National League of Cities. Autonomous Vehicle Pilots Across America: Municipal Action Guide. 2018. The National League of Cities developed this guide to help cities take action in regu- lating, tracking, and participating in automated vehicle testing, pilots, and demonstrations. The report outlines local governance issues associated with AVs and gives examples of AV pilot pro- grams and what goes into implementing them. It also includes recommended strategies for city leaders. The report outlines planning for and implementing AV pilots in cities in the context of new technologies, transportation and land-use trends, and state and federal policy frameworks. One of the examples highlighted in the report is the Arlington, Texas, LSAV shuttle. NTSB. Reducing Speeding-Related Crashes Involving Passenger Vehicles. 2017. The NTSB conducted a seminal study on the risks associated with high-speed vehicle travel. In particular, the study examined the causes of and trends in speeding-related passenger vehicle crashes. From 2005 through 2014, vehicle speed was the cause or contributing factor in 31 percent of traffic fatalities. The NTSB found speed to increase both the likelihood of being involved in a crash and the severity of crash injuries. The study recommended five countermeasures for speeding, including (1) speed limits, (2) data-driven approaches for speed enforcement, (3) automated speed enforce- ment, (4) intelligent speed adaptation, and (5) national leadership. LSAVs were ignored as a direct potential solution. This report emphasizes the implications of implementing low- speed service. See also Elvik, Rune. The Power Model of the Relationship Between Speed and Road Safety. 2009. Navya Safety Report: The AUTONOM Era. January 2019. Navya submitted its Voluntary Safety Self-Assessment (VSSA) to U.S. DOT in January 2018. The document is similar to other VSSAs, including basic information about the company, the technologies they use in their vehicle, safety technologies and testing methodologies, and relevant policies. The VSSA includes an entire section on operational design domain, including the variables that must be assessed and the requirement of a site feasibility study before implementation. Nielsen, T. A. S., and S. Haustein. “On Skeptics and Enthusiasts: What Are the Expec- tations Towards Self-Driving Cars?” Transport Policy 66: 49–55. 2018. This academic study includes a survey of Danes on perceptions of AVs and vehicle sharing. The results of the survey of 3,040 people separate the respondents into three categories: 38 percent fell into the category of “skeptic,” 37 percent “indifferent stressed drivers,” and 25 percent “enthusiasts.” This study found, consistent with other studies, that the enthusiasts are more likely to be young, highly educated, urban-dwelling males. Skeptics are more likely to be older, car reliant, and live in less dense areas. Personal ownership of AVs was preferred over shared fleets. Nuro. Delivering Safety: Nuro’s Approach. 2018. This is the U.S. DOT–recommended Voluntary Safety Self-Assessment form from Nuro, an LSAV manufacturer whose vehicles are

Literature Review and Survey 41 designed to carry only goods and no passengers. The VSSA includes standard elements of the company’s vision, vehicle specifications and technologies, safety testing methods, the opera- tional design domain specifications, and high-level policy considerations. In 2020, NHTSA granted Nuro a temporary exemption from the Federal Motor Vehicle Safety Standards (85 FR 7826, February 11, 2020). Pettigrew, S., Z. Talati, and R. Norman. “The Health Benefits of Autonomous Vehicles: Public Awareness and Receptivity in Australia.” Australian and New Zealand Journal of Public Health. 2018. http://doi.org/10.1111/1753-6405.12805. This was an academic study surveying Australians about AV acceptance, salience of AV health benefits, and prompted awareness of these health benefits. Quantitative and qualitative data were generated using an online survey targeting Australians 16 years and older. Results indicated neutral levels of receptive- ness towards AVs and very low salience of health benefits. However, there were more substantial levels of prompted awareness of positive health benefits. The study implied Australians’ like- lihood to accept AVs when understanding the associated health benefits. The study did not focus on low-speed vehicles. SAE J3016TM Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles, Publication Date June 2018. This is an SAE Recommended Practice describing motor vehicle driving automation systems that perform part or all of the dynamic driving task. SAE J3163TM Taxonomy and Definitions for Terms Related to Shared Mobility and Enabling Technologies, Publication Date September 2018. This is an SAE Recommended Practice providing a taxonomy and definitions for terms related to shared mobility and enabling technologies. Securing America’s Future Energy. America’s Workforce and the Self-Driving Future. June 2018. The report commissioned by SAFE supported academic economists’ research on the impact of AVs on the United States workforce. Implications of automated vehicle technologies affecting the workforce are still unknown, as the capabilities and applications of the technology are still in development, but the research suggests that while there may be shifts in workforce skill needs, the overall number of jobs will not change much over time as the transportation sector adopts automated technologies. Simulations in this study show that impacts on the workforce from AVs would be felt starting in the early 2030s but would only increase the national unemployment rate by 0.06–0.13 percentage points at peak impact sometime between 2045 and 2050 before a return to full employment. The report does not look specifically at LSAVs but notes the possibilities of shared fleets. Swigart, Sarah. AAA Survey Summary. 2018. As a sponsor of the Las Vegas Phase I LSAV pilot project in downtown Las Vegas, AAA distributed a survey to riders. The survey asked riders how many stars they would give the experience, if they would recommend it to friends, their perceptions of AVs before and after riding, and gave space for additional comments. The document indicates an overwhelmingly positive response. AAA did not survey people who chose to not board the vehicle or participate in the pilot experience. University of Michigan. Mcity Driverless Shuttle: A Case Study. 2018. This document provides an overview of the Mcity driverless shuttle testing on the University of Michigan campus. The testing included studies on human behavior and AV technologies. The project included pilots with a Navya LSAV on the University of Michigan campus. The report also includes recommendations for others interested in deploying comparable services. The identi- fied lessons learned are (1) set specific project goals; (2) engage stakeholders early; (3) explore legal, regulatory, and insurance options; (4) identify operational environment constraints; (5) conduct your own testing; (6) train safety conductors thoroughly; (7) anticipate challenges; (8) develop an incident-response plan; and (9) establish data needs early.

