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.

83 To complement the case studies of completed pilots found in Appendix D, the research team reviewed a wider set of 14 projects that are in various stages of planning and operation. These pilots represent a snapshot of what agencies and stakeholders are undertaking with LSAV shuttles; Table E.1 provides an overview of the projects reviewed. For each of these mini case studies, the research team summarized the stages of planning, procurement/implementation, operations, and evaluation. Interviewees provided insight into planned activities, operations, metrics, and key performance indicators. The research team used both interviews and document reviews to capture the key elements of these pilots. These interviews were less extensive than those for the three main case studies and typically were focused on the project lead. When an interview was not possible, information was gathered from primary and secondary sources. The mini case studies were developed as of October 2019 and statuses are current as of November 2020. A P P E N D I X E LSAV Mini Case Studies # Lead/ Location Status Use Case Partners Principal Objectives E.1 City, Mountain View, CA Planning First/Last Mile VTA Google LEA Elliott Apex Strategies Connect transit employment & residential Provide competitive travel times Provide operational flexibility peak/off- peak E.2 CCTA, San Ramon, CA Testing Completed on Phase 1; Phase 2 underway at GoMentum, with deployment at Babcock in 2021. Circulator First/Last Mile EasyMile GoMentum Station AAA NorCal Local Motors First Transit First/Last Mile Reduce GHG Economic development Minimize Road Widening E.3 RTD, Denver, CO Completed First/Last Mile LC Fulenwider EasyMile Panasonic City/County of Denver Transdev Next Assess LSAV tech in public transit Link employment to transit Table E.1. Mini case studies highlights. (continued on next page)

Table E.1. Mini case studies highlights. E.10 Smart Columbus, Columbus, OH Operational Circulator May Mobility Connect downtown to museums and employment Provide shuttles to inexpensive remote parking Introduce Autonomous Service E.11 RIDOT, Providence, RI Operational First/Last Mile, Shuttle May Mobility First/Last Mile Expand access to economic opportunities Test the feasibility of a public LSAV service # Lead/ Location Status Use Case Partners Principal Objectives E.4 JTA, Jacksonville, FL Operational Circulator EasyMile Navya KPMG Improve safety Enhance economic competitiveness Reduce GHG emissions Increase ridership E.5 HART, Tampa, FL Planning Shuttle Transdev EasyMile Connect riders to transit hub and downtown businesses and attractions Increase transit ridership E.6 City, Chamblee, GA Planning Circulator Stantec Connect to MARTA station and adjacent destinations Reduce parking requirements E.7 Mcity, Ann Arbor, MI Operational Shuttle University of Michigan Navya Understand consumer acceptance Assess various ODDs E.8 Bedrock, Detroit, MI Operational Shuttle May Mobility Improve/increase frequency for an employee shuttle to remote parking Demonstrate viability in mixed traffic Test innovative autonomous technology and brand Detroit as a mobility technology leader E.9 City, Grand Rapids, MI Operational Circulator May Mobility Mobility for the elderly and people with disabilities Test LSAV interactions with other road users E.12 City, Frisco, TX Completed Circulator Drive.ai Frisco Station DCTA Attract tenants with mobility amenities Demonstrate circulator service value for future transit Demonstrate safety of LSAV operations Learn about LSAV technology Reduce car use for midday errands, eating out

LSAV Mini Case Studies 85 E.1 Mountain View AGT Studies for Shared AV Transit City of Mountain View, Valley Transit Authority (VTA), Google Mountain View, CA Project Profile The project will provide fixed-route LSAV shuttles or medium-capacity transit vehicles on a dedicated lane that connects the North Bayshore office district (a.k.a. Googleplex) with downtown Mountain View and first/last mile service to rail stations. This project anticipates a regular or permanent deployment of LSAVs on dedicated roadways, which may include fixed guideways that will mitigate congestion by increasing travel choices. Planning assumptions contemplate headways of approximately 30 seconds with a 20-passenger vehicle. The project goals include • Connect major transit stations with nearby employment/residential. • Limit impacts on existing built environments. • Provide highly competitive travel times compared to auto and traditional transit service. • Provide operational flexibility to change operating modes peak/off-peak. The City of Mountain View (City) and VTA envision a low-speed service using electric vehicles capable of recharging at stations, a capacity of 20–25 passengers/vehicle, and operating in platoons on short headways with planned 25–35 mph operational speeds. VTA is participating in Calstart Next Gen Shuttle to produce a transit-ready, accessible vehicle. Possible routes include North Bayshore to Downtown Mountain View Caltrain/VTA station and Moffet Field VTA station. Planning In 2017, the City, VTA, and Google began planning with a review of automated mobility systems that could provide a “medium-capacity,” fast, and reliable transit service to link office workers in the growing office area with existing Caltrain and VTA stations. Feasibility studies Mini Case Study Highlights Status Planning Use Case(s) First/Last Mile Duration Open-ended Project Lead City of Mountain View Project Partners VTA, Google, LEA Elliott, Kimley-Horn, Nelson\Nygaard, Apex Strategies Vendor TBD Operator TBD Funding Source Local, Private Google Map 2019 Table E.1. (Continued). E.13 Houston METRO, Houston, TX Operational Next Phase Sept. 2021 Shuttle EasyMile Texas Southern University Test LSAV technology integration in public transit Introduce the LSAVs to community E.14 UDOT, Salt Lake City, UT Completed First/Last Mile, Circulator UTA EasyMile WSP Improve access to rapid transit rail Understand how the Utah Transit Agency can integrate this technology with larger network Demonstrate the feasibility of serving each use case # Lead/ Location Status Use Case Partners Principal Objectives

86 Low-Speed Automated Vehicles (LSAVs) in Public Transportation assume automated technology to connect with North Bayshore; expected ridership is 9,000 riders per day with 300 riders every 10 minutes of estimated peak-hour transfers from Caltrain and local ridership. As part of feasibility studies, in 2017 the City held community meetings, and additional City Council study sessions were focused on technology options, corridor characteristics, and evalua- tion criteria. In 2019, the City also held stakeholder meetings with Mountain View Transportation Management Association, VTA, and Google. A project website was set up to provide information and updates on the studies. Safety planning includes emergency evacuation of elevated guideways and inclusion of a dedi- cated guideway and lanes to limit potential hazards and increase safety. Development of protocols will occur at a later stage. Procurement and Preparation Procurement: The City Council approved the next phase of detailed studies and the plans are to issue an RFP in mid- to late 2019. Federal and State Approvals: The State of California now permits on-road operation of AVs subject to appropriate approvals and reporting requirements. Both federal requirements and vehicle types are expected to change before the launch of service. Infrastructure: The team does envision constructing physical infrastructure improvements along the selected route to include dedicated lanes, street redesign, and potential elevated guideway. Operation and Safety Planning Development of an SOP will occur at a later stage. Evaluation and Data Collection The project team is currently designing evaluation criteria for technology options to include system design, configuration, and alignment based on • Capacity, with objectives for capacity-holding peak demand of 330 passengers (capacity must also include on-board bicycles). • Connections to other transport modes. • Travel time, with an objective of approximately 7–15 minutes end to end. • Accessibility, with objectives including ADA compliance and ride comfort. • Expandability and adaptability, including potential new route extensions. • Environmental management factors to include impact on protected wetlands near North Bayshore area. E.2 CCTA GoMentum Station/Bishop Ranch Shuttles Contra Costa Transportation Authority San Ramon, CA Project Profile Beginning in 2016, the Contra Costa Transportation Authority (CCTA) launched a series of LSAV shuttle pilots, the first operated in the Bishop Ranch Business Park. The purpose

