Public Comments on NHTSA’s 2017 Automated Driving Systems: A Vision for Safety 2.0
On September 12, 2017, NHTSA published new guidance for autonomous vehicles titled Automated Driving Systems: A Vision for Safety 2.0 (“Voluntary Guidance”). NHTSA asked for public comments on the new guidance to be submitted by November 14, 2017. Over 2900 public comments were posted to the docket by the deadline. More than 2800 of those comments were submitted by individuals concerned about issues of safety for motorcycles, in a campaign coordinated by motorcycle advocacy groups. Other comments were submitted by industry participants, state and municipal governments and transit agencies, policy groups, advocacy organizations and concerned citizens. The comments continue to be submitted even after the deadline, and can be viewed at www.regulations.gov on docket NHTSA-2017–0082.
This article highlights some key points made in the public submissions, but in light of the number of submissions, does not purport to summarize or include all submissions, or all comments or points made in each of the submissions. Many commenters commended NHTSA on its efforts to promote autonomous vehicles (AVs) and to accelerate their development, and noted the potential safety benefits of AVs or raised general concerns about AVs. This article does not repeat these general points but instead focuses on the specific suggestions and comments offered to the guidance. The article groups the comments into three categories: (1) Scope of Guidance; (2) Voluntary Safety Self-Assessments; and (3) Guidance to States.
Scope of Guidance
Some of the comments address the general scope and approach of NHTSA’s guidance, including the types of vehicles and stages covered, the voluntary nature of the guidance, and the standards and format of the safety assessments.
Levels 3–5, Not Level 2
Association of Global Automakers (Global Automakers): The Association of Global Automakers, Inc. is an industry group including American Honda Motor Co., Aston Martin Lagonda of North America, Ferrari North America, Inc., Hyundai North America, Isuzu Motors America, Inc., Kia Motors America, Inc., Maserati North America, Inc., McLaren Automotive Ltd., Nissan North America, Inc., Subaru of America, Inc., Suzuki Motor of America, Inc., and Toyota Motor North America, Inc. Global Automakers “agrees that the Guidance should only apply to automated vehicles classified by the manufacturers as SAE Level 3 or above, as these systems are generally intended to operate without continuous supervision by a human driver or operator. We have concerns, however, with the discrepancies in how each level of automation is described in the Guidance (on Page 4), and we recommend that NHTSA consider more closely aligning its descriptions with the definitions established in SAE J3016.”
Alliance of Automobile Manufacturers (Auto Alliance): The Alliance of Automobile Manufacturers is an industry group including BMW Group, FCA US LLC, Ford Motor Company, General Motors Company, Jaguar Land Rover, Mazda, Mercedes-Benz USA, Mitsubishi Motors, Porsche, Toyota, Volkswagen Group of America and Volvo Cars North America. The Alliance “notes that the scope of the updated Guidance has been refined to focus on SAE Levels 3- 5 ADS-equipped vehicles instead of including Level 2 driver support system-equipped vehicles that are currently on the road. This is an important clarification and addresses the potential retroactive approach previously taken in the 2016 Federal Automated Vehicles Policy.”
Volkswagen Group of America (VWGoA): “VWGoA agrees with the exclusion of SAE Level 2 (L2) driving automation systems from the automated vehicles guidelines. This removes a huge burden from the industry and makes the application of the Voluntary Safety Self-Assessment (VSSA) much clearer to those who have already had L2 systems available in the market prior to the initial publication of NHTSA’s automated vehicle guidelines.”
General Motors (GM): GM “appreciates that the Guidance applies only to HAVs with systems capable of operating at SAE Levels 3–5.”
Ford Motor Company (Ford): Ford appreciates NHTSA’s “clarification that the safety self-assessment (VSSA) is voluntary and applies to SAE Level 3-and-above-capable vehicles.”
US Chamber of Commerce/CTEC (CTEC): CTEC believes that SAE automation Levels 3–5 “is the proper scope for the Voluntary Guidance for a couple of reasons. First, CTEC believed that some of the safety elements outlined in the FAVP were not applicable to Level 2 advanced driver assistance systems (“ADAS ”). Second, ADAS are currently in the marketplace and have already proven to be effective in helping consumers avoid accidents, and are currently regulated by NHTSA through its existing investigation and defect authority.”
American Trucking Associations (ATA): The ATA “supports NHTSA’s decision to focus the Voluntary Guidance on SAE Automation levels 3 through 5, rather than 2 through 5 as in the original FAVP” and notes that “SAE level 2 requires the driver to remain engaged with the driving task and monitor the environment at all times, in contrast to levels 3 through 5 in which the automated driving system monitors the driving environment and performs the driving task.” Additionally, ATA “agrees with NHTSA that it is the responsibility of the entity developing automated driving system (ADS) technology to determine the system’s level of automation in conformity with SAE’s published definitions.”
Maryland Department of Transportation (MDOT): MDOT suggests “more focus to address the lower levels of automation (Level 1–2), including some of the advanced driver assistance systems already being deployed by automakers today.”
National Safety Council (NSC): NSC “calls on NHTSA to address lower-level automation systems, including Level 2, in the policy.” NSC supports “all active safety technologies being held to the same rigor of documentation as proposed in the policy.”
Consumer’s Union: Consumer’s Union notes that “given the demonstrated safety hazards associated with foreseeable use and misuse of Level 2 driving automation systems — which contain combined automated functions, like acceleration and steering, but require the driver to remain engaged and monitor the environment at all times — we do not support this change.” Consumer’s Union also suggested there is confusion among consumers about the different automation levels: “It appears that some consumers are foreseeably operating some Level 2 vehicles as if they were actually Level 3, reinforcing the need for consumer protection for these vehicles.”
National Transportation Safety Board (NTSB): NTSB “believes that NHTSA missed an opportunity by not including Level 2 systems in the guidance document, especially given that the automated vehicles currently commercially available meet the Level 2 definition.” NTSB notes that “Level 2 automated vehicle systems have narrower operational design domains and more limited event detection/response capabilities than vehicles operating at higher levels of automation, but they have fewer restrictions on their operations.” The NTSB “recognizes that Level 2 systems are already in the market” but suggests NHTSA policy could “establish a future model year date by which manufacturers would be required to offer Level 2 automated control systems that have defined and documented operational design domains and that have undergone safety assessments” which would “give both NHTSA and manufacturers an opportunity to normalize industry practices.” The NTSB states that “Level 2–5 vehicles rely on advanced vehicle systems with proven benefits, such as collision avoidance technologies. Increasing the implementation of collision avoidance technologies, an issue area on the NTSB’s Most Wanted List of safety improvements, could be advanced through these policies.”
Insurance Institute of Highway Safety (IIHS): IIHS “strongly urges NHTSA to request similar information from companies deploying vehicles with partial driving automation technology (SAE level 2)” noting it is “concerned that the distinction between SAE level 2 and level 3 systems may not be apparent from the user’s point of view” and that “existing safeguards to keep the human driver appropriately engaged in the driving task are unproven and inadequate.”
Testing v. Deployment
Ford Motor Company: Ford believes that “during the early stages of development, all 12 guidance areas of the Vehicle Safety Self-Assessment (VSSA) may not be applicable to test vehicles.” Ford notes “this is in part due to the fact that a trained test driver will likely supervise the Automated Driving System (ADS) in the early stages of testing” and therefore “the ADS will function much like an L2 system with the test driver engaging the ADS within the Operational Design Domain (ODD) and being responsible for the correct Object and Event Detection and Response (OEDR) and/or the fallback within the ODD.” Ford states: “It is our hope that NHTSA takes this into account. Moreover, we continue to encourage consistency with SAE J3016 (e.g., terms like system safety, OEDR, and fallback, among others), as well as other industry standards for AVs as they become mature.”