42 Low-Speed Automated Vehicles (LSAVs) in Public Transportation U.S. DOT. Automated Driving Systems 2.0: A Vision for Safety. September 2017. NHTSA released the first version of their Federal Automated Vehicles Policy Statement (FAVP) in September 2016 and replaced the document the following year with a second version, Automated Driving Systems: A Vision for Safety (2.0 ADS). The statements were written in consultation with industry stakeholders, including automakers, technology firms, state government officials, and experts in the field, with incorporation of feedback from public comments. The 2.0 ADS provides information for the industry on AVs. It provides guidance to manufacturers testing and devel- oping AVs on public roads, clarifies the roles of the federal and state governments in regulating operations, and outlines NHTSA’s existing and potential enforcement mechanisms. The 2.0 ADS encourages manufacturers to publish their VSSA and does not specifically address low-speed or transit applications of AV technologies. U.S. DOT. Automated Vehicles 3.0: Preparing for the Future of Transportation. October 2018. This document is the updated federal guidance for AV development and testing. It builds off the federal guide Automated Driving System 2.0, adding to and not replacing the 2.0 guid- ance. The document expands on transit applications of AVs and the FTS’s role in providing technical assistance for AV applications in transit. It also includes a full section on “considerations for public sector transit industry and stakeholders” and a shorter section on recommendations for the private sector transit industry. The document also notes the potential use cases and existing pilots for LSAVs. Volpe National Transportation Systems Center. Low-Speed Automated Shuttles: State of the Practice. 2018. Volpe National Transportation Systems Center in conjunction with the Intel- ligent Transportation Systems Joint Program Office conducted research on the state of the prac- tice both domestically and abroad of the use of LSAVs. Their research spanned from August 2015 to September 2018. It included the review of multiple LSAV demonstrations and pilots, examination of vehicle designs and transportation service provided, and the governance behind the various programs examined. The Volpe team also convened a working group for agencies involved in LSAV demonstrations and pilots for project insight and to provide a platform for technical exchange between agencies. World Economic Forum. Reshaping Urban Mobility with Autonomous Vehicles: Lessons from the City of Boston. June 2018. A study on AV and urban mobility by WEF and Boston Consulting Group culminated in this report. The authors used the City of Boston as their study ground. Findings included intuitive possible outcomes of reduced vehicle ownership, but possible increased vehicle miles traveled if shared AV fleets absorb enough mode share; shared AV fleets could reduce parking demand; and that shared AV fleets could replace trips in both personal vehicles and public transit options. A global survey deployed to urbanites by the researchers found that respondents were most enthusiastic about AVs to eliminate the need to manually search for and execute the parking task, and overall about 60 percent of respondents said they would be likely or very likely to ride in an AV. The report mentions “autonomous mini- buses,” which are akin to LSAVs, and notes that many German cities and other cities around the world are implementing LSAV pilots. WSP and AECOM. Draft MnDOT Autonomous Bus Pilot Project: Testing and Demonstra- tion Summary. June 2018. To prepare for the operations of an automated shuttle bus in mixed general traffic and cold weather climate conditions, the Minnesota Department of Transportation (MnDOT) conducted an Autonomous Bus Pilot project. The project aimed to safely demonstrate AV technology to stakeholders and capture public interest. Testing focused on identifying challenges and solutions for operating AVs on the MnDOT transportation system. The demonstra- tion of the automated shuttle bus was conducted by EasyMile, a vendor chosen by MnDOT, with oversight from WSP and AECOM. First Transit and 3M also partnered with MnDOT for

Literature Review and Survey 43 operational support. Demonstrations occurred on a 2.5-mi closed low-volume loop. Winter weather testing found snowfall and snowbanks alongside vehicle routes were often detected as obstructions, causing the vehicle to slow down or stop completely. When core temperatures of the shuttle dropped significantly, operations were negatively impacted and charging times increased. Dry pavement conditions and salt spray from treated sections of roadway did not alter shuttle performance. Overall, survey data from the demonstrations showed the public in favor of the Autonomous Bus Pilot project. A Feasibility Study to Explore the Potential for Running Autonomous Vehicles Trails in Cambridge Utilizing the Unique Aspects of Guided Busway. n.d. It is unclear who conducted this feasibility study for AV testing in Cambridge, UK, regarding sharing existing busway right- of-way with autonomous driving systems. The report focuses on the interaction between trans- portation, land-use, and economics, and the potential for new technologies to increase access and mobility. The report touches on LSAV applications, noting the added expense of dedicated right-of-way, which is sometimes, but not always, necessary. The report includes a feasibility analysis and various options for possible programs in Cambridge, including LSAV programs. 