USDOT Unveils Ambitious Multimodal Automation Initiative, Automated Vehicles 3.0
March 9, 2018
Last week, the U.S. Department of Transportation publicly announced the next phase of its bold initiative to encourage the adoption of automation technologies in transportation.
At the center of this effort is the third iteration of USDOT’s automation policy document, titled Automated Vehicles 3.0. This will build on last year’s Automated Driving Systems 2.0: A Vision for Safety, which provided guidance for automated vehicle (AV) manufacturers and outlined the respective roles of federal and state governments in regulating AVs.
As ETW first reported last year, 3.0 will move USDOT’s role beyond simply overseeing the safe development of AVs to assisting with the actual implementation of automation technologies across all modes of surface transportation – passenger vehicles, trucks, buses, rail, and maritime.
At USDOT’s AV Summit on March 1, Transportation Secretary Elaine Chao touted the benefits of automation highlighting the potential for AVs to reduce the 94 percent of traffic collisions caused by human error, increase efficient use of roadways, and expand access to transportation for the elderly and people with disabilities.
“It’s an exciting time for the future of our nation’s transportation,” she said. “Autonomous technology – including automated cars, trucks, and drones – has the potential to revolutionize the way we travel, transport goods, and connect with one another.”
The deployment of AVs will require not only a significant amount of technological progress in the private sector, but also a concerted government effort to update policies that are inherently incompatible with AVs. Many of the decades-old regulations around motor vehicle safety and operations, including those for trucks and buses, were originally written with human drivers in mind.
But before USDOT goes about updating or eliminating regulations, it must first take inventory of which ones might actually be in conflict with AVs – and the repercussions of modifying existing regulatory structures.
For this reason, Chao has taken a multimodal approach to 3.0 that will coordinate across multiple agencies to identify and address regulatory barriers to AVs, and execute pilot programs to close research gaps. This is a departure from the last two iterations of USDOT’s AV policy documents, which were written primarily by the National Highway Traffic Safety Administration (NHTSA), the nation’s primary regulator of motor vehicles.
During her keynote presentation at the summit, Chao identified 6 department-wide principles that will guide USDOT’s work on AV policy:
- Safety will always be the top priority.
- Policies will be “flexible and tech-neutral, not “top-down, command and control.” USDOT will not pick winners and losers, but will instead allow the market to pick the best solutions and technologies.
- Regulations will be as non-prescriptive and performance-based as possible. In all future regulatory actions, USDOT will not automatically assume that a “driver” of a vehicle is a human.
- USDOT will work with states and other authorities to avoid a patchwork of regulations that could make it difficult for AVs to cross state lines.
- Provide support to stakeholders through guidance, best practices, pilot programs, and other assistance needed to safely integrate AVs.
- Recognize that there will always be a mixed fleet where AVs operate side-by-side with traditional, human-driven vehicles.
“The goal is to develop common sense regulations that do not hamper innovation, while preserving safety,” she said.
But despite their potential benefits, Chao indicated that AVs also present a set of challenges.
“The public has concerns about the security and privacy of automated technology as well as the potential for hacking. Can they operate safely alongside human-operated systems?” she asked.
As a former Secretary of Labor, Chao has also been sensitive to the potential workforce challenges presented by AV technology as it is integrated into trucks and buses. “New technologies create jobs, but the transition period can be very difficult for dislocated workers,” she said. “So, this needs to be addressed to help workers adapt to this new world.”
(Note: Although there is a potential for AVs to trigger worker displacement, some of these fears may be overblown due to the expected gradual implementation of automation in trucking, the current shortage of truckers, and other factors. See ETW’s analysis of the issue here.)
Chao indicated that 3.0 could be released as early as summer 2018 and “will be revised as often as needed.” She also noted that its predecessor, 2.0, was one of the most-viewed USDOT policy documents ever – it has been downloaded more than 125,000 times since it was released in September 2017.
In a display of USDOT’s dedication to incorporating each of the agencies involved in surface transportation, the final panel of the summit managed to cram the leaders of seven modal administrations and USDOT’s Chief Information Officer on stage to discuss their work.
The following is an overview of those engaged agencies and their respective contributions to 3.0.
National Highway and Traffic Safety Administration (NHTSA)
NHTSA is responsible for setting and enforcing nationwide standards around the design, construction, and performance of motor vehicles, known as Federal Motor Vehicle Safety Standards (FMVSS). These universal standards provide automakers with the certainty that they can sell FMVSS-compliant vehicles in any state across the U.S.