LSAV Mini Case Studies 87 was to test in order to evaluate LSAV technology and validate its safety. The eventual vision is using these shuttles in a first/last mile use case, especially for BART stations with limited parking. The program has not yet delivered rides to the public, and the EZMile shuttles testing at Bishop Ranch is completed. Local Motors and others are still undergoing testing at GoMentum Station. Deployment at Bishop Ranch is scheduled for 2021. The project goals include • Address the key first/last mile connections at BART with parking lots full by 6:00 a.m. • Increase transit ridership. • Reduce single-occupancy vehicle trips. • Reduce greenhouse gas emissions. • Increase economic development. • Provide alternatives to freeway widening. From 2016 to 2019, the EasyMile EZ10 Gen-1 vehicles were tested by CCTA and its consultant at GoMentum Station and on a route within Bishop Ranch Office Park. When the vehicle leases ended in July 2019, CCTA and AAA NorCal (serving as manager of GoMentum Station and project manager) took delivery of three Local Motors Olli LSAVs to continue the project. CCTA intends to expand the routes to serve BART rapid transit stations in the Bay Area. Planning CCTA and GoMentum Station planning included • Identifying the federal, state, city, and local approvals and permits required to operate LSAVs on public roads. • Holding two public hearings under the auspices of the City of San Ramon. The hear- ings introduced CCTA’s transportation goals for the LSAV testing/pilot, provided information on expanding transit accessibility (especially for underserved communities), and described the potential for increasing economic development. At that time, the Califor- nia Department of Motor Vehicles regulations required a plan for interaction with local law enforcement. According to CCTA, these hearings were also done to comply with guidance from the California Public Utilities Commission (CPUC), which has the authority to issue permits for AV carriers that provide transportation services to the public. • In 2016 and 2017, CCTA and EasyMile operated in a closed circuit at GoMentum Station, an automated vehicle proving ground, and on a closed course at Bishop Ranch. • As part of the hazard assessment, EasyMile and CCTA identified concerns with unprotected left turns. To mitigate this risk, the required procedure for the EasyMile safety driver was to take manual control of the LSAV when not on a predetermined route and/or in the case of any unsafe behavior (of the vehicle) or external situation. • CCTA, EasyMile, and the San Ramon Police Department developed an Emergency Plan and Accident Report Plan for pilot activities, as required by the CPUC. The outline of the emer- gency response is included as a resource. The project team is assessing and planning for future pilot phases to open LSAV shuttle services to public ridership and longer routes to connect with BART and other rapid transit in the Bay Area. Specifically, the project team is pursuing approval to cross a major suburban arterial road (Camino Ramon). CCTA proposed a modified signal time to allow for safe crossing of the LSAV through the intersection. Mini Case Study Highlights Status Testing Use Case(s) Circulator, First/Last Mile Duration Open-ended Project Lead CCTA Project Partners EasyMile, GoMentum Station, Advanced Mobility Group AAA NorCal Vendor EasyMile Local Motors Operator CCTA, EasyMile First Transit Local Motors Funding Source State, Regional, Private Google Map 2019

88 Low-Speed Automated Vehicles (LSAVs) in Public Transportation Procurement and Implementation Procurement: The original EasyMile EZ10 vehicles were procured through an informal consul- tative process. At the time of the original lease in 2016, it was very early in the development of LSAV shuttle technology, and EasyMile was one of the few vendors providing vehicles. Funding: The public and private funding includes direct and in-kind resources from local/ regional and state agencies, a private landowner/developer, an engineering firm, and EasyMile, the vehicle vendor. These included • Sunset Development (owner of Bishop Ranch Business Park): in-kind storage and charging, funding for EasyMile vehicle lease. • Stantec (engineering firm): in-kind engineering services. • EasyMile (vehicle vendor): provided “additional dollars” and in-kind resources to support the project. • BART (regional transit agency): $250,000 contribution, through Measure J sales tax. • Bay Area Air Quality Management District: $1 million grant contribution. • State of California: $3.5 million discretionary line item in budget. Federal and State Approvals: The vendor’s LSAVs are imported and did not meet the FMVSS regulations. CCTA was the first agency to apply for FMVSS waivers for LSAVs in July 2016; the waiver expired in July 2019 after the time limit of three years. At the state level, the California Department of Motor Vehicles, the California Environ- mental Protection Agency, and the California Air Resources Board approvals were sought. Before the pilot carried members of the public, the operator needed to seek approval from the CPUC (as discussed). CCTA will require a temporary FMVSS exemption for the Local Motors Olli vehicles to operate on public roads. The agency is working with its federal lobbyist and expects regulatory approval within 3 months to 1 year. They are testing these vehicles at GoMentum Station in the interim. Infrastructure: CCTA installed temporary signage to alert pedestrians and drivers of the presence of AVs at the Bishop Ranch office park. CCTA is installing transit signal prioritization technology at key intersections as well as RSUs and DSRC. The latter will provide V2I communication. Operations LSAV shuttles are being tested and evaluated under controlled conditions in parking lots and circulator roads in the office park, as well as at a dedicated testing facility at GoMentum Station. First Transit was the primary operator of the two EZ-10 vehicles at Bishop Ranch. CCTA and EasyMile conducted field training with the police and the city/traffic engineer in accordance with the “Emergency Plan and Accident Plan.” The project team conducts monthly meetings with the Bay Area’s metropolitan planning organization to keep stakeholders informed. With the knowledge it has gained with the EZ10 vehicles, CCTA expects to be able to move more quickly in testing the Local Motors Olli vehicles for the same functions. The plan is for commercialization and wider deployment after the testing phase has been completed.

LSAV Mini Case Studies 89 Evaluation and Data Collection The project team is evaluating how the vehicle can be verified for safe operation, based on meeting a list of functional challenges in real-world conditions to ensure the LSAV vendor provides agreed-upon deliverables. The project team partnered with the LSAV vendor to develop evaluation criteria, test scenarios, and safety goals. The functional challenges include a list of 155 maneuvers for the vehicle to perform safely, including making basic turns, approaching and leaving a station, responding to obstacles, handling intersections like T-junctions and roundabouts, and responding to traffic signals. A full list of test elements is provided in the Practitioner Guide of this report. Building from its work at Bishop Ranch, CCTA and its partners were awarded an Auto- mated Driving System Demonstration Grant from U.S. DOT in September 2019. The funding will support the testing work at the GoMentum Station along with the deploying of vehicles and technology in three projects: First/last mile shuttles in Walnut Creek, hospital-linked accessible transportation in Martinez, and connected vehicle infrastructure along I-680 in San Ramon. E.3 Denver RTD 61AV Shuttle Denver Regional Transportation District (RTD) Denver, CO Project Profile The 61AV Shuttle was an operational 6-month pilot designed to evaluate the capa- bilities of an LSAV shuttle within one public transit integration use case and demonstrate how RTD could safely implement an AV project on public roadways in the State of Colorado. The project goals included • Safely introduce automated vehicle technology on a public roadway in the Denver metro area. • Assess the reliability and availability of an AV shuttle vehicle and its suitability for a public transit application. • Provide additional first/last mile service to/from an RTD bus/rail station, business and residential areas. • Align the interests of multiple stakeholders in order to advance the LSAV project. The LSAV selected for the program was an EasyMile EZ10 Gen-1 with a capacity of 10 to 12 passengers and a built-in automatic access ramp. The LSAV shuttle started January 29, 2019, traveling on a 1-mi one-way loop route connecting the 61st and Pena Station on the RTD East Rail Line to the Panasonic build- ing, an emerging apartment complex, and the Pena park-n-ride owned by Denver Inter- national Airport. The route is on public roads with the LSAV traveling at approximately 10 mph to 12 mph. The LSAV shuttle was a free service for public use and intended trip purposes included employment access and access to transit. This was one of the first deployments nationally of an LSAV operating on a public roadway and integrated with a transit agency’s service offering. The 61AV Shuttle pilot project was completed on August 2, 2019. Mini Case Study Highlights Status Complete (August 2019) Use Case(s) First/Last Mile Duration 6 Months Project Lead RTD Project Partners EasyMile, Transdev, Panasonic, LC Fulenwider, City and County of Denver Vendor EasyMile Operator Transdev Funding Source Transit, Private Denver RTD website 2019

90 Low-Speed Automated Vehicles (LSAVs) in Public Transportation Planning The demonstration project was a partnership between RTD, EasyMile (the automated vehicle provider), TransDev (the operator and provider of the on-board Customer Service Ambassador), Panasonic, and LC Fulenwider (co-developers of Pena Station Next), and the City/County of Denver (including Public Works and Denver International Airport). Stakeholders, including members from EasyMile, RTD, City of Denver, and local developer LC Fulenwider, reviewed and assessed alternative routes. EasyMile also conducted a Site Assess- ment Report as part of its standard project planning process to assess hazards and recommend risk mitigation measures to include infrastructure modifications. As a result of these assessments, Denver RTD selected the route for this project to avoid higher-speed arterial roads. Both RTD and the Colorado Autonomous Vehicle Task Force met with the Denver Police Department and Denver Fire Department before the launch. These briefings provided information on the vehicle, including battery location and emergency-response protocols. Although this project has concluded, RTD has submitted a grant application to U.S. DOT and is planning to partner with the City and County of Denver as well as EasyMile on another project to provide a first/last mile connection between the University of Denver and the adjacent light- rail station. Procurement and Implementation Procurement: RTD contracted directly with EasyMile as the vendor for this pilot project. At the time of the pilot, EasyMile was establishing its North American headquarters in Denver (at the location of the pilot), and the company offered to contribute in kind through reduced lease costs. Funding: RTD leveraged existing relationships and contracts with a private developer/property owner, a transit service provider, and an LSAV vendor to develop, fund, and provide services for the pilot. The private developer provided cash and in-kind support. The contracted transit service provider reduced operation and maintenance costs related to the project. The vendor provided reduced lease costs. RTD contributed cash and staff time. Federal and State Approvals: The EasyMile EZ10 vehicle was imported from France; RTD and the vendor obtained the required NHTSA approval. Under legislation passed in 2017, Colorado State’s Autonomous Vehicle Task Force developed the regulatory framework and approved RTD and the vendor’s application to operate on public streets.1 The City of Denver also participated in site/route review, assessment, and selection through its Public Works group. Infrastructure: To mitigate safety risks identified in the traffic engineering analysis, RTD chose a route that excluded an arterial with high traffic volumes and 45 mph speed limits. The project team worked with the Colorado Department of Transportation to determine and post appro- priate signage that was compliant with the Manual on Uniform Traffic Control Devices for a slow-moving vehicle. RTD/City of Denver posted signage noting the presence of AV operations and that identified LSAV shuttle stop locations. 1 The Colorado Autonomous Vehicle Task Force, comprising representatives from the Colorado Department of Transporta- tion, Colorado State Patrol, and Colorado Department of Revenue/Division of Motor Vehicles, is charged with reviewing applications to operate autonomous vehicles on public roadways. Colo. Rev. Stat. § 42-4-242.