Volkswagen Group (VWGoA): Regarding the application of the guidance and the VSSA to test vehicles, VWGoA notes “this is problematic as an ADS will, as part of the normal process, experience many different iterations, changes, and modifications throughout development” and therefore “continuously providing updates to a test vehicle’s VSSA will unnecessarily impede the development process and create huge burdens for the industry where little benefit will be gained by the Agency or the public.” VWGoA also notes the difficulty in providing VSSAs for ADS testing “without revealing competitive and confidential business information” because in the early stages of development, “describing things like the intended use or ODD of an ADS in testing will disclose sensitive information that companies typically do not share with other manufacturers or the public.” VWGoA also believes “providing VSSAs for test vehicles could cause confusion to consumers for two reasons. First, it would inundate the public with continuously changing VSSAs that do not directly apply to a future production vehicle. Second, it would not be possible for a test vehicle’s VSSA to cover all 12 sections of the guidance which could give an incorrect impression of incompleteness to the public. As such, VWGoA recommends that NHTSA amend the guidance to specify that the VSSA only need to be applied to deployment vehicles.”
Auto Alliance: Auto Alliance members “remain concerned regarding the applicability of the VSSA to test vehicles.” Auto Alliance notes that the Guidance “recognizes that not all of the safety elements in the VSSA will be applicable to test vehicles” and “would like to emphasize that providing a VSSA for each variant of an automated test vehicle will quickly become unwieldy.” Auto Alliance states: “Not only do some of the safety areas clearly not apply for automated test vehicles — for instance, Consumer Education and Training — but providing an update to the VSSA for each modification to prototype technology would greatly encumber the development process and would likely delay deployment without providing additional benefit. We ask that NHTSA revise the applicability of the VSSA to focus on deployed ADS-equipped vehicles and ADSs only.”
Robert Bosch LLC (Bosch): Bosch remains “concerned that the document does not differentiate between testing and deployment of Automated Vehicles (AV) and Automated Driving Systems (ADS ).” Bosch notes that “test vehicles are, by their very nature, prototype or evaluation vehicles” and thus “various aspects of the vehicles and the ADS are constantly being changed and adjusted toward the overall goal of enhancing their performance” and “several of the individual safety design elements identified by NHTSA for inclusion in the VSSA will not be determined or finalized in a test vehicle.”
Automotive Safety Council: ASC “would like to reiterate the need to potentially differentiate between prototype vehicles in the test/development phase and commercially deployed/production vehicles” and notes the “HMI, Components and Control software and the ODD in these test vehicles may not be representative of the mature production state.” ASC believes that “additional clarity is needed relative to the differentiation between test /development vehicles being tested with employees and contractors versus testing with public. Clearly production vehicles need to meet higher safety requirements, but it is also critical for test vehicles which are carrying members of the public to meet higher safety requirements.”
Voluntary v. Mandatory Safety Assessment Process
Auto Alliance: Auto Alliance notes that “members appreciate that the Voluntary Safety Self-Assessment (VSSA) is a voluntary publication process” and that the process “provides both transparency to the public about critical safety elements as well as flexibility for each automaker or ADS supplier to customize its assessment and publish it in the form that makes the most sense.”
Global Automakers: Global Automakers notes that “although the Voluntary Safety Self-Assessment set forth in the Guidance is voluntary — as was the Safety Assessment Letter envisioned in the earlier FAVP — manufacturers and other entities are nonetheless encouraged and incentivized to provide documentation on how they are addressing safety for the testing and deployment of ADS on public roads.”
Ford Motor Company: Ford appreciates NHTSA’s “clarification that the safety self-assessment (VSSA) is voluntary and applies to SAE Level 3-and-above-capable vehicles.”
Niskanen Center: The Niskanen Center notes that “the repeated emphasis on ‘voluntary guidance’ and ‘best practices’ is a welcome shift in NHTSA’s approach to governing autonomous vehicle research, testing, and deployment.” Niskanen suggests “the Agency should continue on this steady course and avoid the hue and cry to ‘do something more’ that may be outside the purview of its statutory mandate, such as attempting to craft privacy rules. There will undoubtedly be such calls moving forward. NHTSA should resist those demands, and continue on its current course by extolling the benefits of voluntary, industry-led best practices and standards.”
National Safety Council: NSC notes that “should NHTSA find that automakers and technology developers consistently do not disclose their safety design elements in a manner that allows public scrutiny, NHTSA should consider making these voluntary assessments mandatory.”
Oregon Department of Transportation (ODOT): ODOT expresses concern “about the voluntary nature of the existing guidance,” which “relies on manufacturers voluntarily providing information about their technologies as they engage in testing.” ODOT notes that “while the policy guidance provides some recommendations in terms of best practices or expectations,” they are “described more as general principles rather than explicit requirements for what constitutes safety in each area.” ODOT believes that “the establishment of a process for safety determination is especially important as vehicles move from testing into commercial deployment. Currently, the non-binding policy guidance is applicable to both testing and deployment on public roads. While testing with limited numbers of vehicles, trained test operators, and controlled conditions mitigates some of the risk associated with AV operation, the large-scale commercial deployment of AVs will necessitate a more rigorous framework than voluntary guidance to ensure that AVs meet reasonable safety standards.”
New York City Mayor’s Office: The office of the NYC Mayor states that “as legislators in both the House of Representatives and the Senate have recognized, NYC feels it is critical that the Safety Assessments be mandatory and required at regular intervals.” The NYC Mayor believes that while “NHTSA has historically required automakers’ self-certification of their vehicles’ safety and compliance with FMVSS only after they were made commercially available,” the complexity of the “HAV development process” and the “potential for harm” represent a “major divergence from the established regulatory model, and necessitate a more active oversight role for the nation’s vehicle safety regulators.” The NYC Mayor notes that “while mandatory Safety Assessments would not equate to pre-market certification, they would help NHTSA anticipate risks and act more quickly and decisively should companies deploy their vehicles prematurely.”
Consumer’s Union: Consumer’s Union urges NHTSA “to eschew its voluntary approach and instead require companies to submit safety self-assessments and make them public.” Consumer’s Union notes that while “Congress may separately choose to require the submission of safety assessments,” they “nonetheless urge NHTSA, at the earliest practicable date, to start a rulemaking that would require submission of the safety self- assessment and its public disclosure by manufacturers.” In the meantime, Consumer’s Union believes that safety assessments “should not be considered voluntary for companies as they seek to build consumer trust” and that “manufacturers and other entities should submit and make public the assessments, and go beyond what is listed in the guidance to include meaningful evaluation of all factors that could affect the safe functioning of an ADS, including those addressed in the remainder of these comments.”
Insurance Institute for Highway Safety: IIHS believes that “companies should be required to submit descriptive information about each automated driving system, its level of automation, and its operational design domain.” IIHS notes that “the safety self-assessment could help the agency and other stakeholders evaluate the safety and performance of automated driving systems, but will provide unreliable and incomplete information if companies are asked to volunteer information about their product and assert its safety without supporting evidence.” IIHS believes the safety assessment submitted by Waymo “provides a good introduction to self-driving vehicle technology and Waymo’s approach in particular, but offers no evidence that testing on public roads, simulated driving, and various hardware and software safety assessments performed during the development and deployment of the automated driving system has yielded a system as safe or safer than a human driver within its operational design domain.”
Center for Auto Safety (CAS): CAS states that “it would be in the best interest of all stakeholders to make sure that NHTSA, researchers, and the public have access to all the necessary data to assure the vehicles are performing as promised — and when there are problems — providing enough information for everyone to understand what happened.” CAS recommends “making the type of information that is listed in the ‘Voluntary Safety Self-Assessment Template’ on crashworthiness mandatory — and making the same true of the other 11 priority safety design elements. Currently, ADS 2.0 states that Safety Assessment letters are neither required nor is there any mechanism to compel entities to submit them — this must change.”