In the interest of avoiding a patchwork of conflicting motor vehicle regulations, states and localities have historically been preempted from setting their own motor vehicle standards. The reason for this is that, in the worst-case scenario, automakers would have to design at least 51 different vehicles to sell in each state and Washington, D.C. if they all set different standards.
It all boils down to the fact that motor vehicle standards are an interstate commerce issue – and therefore an inherently federal issue.
However, NHTSA has not yet set federal standards around AVs. Historically, FMVSS has been a prescriptive set of standards around the physical components of vehicles and how their controls respond to human drivers. This framework has worked well for human-driven vehicles, but is rather inadequate (at least in its current form) when it comes to regulating the safety of AVs, which are driven by a complex mix of sensors, computers, software, and physical components.
Since AV technology is still developing and manufacturers are taking a multitude of different approaches to designing them, NHTSA has held off on setting standards until the technology matures.
But this does not mean that there is a regulatory vacuum, as some critics have suggested. NHTSA is actively overseeing the safe development of AVs through voluntary interactions with manufacturers and its continued authority to investigate potential or conduct recalls.
In 2.0, NHTSA provided voluntary guidance for AV manufacturers who opt to conduct their own 12-point Voluntary Safety Self-Assessments (VSSAs), which can then be published for NHTSA and the general public to review. (Read more about NHTSA’s voluntary approach to AVs here.)
NHTSA also has broad authority to conduct investigations of potential safety defects in all motor vehicles in the United States, regardless of whether there are regulations in place or not. This was the case for the fatal Tesla collision in May 2016, when the NHTSA Office of Defects Investigation (ODI) looked into whether the crash was caused by a design defect relating to Tesla’s semi-autonomous Autopilot feature (ETW summarized ODI’s full findings here). Although NHTSA did not find a defect in the Autopilot case, it does have the authority to initiate a recall of affected vehicles or take other disciplinary actions when defects are found.
NHTSA published a request for comments and conducted a public meeting to inform its work on 3.0:
Request for Comments on Removing Unnecessary Regulatory Barriers for Automated Vehicles
(Comment period extended through March 20, 2018)
“The National Highway Traffic-Safety Administration (NHTSA) seeks comments to identify any unnecessary regulatory barriers to Automated Safety Technologies, and for the testing and compliance certification of motor vehicles with unconventional automated vehicles designs, particularly those that are not equipped with controls for a human driver. Further, NHTSA seeks comments on the research that would be required to remove such regulatory barriers.”
Public Meeting on Removing Regulatory Barriers for Vehicles with Automated Driving Systems
Held on March 6, 2018
NHTSA held a public meeting on March 6 “as part of the Agency’s effort to seek public comments to identify any regulatory barriers in the existing Federal Motor Vehicle Safety Standards (FMVSS) to the testing, compliance certification, and compliance verification of vehicles with Automated Driving Systems (ADSs) and certain unconventional interior designs.”
What to expect: NHTSA will not revise the VSSAs in 3.0 – in fact, most of 2.0 is expected to be simply rolled into 3.0 as-is. What will be new is a discussion of the unnecessary regulatory barriers pertaining to the design, construction, and performance of AVs.
Federal Highway Administration (FHWA)
As discussed in Eno’s report last year, Beyond Speculation: Automated Vehicles and Public Policy, the nation’s roadways will likely not need to be retrofitted with advanced vehicle communication technologies to operate properly. Instead, manufacturers are designing to operate on the roadways we have today by detecting lane markings, reading signs, and detecting stoplights.
Maintaining this traditional infrastructure is already a challenge of its own. In the face of funding challenges, many cities and states will likely find it difficult to ensure their lane striping is refreshed often or that all of their signs are in good condition.
But while AVs will not necessarily require advanced technologies that enable vehicle-to-vehicle (V2V) or vehicle-to-infrastructure (V2I) communication, both solutions could potentially enhance their safety and efficient use of the roadway.
One application of V2V technology, platooning, could allow trucks to follow each other at shorter following distances – which could increase fuel efficiency and potentially improve the job of driving.
FHWA is already conducting a study of how connected infrastructure could be installed along highways to facilitate vehicle connectivity, called the Connected Vehicle Pilot Deployment Program. Three pilots are currently underway in New York City, Wyoming, and Tampa, Florida.