LSAV Mini Case Studies 91 The developer added ADA-compliant bus stops with concrete pads for enhanced access. There was no need for connected vehicle technology for vehicle-to-infrastructure connectivity. Operations The vendor and RTD prepared an SOP that included a daily operations schedule, operations protocols, role descriptions, and detailed operational checklists. The LSAV operation was limited to one charge cycle for one vehicle ranging between 10 and 12 hours and enough reserve for contingencies. The charge cycle was dependent on the temperature and level of air conditioning utilized. Operations were based at the vendor’s North American headquarters located along the route. A contract with Transdev provided an on-board safety attendant known as a Customer Service Ambassador, who took manual control of the vehicle when necessary to deviate from the AV’s predetermined route for obstructions such as construction vehicles. RTD and EasyMile publicized the service through traditional and social media. RTD created a specific 61AV page with scheduled information and frequently asked questions. 61AV informa- tion was included in RTD’s trip planner and could be accessed through RTD’s real-time vehicle location tracker for rider-monitoring. RTD and the Colorado Cross-Disability Coalition completed a successful test of the auto- mated ramp for people with disabilities. Evaluation, Data Collection, and Lessons Learned Metrics included uptime, number of emergency stops, errors, and battery charge at the start/ end of the service day. On a daily and weekly basis, the vendor and operator provided RTD data on ridership, per- centage of LSAV uptime, and vehicle errors. (An example of one of these reports is provided in the Practitioner Guide.) The project team did not conduct customer surveys because the sample size would be statistically insignificant. During the period of operation, current service availability ranged from 45 percent to 99+ percent, with weather and road construction being the primary factors in limiting LSAV shuttle service availability. After the pilot concluded, the project team developed a list of lessons learned. They noted that stakeholder alignment was a key element. The project team explained that to launch each stage of the project takes significant time, including regulatory requirements, contractual, project pre- planning (route and schedule), and marketing/communication, as well as the implementation itself. It was important that marketing and communications activities were proactive between project team members, stakeholders, and the public. The project team also noted that the budget needed to include all project elements at the outset, as there may be small components that are overlooked. The project team noted that the EZ10 Gen-1 vehicle may not yet be reliable enough; potential improvements discussed included enhanced sensors (which are available in the Gen-3 vehicle), improved battery life, and the addition of a heater. The RTD 61AV Autonomous Shuttle Demonstration Project Final Report (August 2019) can be accessed in the Practitioner Guide Resources, Section 5.6.

92 Low-Speed Automated Vehicles (LSAVs) in Public Transportation E.4 JTA Ultimate Urban Circulator Jacksonville Transit Authority (JTA) Jacksonville, FL Project Profile The Ultimate Urban Circulator (U2C) is a multiphase project to test and pilot LSAVs as a replacement of Jacksonville’s existing automated people mover (APM) system (the Jacksonville Skyway). Phase 1 is operational and the focus of this study; notes are also included for ongoing or planned activities in the later phases. The project goals include • Improve safety. • Maintain a state of good repair for transit infrastructure. • Enhance Jacksonville’s economic competitiveness. • Protect the environment and reduce emissions. • Improve quality of life through innovative technologies and new partnerships. • Increase ridership (expect growth to 2,500 passengers daily). EasyMile and Navya fielded one LSAV shuttle each for Phase 1 activities. The EasyMile EZ10 and Navya AUTONOM shuttles both have a capacity of 12 to 15 passengers/shuttle. The Phase 1 route is a short test track located behind TIAA Bank Field and consists of a 0.3-mi dedicated road segment. JTA is planning for a 3.2-mi route along Bay Street (Phase 2) and a 2.5-mi dedicated elevated guideway on the existing Skyway with undeter- mined mixed-traffic extensions (Phase 3). Testing conducted in Phase 1 will inform the final determination of later phase routes consisting of public roadways with a dedicated lane and multiple operational domain designs. Planning JTA completed a policy study as part of the North Florida Transportation Planning Organization’s “Smart Region Master Plan” to evaluate the potential of automated vehicles in the region. Separately, there is an ongoing JTA planning for modernization of the Jacksonville Skyway titled “Transit Concept and Alternatives Review” that is being conducted in cooperation with the Florida Department of Transportation (FDOT). The ridership projections for the U2C transit route concepts were developed using the FTA’s Simplified Trips-on-Project Software travel demand model. The North Florida Transportation Planning Organization adopted the outcomes of this work in June 2018 resulting in the Bay Street Innovation Corridor being adopted as a locally preferred alternative. The JTA secured funding through a U.S. DOT BUILD grant for their Phase 2 Bay Street Innovation Corridor, and this funding is the foundation for its program of pilots. KPMG conducted a risk assessment of the entire U2C program in November 2017 and a copy is included in the Practitioner Guide. JTA plans to mitigate the identified risk factors by develop- ing detailed project risk registers at key stages of project development and will include planning for concepts, permit, PR procurement/procurement activities, constructions, and operations/ maintenance. JTA coordinated with local law enforcement to develop a safety operation outline, which includes the most common scenarios during vehicle testing (Phase 1), and shared the outline Mini Case Study Highlights Status Operational (Ph. 1) Planning (Ph. 2 and 3) Use Case(s) Circulator Duration Phase 1: 18 months Phase 2: 24 months Phase 3: APM deployment, open-ended Project Lead JTA Project Partners EasyMile, Navya, KPMG Vendor EasyMile, Navya, TBD Operator JTA Funding Source Federal, State, Transit

LSAV Mini Case Studies 93 with the vendors’ vehicle safety operators. A copy of this outline is provided in Section 5.5 of the Practitioner Guide. Procurement and Implementation Procurement: Procurement is through a formal process by JTA, utilizing a “Request for Information” and RFPs for vendors to participate. Funding: The project team combined local JTA funds with a U.S. DOT BUILD grant (for Phase 2) and supplemental FDOT funding. Federal and State Approvals: Vendors and JTA applied for FMVSS exemptions for Phase 1 testing of the dedicated LSAVs. The State of Florida specifically allows automated vehicle opera- tion (316.85 of the State Uniform Traffic Control) and no additional permits are required. Infrastructure: JTA is the primary project sponsor and is pursuing approvals from the City of Jacksonville on civil infrastructure on the corridor. Signage for testing of LSAVs was installed for Phase 1 operations. Operations A formal SOP has not been developed for Phase 1 but is under development for Phase 2 and 3 by leveraging lessons learned from the ongoing LSAV testing. JTA has conducted online surveys for riders and provided presentations to stakeholder groups, including elected officials, policymakers, community leaders, and citizens. Multiple public forums were held to provide more opportunity for public input. The Jacksonville Chamber of Commerce developed a supporting press release and one of the third-party operating vendors published a press release regarding Phase 1. Evaluation and Data Collection JTA currently monitors the performance of the three vendors based on automated control performance (e.g., average speed, consistency, disengagements). This is accomplished through frequent testing events/demonstrations where the public is invited to participate. JTA has designed the phased project to test out LSAV technologies with the public to determine their willingness to ride in LSAVs before committing to a large change of fixed infrastructure on the Skyway. JTA is planning for an “integrated data exchange” to collect, manage, and analyze information from sensors and automated vehicles. During this planning process, development of the exact process, access, and sharing protocols will occur to provide a streamlined way to collect and share real-time data between operators, the transit agency, and City of Jacksonville government along the Bay Street Innovation Corridor. E.5 HART Marion Street Transitway AV Shuttle Hillsborough Area Regional Transit Authority (HART) Tampa, FL Project Profile In 2015, the Hillsborough Area Regional Transit Authority began planning the Marion Street Transit AV Shuttle, a fixed-route service from HART’s central transit hub to the commercial