Standards/Format for Safety Assessments
General Motors: GM commends “the format of the Voluntary Safety Self-Assessment, which asks manufacturers to acknowledge consideration of each safety element as opposed to prescribing a ‘meet/does not meet’ designation,” which reflects NHTSA’s “responsiveness to stakeholder suggestions.” GM believes “this format facilitates development of industry best practices that both protect safety and adapt to innovation.” GM appreciates “that the Guidance shows firm commitment to vehicle safety, but remains flexible in welcoming many methods of enabling safe testing and deployment of HAVs.”
Volkswagen Group: VWGoA supports “the decision to change the way in which a submitting entity is encouraged to acknowledge their consideration of the guidance during the development process.” VWGoA finds that “this type of acknowledgment is more in line with the intended purpose of the guidance which is to ‘help designers of ADSs analyze, identify, and resolve safety considerations prior to deployment…’ and not create a standard to which manufacturers must self-certify.” VWGoA notes this is “further supported by NHTSA’s reiteration that the guidelines are purely voluntary and do not serve as a ‘compliance requirement or enforcement mechanism.’” In order to “further expand on what NHTSA already provided,” VWGoA also “asks that NHTSA provide a complete template as it will help to simplify and standardize the document for integration into existing documentation processes and make the VSSA much easier for the public to understand.” VWGoA also suggests that NHTSA “consider assuming the responsibility of hosting a website that would serve as a repository for all VSSAs in order to have a single source where anyone can easily access these documents for deployed ADS-equipped vehicles.”
Uber: Uber endorses “NHTSA’s general approach” and offers comment “in order to highlight certain issues that, in our view, will promote the achievement of these shared goals.” Uber notes that “in particular, we strongly support the flexibility embedded in NHTSA’s safety design elements to apply innovative methods to responsibly mitigate safety risks, and believe there is opportunity to further emphasize this approach in certain targeted areas.”
Zoox: Zoox supports NHTSA’s approach and notes that “given the early stage of the industry, we are pleased to see that this policy takes the appropriate posture for allowing the technology to evolve.” Zoox believes “future standards in this domain will be informed by robust data collection. The data will inform best practices and the best practices will inform standards and future policies.” Zoox states that “the Voluntary Safety Self-Assessment (SSA) will also help inform the public about the technology. Responsible developers will take the letter seriously and will convey their safety principles in a way that builds trust in the technology. Additionally, it is likely that third parties such as consumer groups, professional services firms, think tanks, and media outlets will create tools to evaluate and compare SSA’s — creating an environment that generates transparency and candor.”
Global Automakers: Global Automakers notes that “while providing a template may be useful for some companies in helping to determine the level of information to provide, the Guidance should continue to encourage freedom for manufacturers to disclose relevant information in a format that works best for them.” Global Automakers states that “this is an emerging area, and how manufacturers or other entities may communicate relevant safety information to the public is likely to evolve as we gather more experience and a greater understanding of consumer expectations.” Global Automakers notes the concern that “if NHTSA were to issue further templates (or expand upon the existing template), any inconsistencies between the Guidance and the template(s) could create uncertainty with respect to how the guidance should be considered.”
American Trucking Associations (ATA): ATA “does not believe that there should be a standard format for the Voluntary Safety Self-Assessment at this time” in light of “the differences in approach to the design of automated driving systems in general, as well as differences between commercial and passenger vehicles.”
Connecticut Department of Transportation (CDOT): CDOT “recommends that NHTSA begin to work with industry to revisit the Federal Motor Vehicle Safety Standards (FMVSS). For deployment, minimum standards will be needed, like that provided for mechanical safety under FMVSS.”
Oregon Department of Transportation (ODOT): ODOT notes that “manufacturers are not required to have addressed each area in the guidance, and can simply indicate in their Safety Assessment Letter that they do not meet the recommendations of the guidance area.” ODOT “understands that the policy guidance intentionally avoids being too prescriptive in order to encourage innovation in this rapidly developing sector,” but notes that “this framework presents some difficulty for states that usually look to the federal government for the establishment of motor vehicle safety standards.” ODOT states that “manufacturers are recommended to submit documentation that their vehicles follow NHTSA’s performance guidance, but without clear determination from NHTSA as to whether a vehicle system adequately or inadequately addresses each of the safety areas, states are left to make their own determination as to the safety of the vehicle.” ODOT asks that “NHTSA establish minimum safety criteria for the 15 areas, and a certification process for determining that vehicles meet those criteria, potentially involving certification by third-party organizations with the expertise to make such determinations.” ODOT is concerned that “authorization to test a vehicle technology which has not been determined to be safe by a qualified party could inaccurately imply a determination of safety by ODOT, or put the agency in the position of being liable if damages occur in association with an authorized vehicle test.”
Travelers: Travelers is a U.S. insurance provider. Travelers notes that “a key role for NHTSA, recognized in the Guidance and in federal legislation currently being considered by Congress, is the development of new safety standards for ADS-equipped vehicles.” Travelers supports “prompt development of appropriate standards that will support a level of safety for such vehicles equivalent to that achieved through existing standards for conventional vehicles.” Travelers “does not oppose the issuance of exemptions from today’s safety standards as needed to facilitate testing of vehicles until new ADS standards are in place,” but believes “such exemptions should be conditioned on the party seeking the exemption demonstrating a level of safety of the ADS equivalent to existing safety standards.” Travelers also suggests that “any such exemptions or other authorizations to test and deploy ADSs must also be conditioned on the ADS operator collecting and reporting data on the performance of the ADS, including data on crashes or other incidents” and that “such data should be made available to insurers for all of the reasons noted in these comments, including allowing appropriate insurance/risk decisions to be made based on the vehicle’s safety performance.”
American Association of State Highway and
Transportation Officials (AASHTO): AASHTO notes that “the ADS 2.0 policy guidance might or might not be good for the preliminary testing phase of vehicles equipped with ADS, but it is wholly inadequate for the deployment phase of automated vehicles because minimum standards will be needed, like that provided for mechanical safety under FMVSS.” AASHTO agrees “the regulation of the design, construction and performance (in the traditional, mechanical manner as defined in Title 49 Section 30102) of a motor vehicle is a federal obligation.” However, AASHTO believes the “application of the term ‘performance’ to the autonomous context is now more complex due to the increasing merger of the vehicle and the operator” and that “reexamining the definition of ‘performance’ in the context of ADS” will enable “focus on the operational safety laws regulating motor vehicles and their operators after such vehicles have been constructed and introduced to public roadways.”
American Association of Motor Vehicle Administrators (AAMVA): AAMVA notes that “while AAMVA understands the need for discretion in addressing the submission of what can be considered proprietary data, AAMVA also recommends that NHTSA consider inclusion of supporting data or information that extends beyond the scope of whether or not an identified safety element was considered during product development or is considered ‘not applicable.’ This generalization in submitted information could dilute the substantive accountability chain between what entities report versus the documented capabilities of an ADS — which could ultimately impair public acceptance of the technologies as a whole.”
Electronic Privacy Information Center (EPIC): EPIC believes that “Automated Driving Systems: A Vision for Safety does not further NHTSA’s mission of protecting drivers.” EPIC notes that while “new vehicle technologies offer a variety of beneficial services to American drivers, and are being quickly implemented by car manufacturers,” these new technologies “also raise substantial privacy and safety concerns that must be addressed through meaningful, legally enforceable safeguards.” EPIC believes that “current approaches, based on industry self-regulation, are inadequate and fail to protect driver privacy and safety” and that “NHTSA must issue mandatory rules to address the myriad risks posed to drivers operating vehicles in the United States.”