FHWA is also separately evaluating how roadway designs could be modified for AVs and published the following request for information:
Request for Information on Integration of Automated Driving Systems (ADS) into the Highway Transportation System
Comments closed on March 5, 2018
“To better understand what is needed to accommodate ADS technologies, and maximizing their potential benefits, the Federal Highway Administration (FHWA) seeks the public’s input through a formal RFI to supplement strategy development. Advancing the next generation of America’s transportation network can only happen with input from an array of informed sources, including stakeholders, industry experts and the public at large.”
What to expect: FHWA will likely provide an overview of its ongoing research into the need for uniform traffic control devices for AVs, the role of connectivity in AV operations, insights into the potential for automated trucks to platoon, and potential future research opportunities.
Federal Motor Carrier Safety Administration (FMCSA)
While NHTSA sets standards for the design, construction, and performance of all motor vehicles including trucks and buses, FMCSA sets national regulations around the operation of those vehicles.
Since human drivers have always operated trucks and buses up to this point, FMCSA’s regulations around their operation – called the Federal Motor Carrier Safety Regulations (FMCSR) have always assumed that a human driver is behind the wheel.
FMCSA has already started to research how automating the task of driving could impact truckers and trucking companies, and now plans to gather industry feedback on the regulatory obstacles that exist for automated trucks and buses.
During her speech, Chao indicated that some of the existing FMCSRs “may need to be updated, modified, or eliminated to facilitate the safe introduction of automated technology on commercial motor vehicles.” This could include the current restrictions on driver hours of service and other regulations pertaining to the presence of a driver.
Chao said that FMCSA would request public comments on this issue “in the near future.”
What to expect: FMCSA will likely discuss existing regulatory barriers to the operation of automated trucks and buses in 3.0 and describe how it may remove those barriers.
Pipeline and Hazardous Materials Safety Administration (PHMSA)
One million shipments of hazardous materials occur daily on every mode of transportation. Because of the intermodal nature of hazmat transportation, PHMSA Administrator Howard “Skip” Elliott told summit attendees that it was important for his agency to work with the other modes to proactively research and – if needed – prepare its regulations for the advent of AVs.
“Over the last 20 years, we’ve seen a huge increase in safely transporting these materials. In many modes today, 99.99 percent of hazmat is transported safely from origin to destination,” Elliott said. “That last bit of safety, that sliver of a percent, will come from innovation and automation.”
Elliott promised that, As PHMSA considers the role of AVs in transporting hazardous materials, it “will always put the safety of people and the environment first.”
PHMSA will soon release a request for information on how it can reduce regulatory roadblocks to using AV technology to transport hazardous materials safely and efficiently – including processes for electronically reporting incident information to hazmat responders before they even arrive on the scene.
What to expect: PHMSA is anticipated to provide a broad overview of the comments it received on the RFI and regulations it will review moving forward.
Federal Railroad Administration (FRA)
Compared to the on-road modes of transportation, the rail industry faces a much different set of regulatory, funding, labor, and technological challenges in implementing emerging technologies.
Perhaps the most illustrative example of these challenges is the sluggish rollout of a kind of autonomous braking technology, positive train control (PTC) – which, as ETW has written, could have prevented two recent train accidents that had multiple fatalities. (But, as Jeff Davis also wrote, PTC implementation could be worse.)
PTC aside, FRA Chief Counsel Juan Reyes said that the agency is continuing to explore how railroads can and should be automated. He said that FRA would soon release an RFI to assess the prospective introduction of new automation technology on railroads.
Similar to the requests from FHWA, NHTSA, and FTA, this will seek information related to safety, security, infrastructure, workforce, as well as legal and regulatory matters.
Maritime Administration (MARAD)
If there were an award for calling out the elephant in the room at the summit, it would have went to Richard Balzano, Deputy Administrator of MARAD.
“I know what you’re all thinking, who invited the sailors?” he joked.
But, as Balzano went on to explain, the inclusion of MARAD actually fits neatly within Chao’s overall objective of breaking down the communications barriers and knowledge gaps between each modal administration.
“The maritime industry is connected to all the modes – train, rail, and trucking are very much a part of unloading and unloading our ships. And our industry is transforming; automation is coming to our industry with automated vessels and port services and self-driving trucks,” he said.