94 Low-Speed Automated Vehicles (LSAVs) in Public Transportation core of downtown Tampa. Starting October 2020, Beep, the autonomous program man- ager and operator, began operation of a Navya shuttle to operate on a dedicated transit- way on a public right-of-way in mixed traffic with conventional vehicle buses. Operation is approved for 12 months, with an option for a 1-year extension. The project goals include • Improve the connection between HART’s central transit hub and downtown busi- nesses and attractions. • Demonstrate and test LSAV technology feasibility. • Increase transit ridership by improving the accessibility of the transit network to key downtown destinations. The LSAV vendor will provide an LSAV shuttle capacity of six or more passengers per shuttle. The shuttle vehicle ramp will provide mobility impaired individuals with full access. The planned route is shared with conventional vehicle buses and will have headways of 10 minutes or less along the Marion Street route. The project will create eight predeter- mined stops at bus shelters to include the Marion Transit Center, Cass Station, City Hall Station, Fort Brooke Station, Washington Station, Kennedy Station, Federal Station, and Floridian Station. The LSAV shuttle is expected to travel at 15 mph. Planning In 2015, HART introduced the concept of an AV shuttle in Downtown Tampa in its U.S. DOT Smart Cities Challenge grant application. In 2015, the Florida Department of Transportation (FDOT) awarded HART funding for the project. In 2016–2017, HART prepared for a 1-year pilot to connect people from their central transit hub to the commercial core of downtown Tampa. HART contracted with technical consultants to determine the feasibility of operating an LSAV shuttle on their busy downtown transitway and required physical/digital infrastructure improvements to include cameras and RSUs. In February 2018, HART introduced the concept of automated shuttles on the Marion Street Transitway through public demonstrations with May Mobility. Procurement and Implementation Procurement: HART issued an RFP to solicit a vendor to provide a turnkey solution, that is to say for a shuttle and operations on the Marion Street Transitway. Before launch of the service, HART is requiring vendors to submit a detailed Quality Assurance/ Quality Control plan detailing vehicle manufacturing, operations, and maintenance. Further, HART is requiring the vendor to provide a Preliminary Hazard Assessment and a Test Plan for the LSAV shuttles to clearly demonstrate safe operations. HART also requires a detailed incident protocol from the vendor before launch and operations. Funding: FDOT is funding this project. Federal and State Approvals: In August 2019, the vendors sought necessary federal approvals. Since 2012, the State of Florida has allowed automated vehicles to operate on public roads and no additional permits or approvals are required. In June 2019, the State of Florida passed a new law which authorized the operation of automated vehicles without a safety driver, if the AVs meet stipulated insurance and safety requirements. Mini Case Study Highlights Status Planning Use Case(s) Shuttle Duration 12 Months Project Lead HART Project Partners TBD Vendor TBD Operator TBD Funding Source State

LSAV Mini Case Studies 95 Operations The LSAV shuttle will operate on a regular schedule between 6:00 a.m. to 7:00 p.m. Monday through Friday. The SOP has not yet been developed, but all operations will comply with requirements in the RFP. Evaluation and Data Collection The vendor will provide HART with agreed-upon operational data over the course of the project. Reported data will include ridership, quality of service, the performance of the auto- mated system, and incident reporting. HART intends to track LSAV ridership, operational hours, quality of service (downtime), number of vehicle incidents, emissions, and AMS energy/carbon intensity. E.6 City of Chamblee Self-Driving Shuttle City of Chamblee Chamblee, GA Project Profile The City of Chamblee is planning a 12-month pilot to deploy an LSAV shuttle to connect its growing downtown to MARTA and adjacent destinations while reducing the parking demand in downtown Chamblee. The project goals include • Determine feasibility of deploying LSAV shuttles on local streets. • Understand how to utilize LSAV shuttles to reduce parking needs. The intended LSAV shuttle fleet size is 1 to 2 vehicles with a capacity of between 6 and 15 passengers per shuttle. The routes under consideration would support two LSAV shuttles operating simulta- neously with headways between 7.5 and 15 minutes based on expected demand. They will likely include core segments along Peachtree Road in downtown Chamblee, with potential extensions to Peachtree Station shopping center, the Assembly, Chamblee Plaza, Keswick Park, IRS/Center for Disease Control, and Dekalb-Peachtree Airport. Planning The City of Chamblee contracted a transportation planning consultant to complete the 2018 Chamblee/Self-Driving Shuttle Feasibility Study (Feasibility Study) and Concept Plan. This Feasibility Study and Concept Plan cover potential routes, operational parameters, and cost estimates. A market assessment was included as part of this process to determine ridership populations, but not a demand analysis; instead, a combination of data sources was used, including MARTA ridership data and ESRI Business Analysis Online. The Chamblee Mobility Plan is currently under review and outcomes from these planning activities are being discussed by the City of Chamblee government for potential implementation. The City of Chamblee submitted a funding application for the recent U.S. DOT ADS Notice of Funding Opportunity and was not awarded a grant. Mini Case Study Highlights Status Planning Use Case(s) Circulator, First/Last Mile Duration 12 Months Project Lead City of Chamblee Project Partners Stantec Vendor TBD Operator TBD Funding Source Federal (Pending), Local

96 Low-Speed Automated Vehicles (LSAVs) in Public Transportation An informal risk assessment was conducted as part of the Feasibility Study and Concept Plan focusing on speeds and intersection complexity along potential routes. Planned mitigation involves a safety driver for each LSAV and improved infrastructure. Stakeholder workshops were conducted to engage community groups and organizations to include MARTA, the Atlanta Regional Commission, the mayor/city council, local businesses, developers, residents, and city staff. An online survey about AVs was conducted with more than 100 Chamblee residents participating. Planned signage included in the Chamblee Mobility Plan will increase marketing and outreach along the proposed main corridor route and at MARTA station. Procurement and Implementation Procurement: This pilot will use a formal procurement process to select the vendors. Funding: The funding mix and amounts are not yet established but may include City of Chamblee general funds and potential MPO funding. They also applied for a U.S. DOT ADS Demonstration grant, but in the September 2019 announcement, they were not one of the selected applications. Federal and State Approvals: If required, vendors and the City of Chamblee will apply for FMVSS waivers for shuttles after procurement and the vendors are selected. The State of Georgia has specifically allowed AV operation in its motor vehicle code and does not require permits. The City of Chamblee controls most streets being considered for implementation. MARTA controls the station site potentially needed for LSAV turnaround and will require approvals before pilot launch. Infrastructure: Physical improvements may be required to include improved signage, lane markings, and signal timings to manage crossings. Operations Initial operational parameters were developed in the Feasibility Report with operations envisioned for a 14-hour service day, 6 days a week. The selected LSAV vendor will work with the City of Chamblee to develop formal operational plans, training, and safety protocols. Evaluation and Data Collection The project team is evaluating potential LSAV pilot routes considering the following: • Number of residents along the route • Number of jobs along the route • Route length (number of trips per hour with two vehicles) • ODD compatibility with LSAV • Increase in transit service coverage (derived from comparative metrics based on other similar projects) The data collection plan is to be determined but will likely include data collected by the vendor and provided to the City of Chamblee for appropriate analysis and reporting. The City of Chamblee will measure consumer satisfaction through ridership and rider satis- faction surveys. It will also use consumer willingness/interest measured during planning through online surveys and workshops.

LSAV Mini Case Studies 97 E.7 Mcity Driverless Shuttle University of Michigan Ann Arbor, MI Project Profile The Mcity Driverless Shuttle is an operational 24-month pilot to deploy LSAVs to provide a live transportation service on the University of Michigan (UM) research campus. Mcity is a UM-led public–private partnership to accelerate advanced mobility vehicles and technologies. This applied research project aims to better understand human accep- tance, trust, and behavior when riding in a driverless shuttle or interacting with one on the road. The project goals include • Study how passengers react to the driverless shuttle to gauge consumer acceptance of the technology. • Observe technology’s performance in the real world, based on specific types of inter- actions in a variety of weather conditions. • Understand how trust in shuttle technology changes over time, and how an inter- action with the shuttle (either riding or as a pedestrian) changes feelings about AVs. • Track ridership to provide insights on parking usage and campus destination. The vendor fielded a fleet of two AUTONOM LSAV shuttles with capacity of up to 11 passengers per shuttle. The LSAV shuttles travel at 10 to 15 mph along circulator roads at UM North Campus Research Complex. A planned extension to Lurie Engineering Center along a mixed- traffic street was canceled because of construction. Planning An Mcity research team selected the specific route based on the manufacturer’s specifications for ODD and then premapped the route. The Mcity research team tested the Navya shuttles in a variety of weather and lighting conditions within the Mcity Test Facility to ensure safe operation. The research team put together a series of challenges like those on the selected route. The challenges included a cyclist moving in the same direction, a car crossing from the right, a car crossing at a roundabout, a pedestrian approaching, and a car driving in the opposite direction in proximity. The Mcity research team did not conduct external public engagement but facilitated meetings with internal university stakeholders. These included campus stakeholders, the Ann Arbor police department, and the University of Michigan police department to ensure the appropriate functional planning was accomplished and emergency procedures were accounted for. Procurement and Implementation Procurement/Funding: The pilot is funded through internal University of Michigan research funding and received required internal approvals. The LSAV fleet was purchased with financial support from Mcity corporate members. Therefore, formal procurement processes that could be compared to public agency processes were not required. Federal and State Approvals: The University Office of General Counsel received an FMVSS waiver from NHTSA. The State of Michigan specifically allows operation of automated vehicles Mini Case Study Highlights Status Operational Use Case(s) Shuttle Duration 24 Months (ending December 2019) Project Lead Mcity Project Partners University of Michigan, Navya Vendor Navya Operator Mcity Funding Source University, Private Route Map (University of Michigan website)