Commercial Trucking
Uber: Uber points out that “although interstate motor carrier operations are beyond the scope of this guidance,” the Guidance “categorically suggests that the Federal Motor Carrier Safety Regulations (FMCSRs) require ‘a trained commercial driver [to] be behind the wheel at all times, regardless of any automated driving technologies’ (Guidance at 2).” Uber notes that “while the FMCSRs were not written to anticipate autonomous vehicles, we do not believe the FMCSRs must be construed to require a human behind the wheel at all times.” Uber encourages NHTSA “to work with FMCSA to revisit this point in future guidance, in order to facilitate development of self-driving technology for all vehicle types.”
American Trucking Associations (ATA): ATA notes that “the new NHTSA guidance states in its Scope and Purpose section: ‘Currently, per the Federal Motor Carrier Safety Regulations (FMCSRs), a trained commercial driver must be behind the wheel at all times, regardless of any automated driving technologies available on the CMV, unless a petition for a waiver or exemption has been granted.’ (2017 Guidance at p. 2).” ATA states that “this reading of the FMCSR indicates that a human operator is required in a commercial vehicle, even a vehicle designed to be operated without a human operator, at SAE Level 4 or 5.” ATA believes this “stands in stark contrast with a response NHTSA provided separately to Google in February 2016, which stated that, ‘If no human occupant of the vehicle can actually drive the vehicle, it is more reasonable to identify the ‘driver’ as whatever (as opposed to whoever) is doing the driving. In this instance, an item of motor vehicle equipment, the SDS [self-driving system], is actually driving the vehicle.’” ATA “encourages NHTSA and FMCSA to work together to determine how FMCSA’s position on Highly Automated Commercial Vehicles (HACVs) without a human operator can best align with NHTSA’s prior conclusion that a self-driving system may be a ‘driver’ to ensure consistency between agencies within USDOT and avoid erecting any unnecessary barriers to the development and deployment of automated vehicle technology for all types of vehicles.”
Securing America’s Future Energy (SAFE): SAFE notes “while we acknowledge that NHTSA has limited ability to address the restrictions on AV operations stemming from FMCSRs — or state limitations on heavy-duty AV operation — we wish to underscore the importance of addressing this lacuna and urge DOT to coordinate with stakeholders on this matter.” SAFE believes that “even if a driver remains in the truck at all times, as many OEMs and experts anticipate will be the case for the foreseeable future, FMCSR restrictions threaten to reduce benefits resulting from Level 3 automation.” SAFE states that “in this context, it is crucial to ensure that policy does not place a ceiling on innovation in the heavy-duty sector. It is concerning that the guidance identifies potential obstacles to the full benefits of automation in trucks without outlining a credible plan for mitigating such obstacles. We have heard from stakeholders considering investment in new trucking technologies reconsidering because of uncertainty in the regulatory environment. Stifling innovation is not in the national interest, and we urge you to convey to the private sector that policymakers will work to create a pathway to all levels of automation for commercial vehicles that are safely achievable. The potential benefits are too great to ignore.”
Voluntary Safety Self Assessments
A number of submissions comment on various aspects of the voluntary safety self-assessments (VSSAs), including the items removed from the assessment and the 12 items which remain in it.
VSSA — Removal of Data Sharing, Privacy, Ethics and Registration/Certification Categories; and Absence of V2V
Volkswagen Group: VWGoA believes NHTSA “has made significant improvements” and that “the exclusion of the sections entitled ‘registration and certification’ and ‘ethical considerations’ help to make the guidelines more realistic and applicable for the technology that is on the immediate horizon with the understanding that these topics will continue to be discussed.” VWGoA notes that “removing the topic of privacy is also supported as we believe this section falls outside of the purview of the Agency and intended purpose of the guidelines.”
Global Automakers: Global Automakers agrees “with the agency’s decision that, while issues such as ethics, data sharing and privacy are important considerations as vehicles become more connected and automated, these items do not necessarily apply directly in the context of a safety self-assessment, but rather should be discussed through broad stakeholder engagement, separately from the NHTSA guidance.”
Connecticut DOT: CDOT states that “NHTSA dropped three important safety areas in the updated policy guidance that are critical to how automated vehicles operate safely. The three areas dropped include: privacy; registration and certification; and ethical considerations. The Department recommends that these aspects be discussed and re-inserted into future policy guidance updates.”
Insurance Institute for Highway Safety: On the subject of registration, IIHS strongly urges NHTSA “to create and maintain a public database of vehicles with automated driving systems and those exempt from Federal Motor Vehicle Safety Standards that is indexed and searchable by VIN.” IIHS believes that “in addition to permitting evaluations of automated driving systems, a public VIN-indexed database would help state motor vehicle administrations and law enforcement agencies accurately identify vehicles with and without automated driving systems in vehicle registration and crash data.” IIHS notes that “coupling state registration and police-reported crash data with information about an automated driving system’s operational design domain described in a safety self-assessment letter would enable researchers to precisely measure the safety benefits of an automated driving system based on the most relevant population of crashes.”
Consumer’s Union: On the subject of privacy, Consumer’s Union states “consumers should be able to know what data their car is collecting and transmitting, and who has access to this information. They should be able to trust that companies are legally obligated to protect their privacy and the security of their data. This trust is important not just for consumers themselves, but also for the broader acceptance and successful deployment of active safety and ADSs across the marketplace.” Consumer’s Union notes that “we understand, as NHTSA notes, that it is the FTC, and not NHTSA, that generally is charged with protecting consumer privacy. However, given the particularly sensitive data generated by self-driving vehicles and NHTSA’s general competence over vehicle safety, we believe abandoning privacy as a priority element in ADS 2.0 is counterproductive.” Consumer’s Union states that “other government agencies besides the FTC enforce privacy protections as part of their general mandate, including the Department of Health and Human Services and the Consumer Financial Protection Bureau.” Without privacy as part of a voluntary self-assessment, Consumer’s Union believes “it is all the more important for NHTSA to play a convening role and help advance binding, enforceable, and universal privacy standards for motor vehicles, including allowances for appropriate access to safety-related data for purposes of crash and defect investigations.”
Electronic Privacy Information Center (EPIC): On the subject of privacy, EPIC believes that “at this time, the FTC is simply not doing enough to safeguard the personal data of American consumers. While we respect the efforts of the Commission to protect consumers, the reality is that the FTC lacks the statutory authority, the resources, and the political will to adequately protect the privacy of American consumers.” EPIC believes “the FTC’s privacy framework — based largely on ‘notice and choice’ — is simply not working. Research shows that consumers rarely read privacy policies; when they do, these complex legal documents are difficult to understand. Nor can industry self-regulatory programs provide realistic privacy protections when they are not supported by enforceable legal standards.”
Consumer’s Union: On the subject of ethics, Consumer’s Union states “we agree with NHTSA that there currently ‘is no consensus around acceptable ethical decision-making’ related to automated driving. The lack of such a consensus is all the more reason for automakers to be transparent about the ethical approaches and tradeoffs they are taking.” Consumer’s Union appreciates that NHTSA “plans to further research the establishment of an industry framework for ethical considerations and collaborate with industry to develop standard test and simulation scenarios,” but still believes “these considerations should be part of the guidance.” Consumer’s Union suggests that NHTSA “convene an expert working group on ethics standards, in order to maximize subject-area expertise and help prevent companies from reaching conflicting or divergent conclusions on their own.”
Connecticut DOT: On the subject of ethics, CDOT states “the topic of ethics is critical to how automated vehicles operate safely and should be re-included in the next edition of the NHTSA automated vehicle policy guidance.” CDOT notes that “an ADS’s OEDR will serve as the foundation for various ethical algorithms to be written. The greater the ability for a vehicle to detect and decipher the difference between objects, the safer or less catastrophic a crash may be when an ADS is confronted with a hard ethical situation of what to do if the ADS has no good alternatives.”