Balzano said that the Europeans and other regions are “quite a bit ahead of us” because they have made greater investments in automated vessels and advanced port technologies. Rolls-Royce has developed completely remotely operated harbor tugs and is currently testing them, Balzano said, alluding to the company’s operations at its Research & Development Centre for Autonomous Ships in Finland.
But Balzano insisted that the U.S. will not be left behind. He noted that the US Navy has developed a 132-foot surface vessel that is completely unmanned and operating on its own. And MARAD is currently working with the Office of Naval Research on a project called Overlord – which will completely automate three commercial cargo vessels, put them off the coast, and operate them for 90 days at a time continuously.
“It’s not a matter of whether if it is coming, it is when and how. That affects our workforce; that affects our regulations,” he said. “So the connectivity and the standardization has to be similar so that our ships can operate in those different places and those different categories.”
MARAD is also standing up an Autonomous Ship and Port Operation Working Group and signed up with the American Bureau of Shipping to participate in the Autonomous Symposium. MARAD is also conducting studies on port staging and other port efficiencies and, for the first time ever, has developed a research and technology strategy for MARAD (but the funding for it has yet to be secured.)
“This is going to be a matter of the right safety regulations and the right policies at the right time to get this done,” Balzano said.
Federal Transit Administration (FTA)
FTA’s planned contribution to 3.0 is perhaps the most ambitious. The following is ETW’s overview of its automated bus initiative, which previously appeared here.
On December 5, 2017, FTA unveiled its 5-year agenda for researching the potential benefits of vehicle automation for public transit. Dubbed the Strategic Transit Automation Research (STAR) Plan, it will serve as a guide for the agency and transit stakeholders through FY 2022. The STAR plan will also inform much of FTA’s contribution to 3.0.
FTA has developed an ambitious roadmap for STAR that entails at least one – and sometimes three – demonstration projects per fiscal year through FY 2022, plus a slew of research initiatives to understand the effectiveness of automated buses, consumer acceptance, and potential workforce impacts, and many other topics.
(Source: FTA STAR Plan Roadmap)
For well over a year, FTA has been communicating with transit agencies and the private sector to learn about how public transit services could benefit from automation and, perhaps more importantly, how far it is from maturity.
As an FTA official noted in the announcement webinar for the STAR Plan, a handful of transit agencies are already experimenting with automated transit solutions. Yet the vast majority of these pilot programs are using small driverless shuttles for first/last-mile connections, rather than automating the functions of traditional buses.
(Source: FTA STAR Webinar – December 5, 2017)
This is partially due to the lack of market-ready technologies to automate buses, but also the inherent financial constraints in the public transit sector. While the average 12-year lifecycle of a bus is close to that of passenger vehicles, agencies face challenges in procurement, liability, and labor relations (among many others) that tend to make them more risk averse than car buyers.
As part of its early research for STAR, the FTA conducted benefit-cost analyses to quantify the opportunities and obstacles to transit automation. Notably, the benefit-cost analyses were limited to the financial costs, which excluded the social impacts (e.g., worker displacement) that are difficult to predict and quantify in advance.
FTA hopes to change this dynamic by supporting pilot projects that implement automated transit services and providing guidance for bus manufacturers and stakeholders. Through FY 2022, FTA will publish notice of funding opportunities (NOFOs) for seven demonstration project grants that will explore a variety of use cases:
- Integrated Demonstration 1 (FY 2018-2019): Implementing advanced driver assistance systems (ADAS) in transit buses (SAE Automation Level 1-2);
- Integrated Demonstration 2 (FY 2019-2020): Low-speed shuttle buses automation (e.g., circulator services, first/last-mile access to transit networks) (Level 4);
- Integrated Demonstration 3 (FY 2020-2021): Automation for maintenance and yard operations such as precision movement for fueling, maintenance, bus wash, automated remote parking and recall (Level 4);
- Integrated Demonstration 4a (FY 2021-2022): Automated ADA paratransit (Level 5);
- Integrated Demonstration 4b (FY 2021-2022): Automated first/last mile service (Level 5);
- Integrated Demonstration 4c (FY 2021-2022): On-demand shared ride service (Level 5);
- Integrated Demonstration 7 (FY 2021-2022): Automated bus rapid transit (BRT) service (Level 4).
Before taking automated buses to the streets and exposing them to the public, an FTA official noted that the agency must first test their safety and capabilities. As a result, FTA plans to construct an indoor/outdoor transit automation testing facility to support its integrated demonstration projects.