49 U.S.C. § 5323(j). 49 U.S.C. §5323(j) requires that federal tax dollars used to purchase steel, iron, and manufactured goods in a transit project are produced domestically in the United States. This includes rolling stock purchases and capital leases. For rolling stock purchases for which the average cost of the vehicle is more than $300,000, the cost of steel or iron produced in the United States and used in the rolling stock frames or car shells may be included in the domestic content calculation, regardless of where it is produced. FTA’s Buy America regulations apply to third-party procurements by FTA grant recipients. LSAVs used for public transit and receiving federal funding must abide by the Buy America law. If the minimum percentage content quote is not met, any FTA funding would be withheld unless a waiver is justified by the FTA. LSAVs not using federal funding do not have to abide by the Buy America laws. 49 U.S.C. § 30114. NHTSA issues the FMVSS; compliance with these standards allows vehicles to operate on public roads. This part of the United States Code allows exemptions to the FMVSS. The FMVSS includes provisions regarding minimum safety specifications for vehicle design, construction, and performance. 42 U.S.C. § 12101 et seq. Public transit agencies (and many private providers) must follow the federal law, the Americans with Disabilities Act (ADA), which includes requirements for transit stations, vehicles, and services. Before any agency or company implements service with LSAVs, they must bring the infrastructure, vehicles, and service into compliance. However, there are no currently available vehicles that meet the specifications of the ADA. The most common shortfall is the lack of securement for a wheelchair inside the vehicle, but some vehicles do not include ramps, a bar to hold on to for stability, or designated areas for people with disabilities. Literature Survey In reviewing the published literature on LSAVs as of 2018, many transit agencies have expressed and demonstrated interest in specific use cases for LSAVs, but few formal studies have been conducted to increase knowledge in the field of LSAVs.1 Understanding how LSAVs operate, their limitations, and their potential is important for transit agencies deciding whether or how to integrate them into their networks. This section examines the limited existing literature 1 Cregger, Joshua, Machek, Elizabeth, et al., “Low-Speed Automated Shuttles: State of the Practice Final Report.” Intelligent Transportation Systems Joint Program Office. FHWA-JPO-18-692. 2018, https://rosap.ntl.bts.gov/view/dot/37060.

44 Low-Speed Automated Vehicles (LSAVs) in Public Transportation on LSAVs and LSAV testing and reflects the information as of the publication of the referenced documents. Topics include the following: • Infrastructure requirements and operational design domain • Vehicle specifications • Safety • Cybersecurity • NHTSA “Temporary Importation of a Motor Vehicle or Equipment Under Box 7 on the HS-7 Form” • Law enforcement • Liability • Accessibility for people with disabilities • Connectivity • Workforce • User acceptance • Stakeholders from the private, public, and research sectors have shown interest in and have paid attention to the short- and long-term tests of LSAVs.2 According to the FTA’s Strategic Transit Automation Research report, AV demonstrations can lead to a better understanding of several important issues, including • Transit operations and maintenance • Fuel and emissions • Service quality • Safety • Passenger experience, comfort, and acceptance • Accessibility • Travel options and mode choice • Fare collection • Cost-effectiveness3 Infrastructure Requirements and Operational Design Domain ODD defines the physical and temporal conditions under which AVs can operate, including roadway types (interstate, local, etc.) on which the ADS is intended to operate safely; geographic area (city, mountain, desert, etc.); speed range; environmental conditions in which the ADS will operate (weather, daytime/nighttime, etc.); and other domain constraints.4 The low-speed shuttles currently or recently deployed generally operate on dedicated or limited access right-of- way and with minimal environmental complexity. Sometimes the added complexity involves only portions of the ODD that are in mixed right-of-way or crossings with vehicular traffic.5 LSAVs require specific infrastructure elements that both support the dynamic driving task by any means of operations (such as clear right-of-way), and that support automated systems. Infrastructure allows the LSAV sensors to understand the environment and inform proper steering, acceleration, and braking. 2 Ibid. 3 U.S. DOT, FTA, “Strategic Transit Automation Research Plan.” FTA Report no. 0116. January 2018, http://www.transit.dot. gov/sites/fta.dot.gov/files/docs/research-innovation/114661/strategic-transit-automation-research-report-no-0116_0.pdf. 4 U.S. DOT, “Automated Driving Systems 2.0: A Vision for Safety.” September 2017, www.nhtsa.gov/sites/nhtsa.dot.gov/files/ documents/13069a-ads2.0_090617_v9a_tag.pdf. 5 Cregger, Joshua, Machek, Elizabeth, et al., “Low-Speed Automated Shuttles: State of the Practice Final Report.” Intelligent Transportation Systems Joint Program Office. FHWA-JPO-18-692. 2018, https://rosap.ntl.bts.gov/view/dot/37060.