V2V Standards Not Included
Global Automakers: Global Automakers states that “it is critical in future iterations of ADS 2.0 to acknowledge the important role of vehicle-to-vehicle (V2V) communications in enhancing safer, more efficient transportation in the context of connected automation. NHTSA should include additional guidance on how entities should deploy connectivity to provide enhanced safety and mobility benefits.” Global Automakers encourages NHTSA to “provide clarity” about its “future intentions on that rulemaking as it will directly influence ADS development and deployment.” Global Automakers notes “it is important that the Department continue its efforts to establish a uniform federal standard for V2V safety communications, and we urge NHTSA to maintain rulemaking on V2V as part of the agency’s regulatory agenda.”
Connecticut DOT: CDOT believes NHTSA should develop standards for V2V, noting “the potential of connected vehicle technologies to save lives, enhance mobility, and serve as the platform of a new generation of transportation management systems is vast.” CDOT states “it is recognized that the V2V standards currently under review by NHTSA will establish the regulatory foundation for many future transportation systems, including V2V, V2I, V2X, and CAV.” As such, CDOT “urges USDOT to ensure that its effort to establish a nationwide standard for V2V safety communications continues unimpeded. Any slowdown in this policy implementation will result in a substantial setback in our nation’s efforts to reduce the number of traffic crashes that result in death or injury.”
ITS America: ITS America believes “vehicle to vehicle, vehicle to pedestrian, and vehicle to infrastructure (V2X) communications can augment and support automated driving systems across all levels of automation allowing for smarter decision-making within a mixed fleet — where both conventional and automated vehicles will be operating on the same roads.” ITS America notes that “V2X communication also supports coordinated driving automation such as vehicle platooning highway speed harmonization and intersection safety.” ITS America believes “V2X will be necessary as automated systems grow in numbers and it is important to addressing safety in the transition period as driverless fleets mix in traffic and at intersections with conventional vehicles, motorcycles, bicycles and pedestrians.”
American Association of State Highway and Transportation Officials (AASHTO): AASHTO “believes the transportation industry must use every tool we can — including Dedicated Short Range Communication (DSRC) to connect vehicles with each other and the infrastructure — to make our vehicles, highways and roads safer. The potential of CV technologies to save lives, enhance mobility, and serve as the platform of a new generation of transportation management systems is vast.”
VSSA Safety Elements — The Remaining 12
System Safety
Global Automakers: Global Automakers believes the Guidance “provides the necessary flexibility by appropriately focusing on the documentation of how manufacturers and other entities have considered applicable best practices, design principles, and other relevant standards, in the design of an ADS.” Global Automakers believes this approach will “help inform consumers and regulators about testing, validation and verification methods that have been leveraged as part of the product development lifecycle to help address safety.”
Auto Alliance: Auto Alliance suggests that the language concerning design safety considerations in the Guidance should be changed from “normal (expected) driving practices” to “driving practices in the operational design domain” in order “to clarify that the intent of the VSSA is to address safety considerations of the ADS as designed.”
Securing America’s Future Energy (SAFE): SAFE suggests that “creating a benchmarked level of safety would increase public confidence and encourage consistency across developers.” SAFE also makes the following suggestions regarding standards for system safety: “Adoption of a performance standard would allow technology developers to self-certify — potentially through the Voluntary Safety Self-Assessment mechanism — that they can reasonably represent their vehicles as meeting the standard. This would have the benefit of giving developers a target to aim for, mitigating the risk that companies would either introduce the technology prematurely or delay AV introduction in a counterproductive pursuit of perfection. As is true today, NHTSA can recall vehicles it believes unsafe; however, with adoption of such a standard, companies would have a clear benchmark by which to defend their product. Additionally, a high-level standard would give companies greater visibility into their liability exposure around AV deployment — while not dispositive, meeting a federal standard would be a significant factor in tort cases.To be clear, we are not advocating a rulemaking to create such a standard in this submission. Rather, we are suggesting that NHTSA begin to explore a performance benchmark by soliciting, within the general framework of the Voluntary Safety Self-Assessment, industry opinions on acceptable levels of AV safety and potential metrics for formally assessing AV safety.”
US Chamber of Commerce/CTEC: CTEC notes that “only a certain subset of entities developing ADS technology have access or input into “overall vehicle design” and therefore recommends that “the Voluntary Guidance be refined such that the system safety design and validation process would anticipate a hazard analysis and safety risk assessment on just the ADS for those entities — including non-Tier 1 equipment suppliers, technology companies, and new entrants — that do not have access to the overall vehicle design.” CTEC also notes more broadly that NHTSA should “clarify this for any of the 12 safety design elements where an entity testing ADSs and/or automated vehicles does not have access to overall vehicle design.” CTEC also notes that “the section on System Safety also states: ‘All design decisions should be tested, validated, and verified as individual subsystems and as part of the entire vehicle architecture. Entities are encouraged to document the entire process …’.” CTEC suggests that NHTSA should “clarify that documenting ‘the entire process’ means an internal documentation of technical details, as a good portion of those details may contain confidential business information.”
National Auto Dealers Association (NADA): NADA states that “for ADS and ADS-equipped vehicles to be successful in the marketplace, prospective purchasers, lessees, and renters must be assured that their safety performance will be at least as good, if not better, as the same vehicle operated under the same conditions by human operators.” NADA therefore “urges NHTSA to set a design/validation performance standard that appropriately establishes and defines this level of system safety.” NADA supports “the Voluntary Guidance’s important suggestion that … The design and validation process should also consider including a hazard analysis and safety risk assessment for ADSs, for the overall vehicle design into which it is being integrated, and when applicable, for the broader transportation ecosystem.” NADA states it is “pleased to see that the Voluntary Guidance requires that … the process shall describe design redundancies and safety strategies for handling ADS malfunctions, effectively requiring that proper design and validation include necessary and appropriate fallbacks and operational limitations.” NADA urges NHTSA “to expressly recognize in its Voluntary Guidance the importance of manual overrides for foreseeable ‘valet’ or ‘service’ scenarios where automated operations may be too risky or impractical, such as on dealership lots and service lanes.” NADA believes that “to ensure safe and effective ADS operation, NHTSA’s guidance should expressly recognize that it may be necessary to require minimum mandatory ADS performance inspection and maintenance and to limit ADS service and repair to facilities that meet certain minimum training, equipment and security standards” and that “NHTSA should encourage states and the FMCSA to add ADS performance to their periodic motor vehicle safety inspection lists.”
American Association of Motor Vehicle Administrators (AAMVA): AAMVA notes that “while the categories of voluntary submissions cover the necessary broad safety elements, the voluntary guidance does not provide any type of criteria for assessing the technology nor any metrics or benchmarks that would assist in demonstrating the technology’s ability to satisfy a common level of safety across entities submitting entries.” AAMVA believes that “while comparisons between technologies may be different given their capacity for performing certain functions, a level of expectation on the types of data that will facilitate certain data-based metrics for submission may assist in making comparative, informed decisions regarding the level of data substantiation accompanying each submission.”
Operational Design Domain
Volkswagen Group: VWGoA “would like to clarify that for a Level 3 system, the determination of whether or not a MRC [minimal risk condition] must be achieved is left up to the fallback-ready user, and not the ADS system. The ADS system should only be responsible for issuing a takeover request when it detects its system limits and for giving the fallback-ready user enough time to resume the performance of the dynamic driving task (DDT).”