The agency is currently developing a set of requirements for the facility, which will be used to study the seven use cases for transit automation and other research areas outlined in STAR.
Roadblocks to Automation
Since FTA does not yet have a comprehensive set of regulations or guidance in place for automating transit services, it has already started researching potential policy barriers and incompatible regulations. In the process, FTA is examining how it can update its policies and, if necessary, create entirely new regulatory frameworks.
This research will culminate in FTA sending a set of recommendations to Congress for policy changes and updates to facilitate transit automation. FTA plans to send these recommendations in FY 2019 so that they can be considered in the next surface transportation bill when the FAST Act expires in 2020.
Introducing automated buses into transit operations can present a number of potential risks, which FTA has split into four categories: safety and security, operations and cost effectiveness, passenger experience, and equity.
The official highlighted equity risks in particular, describing a scenario where automated buses may not have on-board attendants and could be designed to only accept electronic payments. In this scenario, people without debit/credit cards – especially those who are already living in communities underserved by transit – would not be able to use the service since they would not be able to pay in cash.
(Source: FTA STAR Webinar – December 5, 2017)
Transit agencies must also prepare to confront a number of barriers to implementing automation technologies, including product availability, labor relations, financial constraints, and some transit agencies’ risk aversion.
(Source: FTA STAR Webinar – December 5, 2017)
Workforce resistance stands out in particular, as labor unions could mount opposition to agencies procuring automated buses if there are concerns of worker displacement. However, an FTA official suggested that the potential impact on bus drivers is “pretty unclear as of yet” and that even a self-driving bus may need an operator on board for fare collection, rider safety, and customer service.
On the other end, transit agencies will also need highly trained workforces capable of managing and maintaining automated bus fleets – workforces that do not exist yet.
As FTA rolls out this initiative in the coming years, it will research these and other issues and develop guidance for agencies as they navigate the new world of automated transit services.
“We really want to be a clearinghouse, a one-stop shop,” an FTA official said.
The roadmap for the STAR Plan and ongoing updates from FTA can be accessed here.
FTA Requests for Comments for 3.0
Request for Comments on Automated Transit Buses Research Program
Comments closed on March 2, 2018
This RFC will obtain transit industry comment regarding the current and near-future status of automated transit buses and related technologies. Comments received will assist FTA in developing future Notices of Funding Opportunity’s (NOFO) for transit bus automation demonstrations with respect to the industry’s ability and interest. Concepts for the NOFOs on demonstrations and evaluations of transit bus automation usage will be defined in FTA’s Strategic Transit Automation Research (STAR) plan. A premise of the STAR plan is to leverage and apply commercially available technology and products to the public transit industry to hasten early demonstrable outcomes.
Request for Comments on Removing Barriers to Transit Bus Automation
Comments closed on March 2, 2018
This RFC will obtain transit industry comment regarding current or potential institutional, regulatory, or other policy barriers to the development, demonstration, deployment, and evaluation of automated transit buses and related technologies. Some Federal policy issues are addressed in the U.S. Department of Transportation (DOT) Automated Driving Systems (ADS): A Vision for Safety 2.0. Information from the second RFC will help inform FTA’s approach to the planned update of this document, which may include additional multi-modal considerations. Information from the RFC will also provide critical feedback on potential modifications of FTA rules, regulations, and agency guidance.
What to expect: FTA will describe the progress of its STAR Plan, the regulatory barriers it has identified for automating bus operations, and how it plans to support stakeholders in the integration of automated buses into the future.
It must be noted that 3.0 focuses on surface transportation modes and will not include automation technology in aviation, which presents an entirely different set of challenges.
The Federal Aviation Administration (FAA) has consistently struggled to implement the next generation of its air traffic control system, and has sluggishly moved to integrate unmanned aircraft systems (UAS, or drones) into the airspace. Yet Michael Kratsios, Deputy Assistant to the President and Deputy U.S. Chief Technology Officer at the White House, told the summit attendees that the Administration is actively working on aviation issues separately.
“Automation is not limited to our roads,” he said, noting that the FAA’s drone registry program has already registered 1 million drones and that FAA plans to announce the selected applicants for its Drone Integration Pilot Program “in only a short few months.”
USDOT is accepting public comments on the Automated Vehicles 3.0 Summit in Docket No. DOT-OST-2018-0017. A video of the public listening summit can be accessed here.