Literature Review and Survey 45 As of 2018, the literature review indicated that LSAV pilots and demonstrations have largely occurred on existing limited access right-of-way.6 A dedicated right-of-way for AVs eliminates risks from mixed traffic or from unprotected left-hand turns. In developing LSAV projects, project managers need to account for interactions with other vehicle types, including higher-speed transit vehicles and bicyclists or pedestrians (for example, at intersections or on mixed-use paths).7 Literature shows that state agencies are beginning to examine various aspects of the opera- tional environment. The Commonwealth of Massachusetts AV Working Group established multiple classifications of environmental conditions with various degrees of complexity in order to permit vehicle testing based on the appropriate environment, starting in a closed course and graduating to open public roads.8 At least one, MnDOT, has conducted a test of an EasyMile vehicle in snowy and rain-inclement weather. MnDOT found environmental factors such as snow or rain cause increased localization issues and wheel slippage for the EasyMile vehicle.9 Levels of Automation The SAE-established levels for ADS automation levels 0–5, with Level 0 (No Driving Auto- mation) and Level 5 (Full Driving Automation). While this classification system does not include many nuances in advanced driver-assistance systems, it provides a standard framework to consider various types of vehicle automation.10 In a January 2018 report, the FTA states that LSAVs can perform SAE Level 4 capabilities under certain, carefully planned, ODD and contextual circumstances.11 In 2015, FAST required a GAO study on automated vehicle policy, which outlines the basic state of the practice for automated driver-assistance systems and recommends further planning by U.S. DOT to accommodate new technologies, such as implemented in the recent U.S. DOT AV guidance document. The study focused on passenger vehicles and light duty trucks and does not specifically mention shuttles such as LSAVs but notes that standards for vehicles and infrastruc- ture will change and DOT must respond accordingly.12 Under current law, vehicles operating on public roads are required to adhere to the FMVSS, which have been developed by NHTSA, unless they qualify for an exemption. In 2020, NHTSA granted Nuro a temporary exemption from the Federal Motor Vehicle Safety Standards (85 FR 7826, February 11, 2020) and initiated rule- making with regard to Occupancy Protection Standards in Vehicles with Automated Driving Systems (84 FR 24433, May 28, 2019). 6 University of Cambridge, “A Feasibility Study to Explore the Potential for Running Autonomous Vehicles Trails in Cambridge Utilizing the Unique Aspects of Guided Busway”; see also EasyMile, “Site Assessment Report,” Sept. 2017, on behalf of the City of Arlington for a site between the Arlington Convention Center and AT&T Stadium. 7 Ibid. 8 Autonomous Vehicle Working Group, “Draft Report of the Massachusetts Autonomous Vehicles Working Group” v4.0. September 12, 2018, http://www.mass.gov/files/documents/2018/09/12/DraftReport_AV_WorkingGroup.pdf. 9 WSP and AECOM, “MnDOT Autonomous Bus Pilot Project: Testing and Demonstration Summary.” Draft. June 2018, http://www.dot.state.mn.us/automated/bus/finalreport.pdf. 10 Eno Center for Transportation, “Beyond Speculation 2.0: Automated Vehicles and Public Policy.” June 2019, http://www. enotrans.org/events/webinar-beyond-speculation-2-0/. 11 U.S. DOT, FTA, “Strategic Transit Automation Research Plan.” FTA Report no. 0116. January 2018, http://www.transit.dot. gov/sites/fta.dot.gov/files/docs/research-innovation/114661/strategic-transit-automation-research-report-no-0116_0.pdf. 12 GAO, “Automated Vehicles: Comprehensive Plan Could Help DOT Address Challenges.” GAO-18-132, Nov. 2017, http:// www.gao.gov/assets/690/688676.pdf.