Uber: Uber states that “we appreciate NHTSA’s consideration of ODD and OEDR elements as effective concepts to enable the gradual and flexible deployment of ADSs. The Guidance helpfully notes that behavioral competencies of a particular ADS will ‘depend upon the individual ADS, its ODD, and the designated fallback (minimal risk condition) method.’ Guidance, at 7. Similarly, the Guidance notes that an ODD may be tailored ‘to initially limit the complexity of broader driving challenges in a confined ODD.’ Guidance, at 6. To that end, we support efforts by developers to define these concepts with as much granularity as necessary for safe deployment, which may involve variations in taxonomy from one developer to the next. Thus, if certain road categories or types of geographic areas would not permit safe deployment for a particular developer’s ADS, there should be no barrier to a developer constraining the ODD as much as necessary to enable safe deployment.”
Global Automakers: Global Automakers notes the Guidance provides “additional clarification on the key elements that manufacturers and other entities may consider when defining the ODD.” Global Automakers believes “it is important that manufacturers have the flexibility to both define and evaluate the specific conditions in which an ADS is designed to operate” and finds that the Guidance “strikes the right balance” in how such information “may be communicated to consumers through the VSSA.”
ZF Group: ZF “recommends that the ODD limits should also be communicated to the ADS vehicle driver and occupants, both in general at the start of a journey and if conditions change during a journey.” ZF notes that “the Guidance requests that the ADS should transition to a minimal risk condition following a change in ODD or dynamic conditions (e.g. weather), and if not able to transfer to the operator, continue to operate and mitigate risks.” ZF believes “the term ‘minimal risk’ should be used to describe a fallback position as a last resort. The vehicle should be able to transition to a minimal risk condition if the driver fails to take over the requested driving task from the ADS. This requirement should not release the driver from the obligation to respond to takeover requests within the given transition period if the system limits are reached.” ZF notes that “the system may need to operate in an emergency ‘limp-aside’ mode to bring the vehicle to a safe stop, but this could lead to complex situations if a driver attempts to intervene and retake control while the system is operating in an emergency mode.”
Connecticut DOT: CDOT suggests that for “the next update of the NHTSA policy guidance, the first sentence of this section should be revised to say ‘Entities are encouraged to define, document, properly display and enforce the Operational Design Domain (ODD) for each ADS available on their vehicle(s) as tested or deployed for use on public roadways…’.” CDOT believes that “properly displaying a description of the ODD in a vehicle, in an easy to read format for the operator and/or occupant(s) to easily view and understand is extremely important for public safety. Similarly, properly enforcing compliance with the ODD by adding system safeguards that ensure automated systems are used in the conditions for which they are designed are also very important for public safety.”
Disability Rights Education and Defense Fund (DREDF): DREDF suggests that “in addition to when and where the vehicle is designed to operate, the ODD should include who the vehicle is designed to transport. For example, the design domain should include whether a wheelchair user can be accommodated.”
Object and Event Detection and Response
Volkswagen Group: VWGoA notes that “for Level 3 systems with a fallback-ready user in the driver seat, it will still be important for the human operator to be receptive to the prevailing conditions related to the current driving situation and react accordingly.” With respect to a set of minimum behavioral competencies, VWGoA states that “behavioral competencies are heavily dependent on the intended use of the system” and that “it is completely reasonable for behavioral competencies to be excluded from the system’s capabilities if they fall outside of the intended use definition.”
Global Automakers: Global Automakers supports the Guidance on OEDR as it “provides additional clarification that manufacturers and other entities have the flexibility to define OEDR capabilities based on the design and implementation of the system within its ODD.”
ZF Group: ZF makes the following suggestions for OEDR: “The Guidance states that the ADS should be able to detect and respond to vehicles, pedestrians, bicyclists, animals and other objects that could affect safe operation and cause risk for the vehicle occupants and adjacent road users. NHTSA could consider what the boundary conditions for response to an object should be? For instance, NHTSA may define the smallest animal that the ADS system should respond to (a deer, a dog, a squirrel )? This could be set at a level beyond which potential injuries to the vehicle occupants may occur following a collision — e.g. > 100 lbs. This factor may drive the sensor performance requirements and vehicle installation configurations. The fourth paragraph in this section includes a note on ‘reasonable road etiquette,’ but NHTSA may be able to clarify more specifically what scenarios this might involve, and if any specific ADS behaviors are expected. For example, etiquette for merging allows some cars in, but this may vary depending on combinations of vehicles and driver behavior. While hazards may be standard evaluations, etiquette may vary depending on scenarios considered and how reaction mechanism s are prioritized.”
Motorcycle Riders Foundation (MRF): MRF states that “according to recent statistics, there are over 8.5 million motorcyclists on our nation’s highways. To put this into perspective, approximately one of every 36 people in America rides a motorcycle.” MRF believes that “given motorcyclists’ smaller profile on our nation’s roadways in comparison to automobiles, commercial trucks, and other road users, assurances and requirements must be met to ensure that any technology can adequately and appropriately identify and respond to motorcycles in all traffic situations.” MRF states that “for motorcyclists, our main concern includes the reliable detection of our motorcycle on the road by autonomous vehicles at ALL levels of automation. Unfortunately, this is not only a concern for future automated vehicles, but is of concern at this very moment.” MRF notes that the subject of Object and Event Detection “encourages automakers to have a documented process for assessing and testing their vehicles’ ability to avoid pedestrians, bicyclists, animals and other potential road users” and believes “it is unacceptable that this section goes as far to specify animals, and yet fails to mention the 8.5 million human motorcyclists.” MRF “STRONGLY urges NHSTA to press automakers to consider the unique attributes of motorcycles and include this growing population of roadway users to be a key consideration when developing any sort of testing requirement as it relates to safety.”
Fall Back (Minimal Risk Condition)
Volkswagen Group: VWGoA refers to the following statement in the Guidance: “Fallback strategies should take into account that, despite laws and regulations to the contrary, human drivers may be inattentive, under the influence of alcohol or other substances, drowsy, or otherwise impaired.” On this point, VWGoA notes that while “monitoring driver availability for a Level 3 system is important with respect to the functionality of the system,” VWGoA “disagrees that manufacturers should be expected to further determine if a driver is potentially impaired simply because it is an ADS equipped vehicle.” VWGoA notes that “it is not required for non-ADS equipped vehicles to make such determinations” and adds that “assuming such determinations were technically feasible, which they are currently not, it would cause significant liability concerns for manufacturers.” VWGoA states that “for a Level 3 system with a human driver as the fallback-ready user, we believe that the human driver must retain the responsibility of remaining available and unimpaired in the event that the system determines that a request to intervene is necessary. VWGoA recommends that NHTSA remove this section.”
Auto Alliance: Auto Alliance believes that “it is important to clarify that automakers should not be held to a higher standard for ADS-equipped vehicles as compared to conventional vehicles involved in crashes caused by a human driver.” Auto Alliance notes that “the human fallback-ready user of an engaged Level 3 ADS feature is expected to be ready to takeover control when needed.” Auto Alliance states that “the human fallback ready user will be given sufficient lead time to take over after being prompted to do so by a request to intervene” and that “being willing, ready and able to perform the fallback when needed remains a duty of the fallback-ready user.” Auto Alliance notes that “the manufacturer will take appropriate measures to address foreseeable misuse, such as providing a failure mitigation strategy designed to bring the vehicle to a controlled stop in place in the event that the fallback-ready user fails to perform the fallback when prompted by a request to intervene.”