98 Low-Speed Automated Vehicles (LSAVs) in Public Transportation on public streets (Public Acts 332-335 of 2016). The University of Michigan owns all the land used for pilot operations; therefore, no additional permits were required for operation. Infrastructure: Signage was installed at stops and along the route to inform drivers of the presence of driverless shuttles. ADA Compliance: In November 2019, the University of Michigan and the Department of Justice reached an agreement related to the ADA requirements for compliance of autono- mous vehicles purchased by the University and operated on a fixed route. The agreement included, among other requirements, that all future vehicles must comply with Title II acces- sibility requirements for disabled individuals and until that is provided for all autonomous vehicles, equivalent services must be provided to those individuals, including those using wheelchairs. Operations Detailed operational protocols and procedures were developed that covered actions to include shuttle start up, regular operation, and incident response. Operational procedures also detailed employee sign-in steps, vehicle cleaning, shuttle start up, running the data-acquisition system, checking route safety, ridership rules, and other parameters. Daily and weekly tasks were delegated to the conductors as part of their start-up and shutdown procedures. Safety checklists are used to include incident/collision response and provide immediate access to step-by-step emergency procedures and vital information such as proof of insurance, government exemption letters, route maps, and shift schedules. A three-step training process was set up for LSAV operators using the Mcity Test Facility closed course to allow safety operators to become familiar with the vehicle. Mcity developed a website accessible by the public to provide information, schedule, and key safety-related messages. The University of Michigan handled all marketing for the pilot and local media to include MLive, the Michigan Daily, and other state/regional/local outlets that have reported on pilot activities. Evaluation and Data Collection The Mcity research team uploads camera and sensor data into the Mcity Cloud, at which time the processing engines count passengers, anonymize/blur faces, characterize weather conditions, and classify interactions. These data are then provided to researchers and Mcity partner companies via a web browser interface with filtering and analysis tools. Mcity worked with JD Power, an automotive market research firm, to design a user survey for riders of the shuttle. The aim is to understand the attitudes of riders and nonriders toward driverless shuttles, including longitudinal tracking to better understand how perceptions change over time. UM and Mcity are also part of the Michigan Mobility Collaborative that received an award in September 2019 of a U.S. DOT Automated Driving System Demonstration Grant to fund AV safety testing and deployment in the region. In addition to UM and Mcity, the members of the Michigan Mobility Collaborative include the City of Detroit, the Michigan Department of Transportation, the Michigan Economic Development Corporation, the American Center for Mobility, Wayne State University, Ford Smart Mobility, LLC, and Deloitte LLP.

LSAV Mini Case Studies 99 E.8 Bedrock Detroit Parking LSAV Shuttle Bedrock Detroit Detroit, MI Project Profile In June 2018, the Detroit-based commercial real estate firm Bedrock launched an LSAV shuttle service on a public right-of-way in mixed traffic to provide its employees connections from the Bedrock offices to remote parking in downtown Detroit. The 1-year contract for service has been renewed. The project goals include • Improve frequency for shuttle service on a regularly scheduled route connecting employees to offsite/remote parking. • Assess feasibility of extending the roll-out of LSAV shuttle service to other routes currently operated by conventional vehicle diesel-powered shuttles. • Position Detroit as a mobility technology leader. May Mobility fielded a fleet of five LSAV shuttles based on the Polaris GEM-6 with a capacity of five passengers per shuttle (one seat is taken by the safety driver). Planning Bedrock Detroit, the project sponsor, explored opportunities to use innovative mobility technologies to address employee shuttles to parking locations. Bedrock Detroit decided to field and evaluate an LSAV shuttle in downtown Detroit. Bedrock Detroit selected a local LSAV vendor and oversaw testing in October 2018 to determine the vendor’s LSAV shuttle’s ability to operate in mixed traffic in downtown Detroit. Bedrock Detroit and the vendor conducted safety planning and identified risks on the route and determined mitigation through a regular fixed-route shuttle service on a premapped route, an in-vehicle safety driver, and real-time vehicle monitoring from a central control center. The project team led planning with Detroit Police for incident- and emergency-related activities. Procurement and Implementation Procurement/Funding: Bedrock Detroit funded the project through internal approval and procurement processes. No additional external or formal procurement processes were used. Federal and State Approvals: The State of Michigan specifically allows operation of AVs on public streets (Public Acts 332–335 of 2016). No additional permits were required for operation on Detroit public streets. No further federal government approvals were required. Infrastructure: Cameras were installed at key intersections along the route to provide sensing redundancy for the vehicles. No other physical or digital improvements or changes were needed and no DSRC RSUs were installed. Operations The LSAV shuttles travel on public roadways between Bedrock Detroit’s One Campus Martius office building and the Bricktown Garage, with peak-hour service before and after the normal working day hours. Mini Case Study Highlights Status Operational Use Case(s) Shuttle Duration Open-ended Project Lead Bedrock Detroit Project Partners May Mobility Vendor May Mobility Operator May Mobility Funding Source Private

100 Low-Speed Automated Vehicles (LSAVs) in Public Transportation The LSAV shuttles operate in automated mode on premapped or programmed fixed routes. When not on this service route, the safety driver takes manual control of the vehicle. LSAV shuttles operate with an in-vehicle safety driver and are monitored in real time from the vendor’s local central control center. The vendor’s control center also serves as the location for LSAV storage, charging, and maintenance. The vendor published an SOP, which includes a daily schedule for regular operations during the 15-hour service day. The SOP also includes operational protocols, safety procedures, operational checklists, role descriptions, shift changes activities, and procedures for servicing and charging LSAVs. The published SOP also outlines how spare LSAV shuttles are used and swapped in to ensure continuous operations. LSAV shuttles are charged at the vendor’s local base of operations ½ mi south of the service route. The vendor provided safety attendants with training that included familiarization with AV technology, as well as safety operations best practices. Bedrock Detroit provided all marketing for the LSAV shuttle service, project and service information on its website, and posted signage at the two LSAV shuttle stops. Evaluation and Data Collection The vendor collects data at its operations base in downtown Detroit and its engineers analyze the data and refine automated systems accordingly. The vendor provides Bedrock Detroit with reports on ridership, operational performance, and rider satisfaction. The vendor does not provide information about disengagements and automated technology performance. The vendor surveys customers on a regular basis to derive their Net Promoter Score (NPS), an industry-standard measure of whether riders would recommend the service to others. (Customer satisfaction for the service consistently registered a +60 NPS score.) E.9 Grand Rapids Autonomous Vehicle Initiative City of Grand Rapids Grand Rapids, MI Project Profile The Grand Rapids Autonomous Vehicle Initiative (AVGR) is a pilot that began in July 2019, was suspended for COVID-19, and relaunched in August 2020. It offers LSAV shuttle services to the general public in downtown Grand Rapids. The circulator provides service every 20 minutes along the DASH West route of the City of Grand Rapids (City) public transportation system. New sites will be launched in Q1 2021. The project goals include • Determine feasibility and gain insight into how automated vehicles impact existing urban structures. • Address how automated vehicles improve or impact mobility for the elderly and people with disabilities.