46 Low-Speed Automated Vehicles (LSAVs) in Public Transportation LSAV Characteristics According to research by the Volpe National Transportation Systems Center, LSAVs have the following characteristics: • Fully automated driving (SAE Level 4 or 5, intended for use without a driver). • ODD (restricted to protected and less-complicated environments). • Low operating speeds (cruising speeds around 10–15 mph). • Shared service (typically designed to carry multiple passengers, including unrestrained passengers and standees). • Share the right-of-way with other road users, either at designated crossing locations or along the right-of-way itself.13 LSAVs are not a specific vehicle class. Some are NEVs, which are subject to guidance by NHTSA as well as state and local law. Others are purpose-built vehicles. A third category includes retro- fitted vehicles such as the Cushman 6 deployed at Fort Bragg and the Polaris GEM used by May Mobility and Optimus Ride, as well as pod car.14 As recommended by the GAO in 2017, U.S. DOT did indeed refine guidance for AVs in its 2018 document, Automated Vehicles 3.0: Preparing for the Future of Automation. The document includes multiple references to ongoing LSAV testing, while cautioning local governments to maintain expectations that are “realistic about their limitations,” presumably meaning that agen- cies understand the necessity of the on-board safety attendant, including to operate the vehicle, the maintenance needs, and varied battery draw.15 Some LSAVs currently on the roads in the United States are imported and have been operating with the issuance of an HS-7 waiver from the FMVSS. These vehicles qualify by virtue of the fact that the vehicles are imported and used for research, investigations, demonstrations or training, or competitive racing.16 However, if transit agencies using federal funding decide to adopt the vehicles for service, the Buy America requirements could lead to the required use of American- made vehicles that will have to comply with FMVSS regulations.17 U.S. DOT specifies in its recent AV guidance that they will be updating the FMVSS to account for automated technologies, which could further influence the landscape for procurement and operation of vehicles.18 According to the FTA in its Strategic Transit Automation Research Plan, LSAVs are subject to speed and operational environment limitations and low passenger capacities, so are not yet suitable for broad adoption by the transit industry for higher-capacity services.19 Safety Research indicates that the largest contributor to deaths and serious injuries on the roadway is vehicles traveling at high speeds. Therefore, a significant safety benefit of LSAVs is that they oper- ate at lower speeds. There is a direct relationship between speed and the probability of serious 13 Cregger, Joshua, Machek, Elizabeth, et al., “Low-Speed Automated Shuttles: State of the Practice Final Report.” Intelligent Transportation Systems Joint Program Office. FHWA-JPO-18-692. 2018, https://rosap.ntl.bts.gov/view/dot/37060. 14 Carolinas Alliance 4 Innovation, Website, http://ca4i.com/projects.html. Accessed February 12, 2019. 15 U.S. DOT, “Automated Vehicles 3.0: Preparing for the Future of Transportation. October 4, 2018, http://netchoice.org/ wp-content/uploads/DOT-Guildelines-v3.0.pdf. 16 NHTSA, “Importation of Motor Vehicles and Motor Vehicle Equipment Subject to Federal Motor Vehicle Safety, Bumper, and Theft Prevention Standards.” http://www.nhtsa.gov/sites/nhtsa.dot.gov/files/hs7_r.v.7.pdf. 17 49 U.S.C. §5323(j) 18 U.S. DOT, “Automated Vehicles 3.0: Preparing for the Future of Transportation. October 4, 2018, http://netchoice.org/ wp-content/uploads/DOT-Guildelines-v3.0.pdf. 19 U.S. DOT, Federal Transit Administration, “Strategic Transit Automation Research Plan.” FTA Report no. 0116. January 2018, http://www.transit.dot.gov/sites/fta.dot.gov/files/docs/research-innovation/114661/strategic-transit-automation-research- report-no-0116_0.pdf.