Global Automakers: Global Automakers notes that “the process for transitioning to a minimal risk condition is an important factor to consider in the development of automated vehicles” but states there are “a number of clarifications that the agency should consider as part of future updates to the ADS 2.0.” Global Automakers provides the following suggestions: “For example, the Guidance should emphasize flexibility concerning the extent to which a fall back by the ADS is necessary in the presence of a trained test driver that is available to resume control of the vehicle in the event of a system malfunction. Additionally, the Guidance should more clearly distinguish between considerations for Level 3 vehicles, where a fallback-ready user is expected, and those that may be more applicable to Level 4/5 vehicles where it is not necessary for the driver to be available or present to resume control of the vehicle. For example, the Guidance should clarify that the term ‘In cases of higher automation…’ implies Level 4 and 5 vehicles. The guidance also suggests that the ‘[f]allback strategies should take into account that, despite laws and regulations to the contrary, human drivers may be inattentive, under the influence of alcohol or other substances, drowsy, or otherwise impaired.’ While manufacturers will incorporate measures to account for fallback ready driver’s error, it is impractical to take into consideration all of the possible scenarios. In addition, if the driver intentionally abuses the ADS, we have concern that the Guidance implies it is the manufacturer’s responsibility to account for all such potential scenarios.”
ZF Group: ZF notes that “the Guidance requests that the ADS alert the driver to retake control or return to a minimal risk condition independently in the event of system malfunction or degradation or reaching the limits of the ODD, as drivers may be inattentive, under influence of alcohol or other substance, or drowsy.” ZF suggests that “a driver operating a vehicle under the influence of alcohol or other substances which are forbidden by law should be regarded as an intentional misuse of the ADS and outside the responsibility of the providing entity.” ZF also suggests the term “minimal risk” could be “replaced by another phrase, since it will be difficult to define and mathematically prove an optimal minimal risk and, as a practical matter, it may be difficult to explain informally why a particular state is at lower risk than other fallback states.” ZF believes that “NHTSA could consider proposing general recommendations for driver HMI warnings and fallback strategies in the event of an ADS systems failure or system limits being exceeded, to minimize the need for unique user training per vehicle type and the risk of errors in recognition and decision-making during and after a transition — as happens with Electronic Stability Control, where a blinking yellow icon notifies the driver of a system intervention.”
Automotive Safety Council: ASC suggests that “further guidance may be necessary from the agency regarding several elements of the fallback scenario and the transition to a Minimum Risk Condition (MRC): It may be necessary to provide further definition regarding the scenarios under which the Minimum Risk Condition needs to be initiated. For example, test vehicles (under most conditions) will rely on trained drivers as the fall back for vehicle control; the agency should consider whether there is a need to differentiate the MRC in scenarios where test vehicles are being operated by trained drivers.” ASC also suggests that “the agency may benefit from providing entities that are developing automated solutions with more standard definition for what actions, mental states, behaviors, etc., correlate to the driver having positive control of the vehicle. This likely will require further investigation and research into what conditions represent the engaged, driving state (e.g., do eyes need to be open, looking forward, active engagement with the steering wheel, …).” ASC also “reiterates the need for standardized HMI for transition to the MRC so that comprehension time and the time to reengagement in the active driving task is minimized.” ASC also suggests “the agency may wish to evaluate whether there is a need for notification to surrounding traffic during transition to the MRC (e.g., hazard lights) and whether there is a need for such indications to be standardized to assure comprehension by other road users and avoid confusion.”
Validation Methods
Global Automakers: Global Automakers notes that “validation methods for evaluating the performance of ADS are an important component in the development lifecycle.” Global Automakers supports the “additional clarification encouraging entities to consider the extent to which simulation and track testing may be necessary prior to on-road testing.”
Consumer’s Union: Consumer’s Union “strongly support[s] the idea of third-party testing and certification of ADSs by expert, independent entities, provided that companies covered by the guidance also carefully test and certify the safety of their vehicles.” Consumer’s Union notes that “while the section on validation methods in NHTSA’s Performance Guidance already mentions that testing ‘could also be performed by an independent third party’ in addition to the company making the ADS, NHTSA should more strongly encourage such independent testing in the guidance and through interactions with industry. It would benefit safety not just by adding a layer of validation, but also by reducing the potential for safety compromises driven by conflicts of interest.”
Automotive Safety Council: ASC notes that “the guidelines reference the need to design and validate the automated system according to a hazard analysis and safety risk assessment and that all design decisions should be tested, validated, and verified both as individual subsystems and as part of the entire vehicle architecture.” ASC states that while it “fully concurs” it also urges “the Agency to continue to investigate and assess the potential need to drive standard validation methodologies and minimum validation levels.” ASC also “recommends that the agency consider hosting a public data repository where all safety self-assessments can be stored and accessed, potentially also including description of relevant field issues/lessons learned.”
Center for Auto Safety (CAS): CAS believes that “if the ADS 2.0 is to meaningfully protect human beings while simultaneously encouraging the development of robot-cars, section 5, ‘Validation Methods’ must be amended to explicitly prohibit the testing of Level 4 and 5 vehicles on public roads, in noncontrolled environments, unless and until these vehicles have undergone far more simulation testing — both in terms of miles and sophistication.”
Human Machine Interface
Global Automakers: Global Automakers appreciates the “additional clarification with respect to how entities should consider the design of the Human Machine Interface,” but suggests that for vehicles that are intended to operate without a human driver or occupant, the language concerning “the remote dispatcher or central control authority, if such an entity exists” be changed to “a remote operator, dispatcher, or central control authority, if such an entity exists.”
ZF Group: ZF makes the following suggestion regarding HMI: “To minimize the likelihood of consumer confusion in potential danger scenarios, NHTSA could consider recommending the development of general principles regarding HMI layout and messaging for interaction with drivers, vehicle occupants and other road users, either directly or together with standards organization such as SAE or VDA. This is especially important for other road users (e.g. pedestrians and cyclists), who may interact with vehicles from multiple entities simultaneously.”
Disability Rights Education and Defense Fund (DREDF): DREDF notes that “in daily use and in the case of operation failure, the ADSs must be able to communicate clearly with all passengers. People with disabilities will have vital design and process recommendations.” DREDF asks NHTSA to “please encourage the engagement of people with disabilities, in task forces, with organizations, engineers and testers across all 12 safety elements. It is crucial for stakeholders to engage the disability community, especially as non-traditional companies are being encouraged to innovate. There is no substitute to gathering input directly from potential users.”
National Mobility Equipment Dealers Association (NMEDA): NMEDA makes the following suggestions: “While NMEDA supports consultation and engagement of the disabled community, this portion of the guidance does not go far enough to address the specific ADS and automotive mobility needs of people with disabilities. There must be a structured environment, during the design process, to enable both the disabled community and the automotive mobility industry to share their knowledge and experience with, and to provide technical and engineering advice to, the entities engaged in AV design, testing, and deployment.” NMEDA notes legislation may require creation of advisory committees and “recommends that NHTSA not wait for these legislative proposals to become law, but rather act now to create a Federal Advisory Committee (“FAC”) under the existing Federal Advisory Committee Act (Pub.L. 92–463).” NMEDA believes “swift creation of such a FAC will significantly facilitate timely, productive discussion among federal and state governments, AV manufacturers, people with disabilities, and automotive mobility equipment industry providers (i.e., providers that design, manufacture, test, supply, sell, install, maintain, and/or repair automotive mobility equipment including but not limited to vehicle lifts for wheelchairs and scooters; in-floor and foldout vehicle ramps; specialized seating; mobility device securements; and passenger restraints).”
ITS America: ITS America believes that “NHTSA should also consider research or principles addressing the safety of human machine interfaces, in particular how drivers or passengers might handle transitions from automated driving to conventional driver or degraded operations, since these will impact traffic safety. Along these same lines, how ADS users with disabilities could interact with systems in the case of emergency is a topic that has not been fully explored by researchers and automotive designers and should be included to address accessibility implications of the new technology.”