LSAV Mini Case Studies 101 • Gather data and study LSAV operations safety in real-world scenarios to include inter- actions with vehicle traffic, bicycle riders, and pedestrians on city streets. • Engage the local community through open sessions for local stakeholders to provide community visioning, goal setting, plan making, and recommendations for public investments specific to automated vehicles. Planning Community leaders and local businesses came together under the banner of the AVGR. This consortium, led by the City of Grand Rapids, determined that an LSAV shuttle was feasible. This was confirmed through discussions and outreach with the City’s preferred vendor, May Mobility. In the second half of 2018, the City of Grand Rapids, the Rapid (the local transit agency) and their partners developed the concept for a downtown circulator to run along the DASH West route. Procurement and Implementation Procurement: The public–private consortium was formed through a series of discus- sions throughout 2018; these included participation of the selected LSAV vendor. The vehicles were selected through this informal consultative process. Funding: The consortium, consisting of nine companies, the City of Grand Rapids, and the State of Michigan (through the Michigan Economic Development Corporation), provided funding for this pilot. Members of the consortium provided in-kind contribu- tions, cash, and logistics support such as LSAV shuttle storage and charging facilities. Federal and State Approvals: The State of Michigan specifically allows LSAV testing and operations on public roads without further permits. The City of Grand Rapids was part of the project management team and no additional permits were necessary for operating on city streets. No further federal government approvals were required. Operations The fixed-route LSAV circulator runs along a 3.2-mi looped route in downtown Grand Rapids with 22 stops. This route follows an existing downtown circulator (the DASH West) and provides regular service that runs every 20 minutes from 7:00 a.m. to 7:00 p.m. Tuesday through Friday, and 10:00 a.m. to 7:00 p.m. Saturday. In 2020, based on rider requests and trends, the service schedule will shift to operate Monday through Friday. The LSAV shuttle operates in mixed traffic along this route and navigates 30 traffic lights and 12 turns (including three left turns). The LSAV shuttle service is provided to the public free of charge. The vendor prepared an SOP before launch. The SOP includes a daily schedule for regular operations during 12-hour service days 5 days a week. The SOP also includes detailed safety procedures, operational checklists, protocols, role descriptions, shift changes, and procedures for servicing and charging LSAVs. The SOP also outlined how spare LSAVs would be used and swapped in as required to ensure continuous operations. The vendor developed comprehensive training for safety attendants, with theory and practical sections to ensure they understood the technology and could provide a high level of service safely. Mini Case Study Highlights Status Operational Use Case(s) Circulator Duration 12 Months Project Lead City of Grand Rapids Project Partners City of Grand Rapids, Consumers Energy, Faurecia, Gentex, May Mobility, PlanetM/MEDC, Rockford Construction, Start Garden/Seamless, Steelcase Vendor May Mobility Operator May Mobility Funding Source Local, Private Route Map (©Mapbox © Open Street Map)

102 Low-Speed Automated Vehicles (LSAVs) in Public Transportation The signage includes existing DASH bus signage along with a May Mobility sign. Information panels with a route map, customer service information, and hours of operation were installed in December 2019. The City of Grand Rapids manages all marketing and promotion, leveraging its website and social media to describe the shuttle service, and answers frequently asked questions. The initially launched vehicles are not accessible to wheelchairs. However, the May Mobility team will begin offering a wheelchair-accessible vehicle (WAV) in parallel with other vehicles beginning in December 2019. The WAV service will be accessed by calling a customer service number posted on information panels at stops and online. Evaluation and Data Collection The metrics to evaluate project goal achievement include ridership and customer satisfaction. Community workshops held during the year-long pilot will gauge feedback from citizens. All data are collected by the vendor at its operations center in Grand Rapids and provided to vendor headquarters for automated system refinement. The vendor surveys its customers on a regular basis to derive their Net Promoter Score. Information about disengagements and automated performance is kept confidential by the vendor. Key ridership and customer satisfac- tion data are provided to the AVGR consortium. E.10 Smart Columbus Smart Circuit Smart Columbus Consortium (Smart Columbus) Columbus, OH Project Profile The Smart Circuit was a downtown circulator pilot project sponsored by Smart Columbus and DriveOhio that launched in December 2018 and completed in September 2019. The shuttle deployment was the first of two automated vehicle projects in a portfolio of more than 30 projects for Smart Columbus. The shuttle was designed for a downtown circulator use case to give central Ohio residents and visitors a firsthand experience with the mobility technologies of the future and a chance to learn more about automated vehicles’ capabilities and suitability to address real mobility challenges in communities. The project goals included • Conduct pilot, evaluate results, and subsequently scale/sustain/sunset effort. • Engage stakeholders and partners appropriately at every step of the process. • Successfully and safely deploy an automated vehicle technology to better connect residents and visitors in the Scioto Mile area. • Educate residents and visitors to build public trust and support for automated vehicle technology. The second phase will focus on expanding mobility in an underserved community in northeast Columbus. May Mobility fielded a fleet of three LSAV shuttles based on a Polaris GEM-6 with a capacity of six passengers/shuttle. The LSAV shuttles traveled along a circular 1.4-mi route as part of the “Smart Circuit” along the Scioto Mile in Downtown Columbus. The route had four stops including the Smart Columbus Experience Center, Bicentennial Park (along the river), the COSI museum, and the National Veterans Museum. Mini Case Study Highlights Status Completed Use Case(s) Circulator Duration 10 months Project Lead Smart Columbus Project Partners May Mobility Vendor May Mobility Operator May Mobility Funding Source State, Local, Private Route Map (Columbus Partnership)

LSAV Mini Case Studies 103 Planning Smart Columbus competed for the U.S. DOT Smart City Challenge Grant and beat out 77 applicant cities to receive $50 million in funding. Columbus’s plan called for the use of AVs to solve first/last mile challenges in order to improve quality of life for residents. The project team conducted outreach meetings as part of the Smart Columbus overall project plan. Smart Columbus assessed the feasibility of LSAV shuttles in downtown as part of the U.S. DOT Smart City grant submission process. The vendor conducted further assessment and testing in late 2018 to determine the suitability of a specific route, how the vehicle would turn, where it would stop, and if any additional infrastructure would be needed for reliable operation. The pilot project team led incident- and emergency-response planning discussions with Columbus Police before launch. Procurement and Implementation Procurement: A formal RFP for turnkey LSAV shuttle operations was issued by the Ohio Department of Transportation/DriveOhio on behalf of the Columbus Partnership. It was a 1-year contract to mobilize, deploy, and provide 10 months of service with two 1-year options. May Mobility was selected. Funding: The funding for this pilot was provided by private investors who contributed to the Columbus Partnership and DriveOhio. Federal and State Approvals The State of Ohio specifically allows LSAV testing and operations on public roads without further permits. The City of Columbus was part of the project management team and no additional permits were necessary for operating on city streets. No further federal government approvals were required. Operations The fixed-route LSAV shuttle service operates on roads that were remapped, providing a more controlled operation. When not on this service route, the safety driver takes manual control of the vehicle. As a further control to mitigate risk, vehicle status and performance are tracked from the vendor’s local operations center in Columbus. • The vendor prepared an SOP before launch. The SOP included a daily schedule for regular operations during the 16-hour service days, 7 days a week. The SOP also included detailed safety procedures, operational checklists, protocols, role descriptions, shift changes, and pro- cedures for servicing and charging LSAVs. The SOP detailed how spare LSAV shuttles are integrated into the active fleet to ensure continuous operations. • The vendor provided safety attendants training that included familiarization with AV tech- nology, as well as safety operations best practices. Signage was installed at each stop along the route, and information was provided at the Smart Columbus Experience Center downtown. Smart Columbus created a website describing the LSAV shuttle service; it managed user engagement and marketing activities. The operations of the Smart Circuit ended on September 27, 2019.

104 Low-Speed Automated Vehicles (LSAVs) in Public Transportation Evaluation and Data Collection The project metrics included ridership, the percentage of time the vehicle is operating in automated mode, and customer satisfaction. The vendor collected data at its operations center in Downtown Columbus and its engineers analyzed the data and refined automated systems accordingly. The vendor surveyed customers on a regular basis to derive their Net Promoter Score. Key ridership and customer satisfaction data were provided to Smart Columbus. Information about disengagements and automated performance was kept confidential by the vendor. During the operating period, which ran from December 2018 to September 2019, more than 15,000 people rode the shuttles. Using lessons learned, Smart Columbus launched a second phase that concluded in late 2020. Separately, the Ohio Department of Transportation, under DriveOhio, was awarded an Automated Driving System Demonstration Grant by U.S. DOT in September 2019. That project, “Deploying Automated Technology Anywhere in Ohio,” is focused on highway truck platooning to the northwest of Columbus as well as rural ride-hailing and paratransit applica- tions in Southeast Ohio. The project has a data-sharing partnership with Smart Columbus and the Smart Circulator. E.11 RIDOT “Little Roady” AV Shuttle Pilot Project Rhode Island Department of Transportation (RIDOT) Providence, RI Project Profile Launched in May 2019, the “Little Roady” LSAV Shuttle provides first/last mile connection to transit and intercity rail and serves business parks and other employ- ment centers. The almost 6-mi route connects Olneyville with downtown Providence on public right-of-way and operates in mixed traffic. The shuttle concluded in March 2020; the final report is due December 2020. The project goals include • Provide better first/last mile connections with existing transportation modes. • Expand access to economic opportunities in Olneyville along the river corridor. • Demonstrate the feasibility of a connected automated vehicle (CAV) service open to the public. The LSAV vendor fielded six LSAVs with capacity for six passengers/LSAV shuttle that operate at speeds of 20 to 22 mph. Planning Established in 2018 by RIDOT, the Rhode Island Transportation Innovation Part- nership conducted preliminary studies and vendor testing (late 2018 through early 2019) in the Quonset Business Park south of Providence. Following this study and testing, RIDOT began a pilot project in the Woonasquatucket River corridor between Down- town Providence and Olneyville to demonstrate the feasibility of a public LSAV service in a Mini Case Study Highlights Status Operational Use Case(s) First/Last Mile, Shuttle Duration 12 Months Project Lead RIDOT Project Partners May Mobility Vendor May Mobility Operator May Mobility Funding Source State Little Roady route map (RIDOT website)