Literature Review and Survey 47 injury or fatality in a crash, for pedestrian safety as seen in Figure A.1. The escalation of risk with the increase of speeds holds for vehicle crashes with other vehicles as well.20 There is no standard definition of “safety” in the specific context of AVs, including LSAVs.21 A 2018 Rand report explores definitions of safety, safety measurement, and communication about safety, limiting the discussion to highly automated vehicles, or vehicles operating under SAE automation Level 4 or 5. Rand divides safety measures into two types: leading and lagging. Lead- ing measures include behavior such as following traffic laws that may prevent crashes, whereas lagging measures measure actual safety outcomes. Figure A.2 shows the proposed measures and when to implement them. Measures such as the proposed Rand measures could play a role in future safety certification for vehicles. Public education and communication around safety may have an impact on willingness to adopt and utilize automated vehicle technologies. Documents such as the voluntary safety self- assessments outlined in the Federal Automated Vehicle Policy 3.0 inform the public on auto- makers’ approach to safety and the technologies they use.22 As of the date of this review, the only manufacturer of an LSAV that had submitted a Voluntary Safety Self-Assessment is Nuro, who develops vehicles exclusively for the transport of goods, and not passengers.23 On February 11, 2020, the NHTSA granted a temporary exemption to Nuro from the FMVSS, the first granted to any AV manufacturer (85 FR 7826).24 Figure A.1. FHWA speed/pedestrian injury severity correlation. 20 NTSB, “Reducing Speeding-Related Crashes Involving Passenger Vehicles.” NTSB/SS-17-01, 2017, http://www.ntsb.gov/ safety/safety-studies/Documents/SS1701.pdf; Elvik, Rune. “The Power Model of the Relationship between Speed and Road Safety.” TOI Report: 1034/2009. Norwegian Centre for Transport Research, Oct. 2009, ISBN 978-82-480-1001-2. 21 Fraade-Blanar, Laura, Marjory S. Blumenthal, James M. Anderson, Nidhi Kalra, “Measuring Automated Vehicle Safety.” Rand Corporation. 2018. ISBN: 978-1-9774-0164-9. 22 Ibid. 23 U.S. DOT, “Automated Vehicles 3.0: Preparing for the Future of Transportation.” October 4, 2018, https://www.transportation.gov/ sites/dot.gov/files/docs/policy-initiatives/automated-vehicles/320711/preparing-future-transportation-automated-vehicle-30.pdf. 24 Nuro, Delivering Safety: Nuro’s Approach. 2018; Nuro, Temporary Exemption from the FMVSS, 85 FR 7826.

48 Low-Speed Automated Vehicles (LSAVs) in Public Transportation Federal Motor Vehicle Safety Standards NHTSA has a legislative mandate under Title 49 of the United States Code, Chapter 301, Motor Vehicle Safety, to issue FMVSS and Regulations to which manufacturers of motor vehicles and items of motor vehicle equipment must conform and certify compliance. Under these standards, vehicles must include standard features; vehicles that do not conform to standards make seek a waiver of FMVSS from NHTSA.25 Law Enforcement Studies agree that states and cities must develop law enforcement policies around AVs, includ- ing for issues around operations and data sharing and access.26 The state of California, which has some of the most stringent laws for AV testing, requires that manufacturers certify that there is a process to hand over operator information to law enforcement in the event of a crash or “any reason.” California also requires manufacturers to inform law enforcement and other first responders of the ODD and protocols to interact with the vehicle and requires that law enforcement interaction plans are reviewed annually.27 Case studies in Appendix D and Appendix E provide further examples of collaboration with law enforcement. Accessibility for People with Disabilities Public transit agencies (and many private providers), must follow the ADA, which includes requirements for stations, vehicles, and services. In implementing LSAV service, public and private providers must ensure that infrastructure, vehicles, and services are compliant with Source: Fraade-Blanar et al., “Measuring Automated Vehicle Safety.” Rand Corporation, 2018. Figure A.2. Integration safety framework. 25 49 U.S.C. § 30114. 26 Governors Highway Safety Associations, “Preparing for Automated Vehicles: Traffic Safety Issues for States.” August 2018, http://www.ghsa.org/sites/default/files/2018-08/Final_AVs2018.pdf. 27 California Department of Motor Vehicles. Title 13, Division 1, Chapter 1.

Literature Review and Survey 49 applicable laws.28 The regulations vary depending on many factors, and little has been written about the topic. The 2018 report on LSAVs conducted by Volpe did not fully address the issue of ADA compliance because of the levels of complexity and rapid changes in the levels of compliance.29 U.S. DOT’s recent Notice of Funding Opportunity for Automated Driving Systems stated that AVs in public transit should include, at a minimum, a ramp and securement. In addition, U.S. DOT is now requiring an input/output interface to allow communication between the vehicle and those who have a vision, hearing, or cognitive impairment. The U.S. Attorney’s Office for the Eastern District of Michigan (as a component of the U.S. Department of Justice) and the University of Michigan reached an agreement in November 2019 related to the university’s autonomous transportation services (as further described in mini case study E.7, Mcity Driverless Shuttle, in Appendix E).30 The agreement was predicated on the lack of accessible features on two vehicles purchased by the university for its autonomous transportation system, which operates on a fixed route during business hours. Title II requires that all new vehicles operating on a fixed route be accessible and there are no exemptions for autonomous vehicles, pilot programs, or research programs. The agreement included the following elements: • Any future highly automated vehicles (HAV) that the university purchases or leases for the Mcity Driverless Shuttle program or any other fixed-route system must be equipped with accessible features as set forth in 49 CFR, Part 38, and 49 CFR 37.7 (a) including, but not limited to, securement devices and lifts, ramps, and other means of access to vehicles. These must be maintained to be in compliance with 49 CFR 37.161. • Additional requirements for development of policies, procedures, and training materials. • Until all HAVs are accessible the university must provide equivalent services to individuals with disabilities. • Convene a research group to evaluate the accessibility of the autonomous transportation services. As of January 2019, some vehicles have a ramp, one has securement, and one prototype included a human–machine interface. The National League of Cities highlighted the Arlington, Texas, LSAV pilot project, attributing their partnership with the vehicle manufacturer EasyMile to the wheelchair ramp included in the design.31 MnDOT also chose to use an EasyMile vehicle with a ramp.32 The Navya AUTONOM vehicle also has a ramp. May Mobility expects to begin production of a vehicle with ramp and securement later in 2019.33 Local Motors and Perrone Motors each produce vehicles with securement and ramps.34 Connectivity LSAVs, like all automated vehicles, depend on loaded maps as well as sensors such as camera, radar, and LiDAR read-ins to navigate the ODD. Some vehicles also communicate with other 28 Americans with Disabilities Act of 1990. Pub. L. 101-336. 26 July 1990. 104 Stat. 328. 29 Cregger, Joshua, Machek, Elizabeth, et al. “Low-Speed Automated Shuttles: State of the Practice Final Report.” Intelligent Transportation Systems Joint Program Office. FHWA-JPO-18-692. 2018, https://rosap.ntl.bts.gov/view/dot/37060. 30 Ackerman, Shannon M., United States Attorney’s Office Eastern District of Michigan, U.S. Department of Justice, “Letter of Resolution D.J. No. 204-37-328, ADA Compliance Review of the University of Michigan’s Driverless Shuttle Program.” Received by C. Ndu Oor, University of Michigan, November 12, 2019. 31 National League of Cities, “Autonomous Vehicle Pilots Across America: Municipal Action Guide.” 2018, https://www.nlc.org/ sites/default/files/2018-10/AV%20MAG%20Web.pdf. 32 WSP and AECOM, “MnDOT Autonomous Bus Pilot Project: Testing and Demonstration Summary.” Draft. June 2018, http://www.dot.state.mn.us/automated/bus/finalreport.pdf. 33 Alysin Malek, interview by research team, May 10, 2019. 34 Gina O’Connell, interview by research team, April 9, 2019; Paul Perrone, interview by research team, April 7, 2019.

50 Low-Speed Automated Vehicles (LSAVs) in Public Transportation vehicles and/or with the physical infrastructure via several communications technologies, including DSRC and 5G cellular. Vehicle to vehicle and V2I communication can take many tech- nological routes, including dedicated short-range communication (DSCR) and cellular vehicle to anything (CV2X) technologies. U.S. DOT has established a robust connected vehicle program to test DSRC technology, and many universities also research connected vehicles.35 However, CV2X is still a nascent technology and much testing is needed to assess capabilities.36 LSAV pilots have shown that without DSRC V2I communication, AVs need extensive visual cues such as additional signage, decoration, and identifiable objects.37 As LSAVs on the road today have an on-board attendant/operator present, communication can also occur between people operating the vehicle and people in other vehicles or exterior to the driving domain without CV2X or DSRC technologies enabled.38 Workforce Implications of automated vehicle technologies affecting the workforce are still unknown, as the capabilities and applications of the technology are still in development. In the short term, LSAV pilots and demonstrations will likely boost workforce opportunities. LSAVs currently require on-board attendants, will continue to need mechanics and technology development staff, and they do not replace existing transit services.39 In the long run, research has suggested that while there may be shifts in workforce skill needs, the