LSAV Mini Case Studies 105 dense urban area in all weather conditions. The initial pilot is planned for 1 year with an option to extend for an additional 2 years. The pilot project team led incident- and emergency-response planning discussions with Providence Police before launch. Procurement and Implementation Procurement/Funding: RIDOT used funding from the Volkswagen EPA settlement and FHWA research funds for this pilot project and procured the LSAV shuttle services through a formal RFP in mid-2018. Federal and State Approvals: The vendor’s LSAV shuttles are FMVSS compliant and do not require further approval. The State of Rhode Island has no specific AV statute and this pilot project is being coor- dinated by the state government by RIDOT. The project team is in active discussions with the City of Providence and no additional permits are necessary for operating on city streets. Infrastructure: Signage was installed at each of the nine stops along with posted information at Providence Station. Public Engagement: RIDOT conducted public outreach meetings before launch to explain the nature of the service. Operations The vendor prepared an SOP before launch. The SOP provides for a daily schedule for the six LSAV shuttles in regular operations during 12-hour service days (6:30 a.m.–6:30 p.m.), 7 days a week. The SOP also includes detailed safety procedures, operational checklists, protocols, role descriptions, shift changes, and procedures for servicing and charging LSAVs. The vendor provided training to safety attendants that includes familiarization with AV tech- nology as well as safety operations best practices. The fixed-route LSAV shuttle service operates on roads that were premapped, providing a more controlled operation. When not on this service route, the safety driver takes manual control of the vehicle. As a further protection, vehicle status and performance are tracked from the vendor’s local operations center in Providence. RIDOT maintains oversight on the Little Roady project and is handling public marketing and media queries on the project. Evaluation and Data Collection The metrics to evaluate project goal achievement include ridership, operational reliability, and customer satisfaction. The vendor collects data at its operations center in Providence and its engineers analyze the data and refine automated systems accordingly. The vendor surveys customers on a regular basis to derive their Net Promoter Score. Ridership and customer satisfaction data are provided to RIDOT. The vendor does not provide RIDOT information about disengagements or the amount of time the vehicle operates in automated mode.

106 Low-Speed Automated Vehicles (LSAVs) in Public Transportation E.12 Frisco Drive.ai Self-Driving Shuttle City of Frisco, Frisco Station, Denton County Transportation Authority, Drive.ai Frisco, TX Project Profile In July 2018 and March 2019, the City of Frisco sponsored a weekday (on-demand) LSAV shuttle service in an office and retail/entertainment district along a 1-mi route with programmed stops at speeds of under 35 mph. The project goals included • Attract tenants to the district by providing a differentiated service to provide effective mobility as densities increase. • Demonstrate the need for a circulator service and lay the groundwork for future transit service (Frisco currently has no regular transit service). • Demonstrate safe operations to the public. • Understand the requirements of automated vehicle technology, including infra- structure needs. • Determine whether car usage would be lower if LSAV service provided an effective midday circulator for errands and lunch options. Drive.ai operated a fleet of modified Nissan NV-200 vans, which accommodated three riders in a bench seat and provided 6,000 rides over an 8-month period. Planning Under the auspices of the Frisco Transportation Management Association, a partner- ship was formed between the City of Frisco (City), the Denton County Transportation Authority (DCTA), Frisco Station Partners, the HALL Group, and the Star to improve connectivity between several mixed-use developments in Frisco’s North Platinum Corridor. In addition to solutions such as improving walkability, ridesharing, and con- nected vehicles (with traffic signal data sharing), FTMA sought to introduce driverless vehicle service between the Star, HALL Park, and Frisco Station developments. Initial planning focused on identifying and reviewing vehicle providers. The City identified two key factors in its selection: Drive.ai offered the pilot without charge to the City and an alternate vehicle was a neighborhood electric vehicle deemed unsuitable to travel on a section of the route with a posted speed of 45 mph. DCTA executed a memorandum of understanding with Drive.ai, which included insurance requirements to carry public passengers and logistical considerations (such as charging and storage). The final route was shaped around vehicle specifications provided by Drive.ai. The team assessed hazards by reviewing all crash data for this area, parking, and bicycle usage. The project team held a tabletop exercise that included the LSAV vendor, emergency services, and visits to fire stations. Drive.ai prepared information on standard operations and incident response, which was shared with emergency responders and public safety officials. In addition to public briefings with City Council, the project team hosted two outreach sessions with Drive.ai briefing on the project demonstrating the vehicle. Procurement and Implementation Procurement: The vendor was selected through an informal consultative process that included a variety of automated vehicle providers. Mini Case Study Highlights Status Completed Use Case(s) Circulator Duration 8 Months Project Lead City of Frisco TMA Project Partners Drive.ai, Frisco Station, Denton County Transportation Authority Vendor Drive.ai Operator Drive.ai Funding Source Private

LSAV Mini Case Studies 107 Funding: Drive.ai funded the project for the initial 6-month pilot and 2-month extension. Federal and State Approvals: No further federal or state approvals were required. Infrastructure: The project team wanted to include a pilot of DSRC technology to provide redundancy to the LSAV sensors. Drive.ai’s position was that it did not require V2I connectivity and preferred to rely on detailed 3D mapping of the route. DCTA provided new signage on private and public property to inform road users and to direct passengers to pick up spots. Operations The fixed route of the shuttle was operated from 11:00 a.m. to 7:00 p.m. Monday through Friday with adjusted winter times of 10:00 a.m. to 5:00 p.m. The vendor required the project team to have a system to notify them about anything happening in the route area that could affect the right-of-way of the LSAV shuttle (e.g., construction or other work order). Based on that information they adjusted operations accordingly. The presence of delivery trucks in the fire lane was found to affect automated operations. Evaluation A total of 6,000 trips were made on the shuttle during the 8-month pilot. The average number of passengers per ride was 1.8. Information received from riders included requests for an expansion to more locations. The Texas Transportation Institute assessed consumer acceptance and satisfaction through an online survey. The results indicated • Demographics, attitudes, and behaviors influence acceptance and trust of AVs. • If an AV has advanced driver-assistance systems, acceptance is raised. • Interacting with AVs, such as through seeing, entering, and taking a ride in an AV, raises acceptance further. • Acceptance and trust differ significantly across types of automated vehicles. • Level 5 (Full Driving Automation) vehicles as on-demand shuttles or ride hailing are preferred.1 Drive.ai terminated its operations in Frisco in March 2019. The City was a key member of the Texas A&M Transportation Institute grant application for safety performance assessment of on-demand AV service under U.S. DOT’s Automated Driving System grant. Its application was not awarded funding. E.13 Houston METRO University District Circulator Pilot Houston METRO, Texas Southern University Houston, TX Project Profile Houston METRO (the Houston region’s transportation and public transit provider) launched a 6-month LSAV pilot program in June 2019 on the Texas Southern University’s (TSU) Tiger Walk, an off-street mixed-use pathway. The route is approximately 1 mi long and focused on 1 Zmud, Johanna, “What the Public Really Thinks About Automated Vehicles: Evidence from Survey Research,” Auto- mated Vehicle Symposium. Orlando, Florida. 2019, https://s36.a2zinc.net/clients/auvsi/avs2019/Public/SessionDetails.aspx? FromPage=Sessions.aspx&SessionID=3404&SessionDateID=45.

108 Low-Speed Automated Vehicles (LSAVs) in Public Transportation the transportation of TSU students, faculty, staff, and visitors along the path. This is part of a set of projects including the AV grocery delivery Nuro in lieu of paratransit trips and 160 Memorial Express and completed creation of an AV Bus specification for a full-size AV bus using the APTA bus standard, for launch in Sept. 2021. This phase will connect the TSU campus to a nearby light-rail station. The project goals include • Understand how to integrate LSAV technology into public transit as an attractive option for Houston METRO patrons. • Determine insights into human behavior, which can help train METRO staff to work better with these vehicles. • Invite the public to imagine a transportation future which includes LSAV systems. Houston METRO decided to mitigate risk by engaging procurement and legal professionals early, inviting vendors to review their technologies/deployments, and connecting with agency colleagues across the country who had experienced LSAV deployments. Planning Beginning in 2017, Houston METRO, the City of Houston, Houston-Galveston Area Council (H-GAC), Harris County, and the Houston Office of the Texas Department of Transportation established a working group looking at AV deployment as a regional transportation solution. However, after further review, it was decided that AVs on a university campus would be more feasible than a more complex deployment (e.g., open freeway) in the short term and would provide an opportunity to evaluate the new technology in a semi-controlled environment. Stakeholders collectively decided on TSU as a promising first pilot location. The TSU appeal included its well-regarded transportation research institute and a major pedestrian mall in the center of campus. In considering the campus, the team identified the potential for a First/Last Mile use case, with an eventual connection to the nearby light-rail station. First Transit was selected after a competitive procurement process by offering a turnkey approach featuring an EasyMile Gen-2 LSAV. Before the commencement of operations, EasyMile and Houston METRO conducted a workshop on safety design and safety assurance, a workshop on automated vehicle fleet operations and maintenance, and an automated vehicle transit system conceptual design review. Procurement and Implementation Procurement: Houston METRO was working as part of the Texas Innovation Alliance with the H-GAC (the local MPO) on a joint group procurement for AV services for local government entities. While Houston METRO originally intended to use that group procurement, the timing did not align with the project timeline. Houston METRO then issued a formal RFP for AV services, the Autonomous Vehicle Demonstration Pilot. The procurement process closed in July 2018. First Transit was then selected and offered two vehicles for Houston METRO to select, either EasyMile or Navya. METRO selected EasyMile, and First Transit provided an EZ10 Gen-2 vehicle for Houston METRO’s use, according to the agreed-upon contract terms. Funding: The budget for this initial Phase 1 pilot was $250,000 and was provided by Houston METRO. Federal and State Approvals: This EZ10 vehicle was not FMVSS compliant and required a waiver. First Transit was responsible for obtaining the FMVSS waiver for the vehicle. Mini Case Study Highlights Status Operational Use Case(s) Shuttle Duration 6 Months Project Lead Houston METRO Project Partners Houston METRO, Texas Southern University, Houston-Galveston Area Council, City of Houston Vendor EasyMile Operator First Transit Funding Source Private, Transit Texas Southern University Campus