overall number of jobs will not change much over time as the transportation sector adopts automated technologies.40 LSAVs specifically require a manufacturing workforce, maintenance workforce, operations specialists, and on-board liaisons/drivers. With the potential elimination of the human driving feature, on-board attendants still fill the role of education, safety oversight, and assistance, including for people with disabilities. Volpe notes the unclear path forward with on-board attendants and states that they will be present for the time being. Their report further notes that some technology is not ready to oper- ate without an attendant present. In instances in which the technology advances or operations can be handled remotely, the attendant position may be eliminated. Alternatively, an atten- dant may remain to assist passengers, for security, and for other duties unrelated to driving the vehicle.41 U.S. DOT plans to further study the impacts of automated vehicles on the workforce as outlined in AV 3.0. The document announced a study to be conducted by the Departments of Labor, Commerce, Health and Human Services, and Transportation.42 35 U.S. DOT, Intelligent Transportation Systems Joint Program Office, “ITS Research 2015–2019: Connected Vehicles.” 2018, http://www.its.dot.gov/research_areas/connected_vehicle.htm. 36 Federal Communications Commission, “Press Release: FCC Allocates Spectrum in 5.9 GHz Range for Intelligent Transporta- tion Systems Uses.” 1999; 47 CFR § 2.106, NG160. 37 EasyMile, “Site Assessment Report.” Sept. 2017. On behalf of the City of Arlington for a site between the Arlington Convention Center and AT&T Stadium; see also WSP and AECOM, “MnDOT Autonomous Bus Pilot Project: Testing and Demonstration Summary.” Draft. June 2018, http://www.dot.state.mn.us/automated/bus/finalreport.pdf. 38 U.S. DOT, FTA, “Strategic Transit Automation Research Plan.” FTA Report no. 0116. January 2018, http://www.transit.dot.gov/ sites/fta.dot.gov/files/docs/research-innovation/114661/strategic-transit-automation-research-report-no-0116_0.pdf. 39 EasyMile, “Site Assessment Report,” Sept. 2017, on behalf of the City of Arlington for a site between the Arlington Convention Center and AT&T Stadium. 40 Securing America’s Future Energy, “America’s Workforce and the Self-Driving Future.” June 2018. https://avworkforce. secureenergy.org/wp-content/uploads/2018/06/SAFE_AV_Policy_Brief.pdf. 41 Cregger, Joshua, Machek, Elizabeth, et al., “Low-Speed Automated Shuttles: State of the Practice Final Report.” Intelligent Transportation Systems Joint Program Office. FHWA-JPO-18-692. 2018, https://rosap.ntl.bts.gov/view/dot/37060. 42 U.S. DOT, “Automated Vehicles 3.0: Preparing for the Future of Transportation.” October 4, 2018, https://www.transportation.gov/ sites/dot.gov/files/docs/policy-initiatives/automated-vehicles/320711/preparing-future-transportation-automated-vehicle- 30.pdf.

Literature Review and Survey 51 User Acceptance There are no peer-reviewed studies published on consumer acceptance of LSAVs specifically. There are polling studies that consistently indicate that AVs generally do not enjoy high levels of consumer confidence. Some studies focusing on the acceptance of AVs in general reveal wider acceptance among younger people, males, and urbanites.43 Although very few studies have examined acceptance in the context of transit, those also show younger users and males as having a higher propensity toward willingness to ride in a driverless bus.44 JTL/Transit Lab has recently completed a study of stated versus revealed preferences of rides for different types of use cases.45 Researchers at the University of Michigan have been surveying LSAV users to gauge consumer attitudes and plan to publish findings.46 To date, evaluations have been based on anecdotal observations and rider surveys, including collection of Net Promoter Scores. Pilots, such as in the City of Las Vegas, report people waiting in line for the opportunity to ride in a vehicle. AAA commissioned the University of Nevada, Las Vegas, to conduct a survey that revealed overwhelmingly positive feedback by people who chose to use the service (they did not survey nonusers).47 May Mobility queries riders on whether they would recommend the service to a friend or colleague and calculates a Net Promoter Score. NPS is a standard consumer survey technique that allows for a comparison between consumer products and services. 43 Bansal, Prateek, et al., “Assessing Public Opinions of and Interest in New Vehicle Technologies: An Austin Perspective.” Transportation Research Part C: Emerging Technologies, 67: pp. 1–14, June 2016, http://www.sciencedirect.com/science/article/ pii/S0968090X16000383?via%3Dihub; Pettigrew, Simone, et al., “The Health Benefits of Autonomous Vehicles: Public Aware- ness and Receptivity in Australia,” Australian and New Zealand Journal of Public Health, 2018, https://doi.org/10.1111/ 1753-6405.12805; Nielsen, Thomas, and Haustein, Sonja. “On Sceptics and Enthusiasts: What Are the Expectations Towards Self-Driving Cars?” Transport Policy, 66, pp. 49–55, 2018, https://doi.org/10.1016/j.tranpol.2018.03.004. 44 Dong, Xiaoxia, et al., “Transit User Perceptions of Driverless Buses,” Transportation, 46, pp. 1–16, May 2017, https:// link.springer.com/article/10.1007/s11116-017-9786-y. 45 Attanucci, John, and Salvucci, Frederick, Personal Interview, JTL Transit Lab, MIT, 2018. 46 University of Michigan, “Mcity Driverless Shuttle: A Case Study,” 2018, https://mcity.umich.edu/wp-content/uploads/ 2018/09/mcity-driverless-shuttle-case-study.pdf. 47 Swigart, Sarah, AAA Survey Summary, 2018.