LSAV Mini Case Studies 109 Operations The fixed-route LSAV shuttle service goes back and forth along an approximately 1-mi route on the campus of TSU, along the Tiger Walk pedestrian mall. There are four stops, which are indicated with METRO-branded signage: at a technology classroom building, the student center, the recreation center, and the university library. Operations during the initial Phase 1 pilot are for 9 hours per day Monday through Friday, from 8:00 a.m. to 2:00 p.m. and 5:00 p.m. to 8:00 p.m. The service is open to university students, staff, and visitors; all passengers are required to sign a consent form before riding. A process was established for repeat riders, so they only must complete the forms once. First Transit and Houston METRO prepared an SOP based on their prior experience with AVs, which includes operational checklists and protocols. A safety attendant is on board each shuttle to monitor operations and answer questions. Evaluation and Data Collection Data is collected by First Transit and by TSU’s Center for Transportation Training and Research. The TSU team is being funded by a research grant from H-GAC and is focused on getting feedback from users on their comfort and willingness to use the service and recommend it to others. The metrics that Houston METRO requested from First Transit are ridership information, the number of users at each station that get on/off the shuttle, the number of customers that need assistance, and the number of disengagements. The other metric being assessed, in partner- ship with the Idaho National Laboratory, is energy consumption and the battery range of the vehicle while in service. To date, the vehicle has been well received by users and the data collection is going well. Houston METRO believes that the experience on this project will be informative for the region and nationally. E.14 Utah Self-Driving Shuttle Utah Department of Transportation (UDOT); Utah Transit Authority (UTA) Salt Lake City, UT Project Profile UDOT and UTA are conducting a pilot program for LSAVs, originally planned for a year, which launched in April 2019 and concluded in September 2020, following COVID-19. This program consisted of a series of three 5-week pilots designed to evaluate the capabilities of an LSAV shuttle to connect to fixed-route transit in a variety of operating environments and locations. The Final Report is due out by the end of 2020, along with an evaluation of consumer acceptance. The project goals include • Improve access to UTA’s FrontRunner commuter rail and TRAX light-rail networks. • Understand better how UTA can integrate LSAV technology with its larger network. • Demonstrate the feasibility of serving each specified use case. • Have a broad discussion with the public about autonomy, specifically focused on trust and safety. avshuttleutah.com Mini Case Study Highlights Status Operational Use Case(s) First/Last Mile, Circulator Duration 12 Months Project Lead UDOT Project Partners UTA, EasyMile, WSP Vendor EasyMile Operator UTA Lead, EasyMile Assist Funding Source State

110 Low-Speed Automated Vehicles (LSAVs) in Public Transportation The LSAV shuttle selected for the program is an EasyMile EZ10, with a capacity of 12 passengers per shuttle. The LSAV shuttle will travel routes in multiple locations to include Farmington Station Park, circulator roads at the Canyons Resort, an office park in Salt Lake City, the Utah State Capitol, the University of Utah, the Mountain America Expo Center, a hospital site in the Salt Lake City metro area, and one location in St. George in southern Utah. Planning The project team (UDOT and UTA) created an internal project plan to identify the scope of the program, operational hours, the anticipated operational design domain for the vehicles given the use cases, and the logistics required to include charging and maintenance. UTA contracted with WSP to create the SOP and safety management plan. EasyMile conducted risk assessments for each location to identify hazards and recommended mitigation measures for a given route. The elements considered include the width and character- istics of the pathway, close walls and unclear edges, the speed of traffic, interaction with other road users, and vegetation that may block signals. UTA hosted a tabletop exercise for local stakeholders and emergency personnel to walk through various scenarios. UTA and the vendor prepared a safety plan and protocols, based on their prior experience. These covered the steps to switch from manual to automated mode, emergency stop procedures, oversight, and incident response. A communications plan was also prepared, which included the crisis communications plan to guide communications in the event of a serious incident. Procurement and Implementation Procurement: UDOT product-testing procurement processes allowed for consultation with vendors and open selection of products. Two vendors offered proposals and UDOT executed a contract for LSAV shuttle leasing and operations with EasyMile. Funding: UDOT funded the project through state transportation funds, as part of the connected and automated vehicle program. Federal and State Approvals: The vendor LSAV was imported and does not comply with the FMVSS. The vendor and UDOT will obtain any necessary FMVSS waivers for each deployment. The State of Utah allows automated vehicles to operate on public roads; therefore, no permits from the state were required. Property owners’ consent is required to operate an LSAV shuttle on private roadways in Utah and consent was received. The first of the deployments is on a private roadway, including a mixed-use development and a ski resort, therefore requiring consent. Infrastructure: No physical permanent infrastructure improvements are required, and temporary cones/signage is installed at each deployment site for information and localization. The LSAV shuttle is using LiDAR localization and is not requiring DSRC RSU installation for its deployments. A DSRC on-board unit is available on the vehicle and will be used on at least one site to communicate with a traffic signal. Operations UDOT published an SOP prepared by the firm WSP. The SOP contains a description of organizations involved, roles, training, potential venues, applicable regulations and permits, passenger rules, and operator safety checklists. The State of Utah hosts a website (http://www.avshuttleutah.com) with information about deployments, marks the route with branded signage at stops, and conducts marketing efforts

LSAV Mini Case Studies 111 about the individual pilot sites. The marketing efforts were a joint effort from UDOT’s and UTA’s public involvement team and Horrocks, a public engagement consulting firm. The project leaders at UDOT and UTA have been providing “site ambassadors” at each of the route stops to interact with potential riders and answer questions. In 2019, the Autonomous Shuttle Pilot was operated in Park City in May, Station Park in Farmington in June and July, the 1950 West Business Park in Salt Lake City in July, and the University of Utah in Salt Lake City in August and September. The pilot was operated for events at the Mountain America Expo Center in Sandy from October 2019 to January 2020 and was scheduled for additional venues. These phases were interrupted by the pandemic, and the project was concluded in the fall of 2020. On July 16, 2019, a 76-year-old rider suffered minor injuries after the LSAV came to an emergency stop. The vehicle was removed from service while an internal investigation was conducted by safety personnel at Utah Transit Authority. They found no vehicle error. As a safety precaution, the following actions were taken: the top speed was lowered from 12 mph to 9 mph, safety attendants increased outreach to riders and warned of potential emergency stops, and adhesive material was added to the seats to reduce the potential to slide from the seat. There was no NTSB investigation. Evaluation Key metrics will be ridership, qualitative information about the requirements of operating a low-speed automated shuttle, and public feedback. At the Station Park deployment, the shuttle carried a total of 2,613 passengers over the 4-week period (about 125 per day). The peak day was 205 passengers, which is close to full capacity, and the lowest ridership day had 71 riders. Ambassadors collected 318 surveys on a tablet about their experience. UDOT will collect ridership and vehicle trip data from the vendor. UDOT will work with a team of cognitive psychologists from the University of Utah who will gauge the public’s level of trust in the technology and will assess how trust changes over time. The project team is currently designing evaluation criteria for technology options to include system design, configuration, and alignment based on • Capacity, with objectives for capacity-holding peak demand of 330 passengers (capacity must also include on-board bicycles). • Connections to other transport modes. • Travel time, with an objective of approximately 7 to 15 minutes end to end. • Accessibility, with objectives including ADA compliance and ride comfort. • Expandability and adaptability, including potential new route extensions. • Environmental management factors include impact on protected wetlands near the North Bayshore area.