Automotive Safety Council: ASC notes that it “has consistently commented on the strong benefit that we see from standardizing how warning and automated actions taken by the vehicle are communicated to the driver. When any active safety or automated function is activated, it is critically important that the driver comprehend what is happening quickly so that they can take any actions required.” ASC believes that “HMI standardization would also increase the effectiveness of efforts to educate drivers on what active safety features are, how they work and their benefits. As commented earlier, the need for standardization is equally critical for interactions between automated vehicles and other road users including pedestrians and cyclists, as these groups will interact with vehicles from multiple entities simultaneously and need to clearly understand an AD vehicle’s detection of themselves and intentions when maneuvering.” ASC notes that “currently, no such HMI standards exist, and we strongly recommend that NHTSA monitor and consider as appropriate in future updates of the AV policy, the efforts by external working groups (such as ISO TC22SC39 WG8 — HMI for Traffic Information and Control Systems) working in this area to develop appropriate standards addressing both the internal HMI between driver and the vehicle and external HMI between an automated vehicle and surrounding road inhabitants.”
Vehicle Cybersecurity
Uber: Uber makes the following suggestions regarding cybersecurity: “We agree with the importance the Guidance places on achieving robust cybersecurity in the operation of automated vehicles. We also believe that strong cybersecurity practices should be implemented in the development of both automated and conventional vehicles, and we suggest the agency more expressly recognize that universal applicability in future iterations of the Guidance. The Guidance also helpfully recommends drawing from best practices, advice and design principles from an array of organizations. Expanding on this approach, we believe developers in this area also should be empowered and encouraged to implement effective approaches to cybersecurity from a range of industries that are grappling with similar threats, and not just guidance from the auto industry.”
Global Automakers: Global Automakers notes that “cybersecurity is a key component for ensuring consumer trust in the deployment of ADS.” Global Automakers states that “as the automotive industry continues to take a number of proactive steps through the sharing of information and development of best practices within the Auto-ISAC, it is important that the Guidance provides entities with the appropriate flexibility to adapt and respond to changes in the cybersecurity landscape.”
ZF Group: ZF notes that “as in the section on System Safety, the second paragraph in this section suggests that an entity should simultaneously adopt best practices identified by different standard-setting bodies.” ZF suggests that “NHTSA could instead indicate that entities should develop a cybersecurity engineering process consistent with best practices published by NIST, NHTSA, Auto ISAC, etc. The entity could then identify the source of each best practice input as part of their development process.”
US Chamber of Commerce/CTEC: CTEC makes the following suggestions regarding cybersecurity: “CTEC recognizes the fluid and dynamic nature of potential cyber threats and would appreciate if NHTSA would clarify that it is not expecting entities to provide any details regarding cybersecurity processes that may contain confidential business information in the public-facing VSSA. CTEC further recommends that any vehicle cybersecurity guidance be consistent with SAE J3061 (also referenced by the ISO 26262 standard) , which encourages entities to focus product development execution in four critical security areas: Common Engineering Process that unifies functional safety and security; Secure Engineering Framework; Continuous Security Improvement model; and Supply Chain Security process.”
Niskanen Center: Niskanen notes they are “pleased to see that this guidance avoids any attempt to impose mandatory federal cybersecurity standards on autonomous vehicles.” Niskanen states that “running headlong into establishing federal cybersecurity standards runs the risk of incentivizing technological lock-in, resulting in a ‘race to the bottom’ — that is, with a baseline mandate from NHTSA and the FTC, manufacturers will be less inclined to invest in continual improvements to security.” Niskanen states that “a better alternative would be to leave standards of ‘reasonable security’ to engineers and automakers — or delegated to the National Institute of Standards and Technology — who in turn are more likely to effectively address real, proven harms to consumers.” Niskanen notes “there are powerful reputational incentives for companies to adopt strong cybersecurity standards for their vehicles. Embracing an approach that allows for competing standards to vie for market share is key to preventing technological stagnation, while ensuring consumers are better protected in the long run.”
Disability Rights Education and Defense Fund (DREDF): DREDF asks NHTSA to “please encourage transparency regarding the protection and use of data. Information regarding passenger destinations and health or disability status should not be shared with insurers or potential marketers.”
Crashworthiness
Global Automakers: Global Automakers states that “while NHTSA has established Federal Motor Vehicle Safety Standards (FMVSS) for the design and performance of motor vehicles, we appreciate the agency providing additional clarification with respect to how entities may consider alternative or planned seating or interior configurations during use.” Global Automakers believes “it is important to recognize that certain FMVSSs present barriers to the certification and deployment of automated vehicles due to the explicit references to a human driver or other physical component which may not be incorporated in an ADS-equipped vehicle (e.g. steering wheel, brake pedal).” Global Automakers states “it is critical, therefore, that NHTSA continue to aggressively research how best to modernize existing regulations to support deployment of these potentially life-saving technologies in the short term.”
ZF Group: ZF states that “current FMVSS test scenarios are not written with predictive safety systems in mind, and NHTSA may need to adapt them to account for protection systems using pre-crash data. For example, in angular front high-speed events, an ADS system may detect a potential collision and deploy an automated braking function to reduce vehicle velocity and potential crash-related injuries. However, this action may cause the vehicle occupants to move forward from their optimum locations (Out of Position case), resulting in a lower test score for the passive restraint systems unless advanced, e.g. motorized seatbelts are fitted to the vehicle to proactively tighten the seatbelts as braking begins, to maintain optimum occupant position.”
Disability Rights Education and Defense Fund (DREDF): DREDF notes that “current guidelines recommend considering alternative seating arrangements.” DREDF suggests the following: “Please recommend consideration of people of all shapes and sizes, wheelchair users, and guide dogs.”
Automotive Safety Council: ASC believes “NHTSA should keep crashworthiness as an important focus area for continued investigation to explore the implications that increased flexibility in vehicle interior design will have on occupant protection in a crash event.” ASC notes that “with automated vehicles, drivers and other occupants will have the opportunity to sit in numerous positions and postures (which may not be available or feasible in today’s vehicles), and perform a wide variety of activities including using devices like computers or other equipment for entertainment. The potential scenarios are limitless, and the consumer will have an expectation for a baseline level of safety protection irrespective of their position in the vehicle.” ASC believes “there will be a need to exercise and demonstrate a level of due care to protect occupants in non-standard seating and interior configurations and suggests that this be an area for further investigation/exploration.”
Post-Crash ADS Behavior
Volkswagen Group: VWGoA suggests removing the following language in the Guidance: “If communications with an operations center, collision notification center, or vehicle communications technology exist, relevant data is encouraged to be communicated and shared to help reduce the harm resulting from the crash.” VWGoA notes “this matter is already being dealt with within the industry, but it is not specific, nor should it be specific to ADS-equipped vehicles.” VWGoA believes “we must be careful not to confuse a vehicle’s ADS technology with the other peripheral technology that does not necessarily have an impact on the safe operation or performance of the ADS. Non-ADS vehicles can be expected to be developed in much the same way and contain many of the same functionalities with respect to non-ADS dependent technology. As such, we recommend that NHTSA remove this sentence from the guidance.”
Auto Alliance: Auto Alliance recommends amending the “Post-Crash ADS Behavior” section to “acknowledge that even in a low severity crash, vehicle sensors may become misaligned or damaged, thereby reducing the confidence for the system to make a lane change.” Auto Alliance proposes the amended language should read: “When conditions allow, actions such as shutting off the fuel pump, removing motive power, moving the vehicle to a safe position off the roadway (or safest place available), disengaging electrical power, and other actions that would assist the ADSs could be considered.”
ZF Group: ZF notes that “immediately following a collision, if the ADS detects that system performance has been impacted (e.g. due to sensor damage or misalignment), the ADS system could provide an HMI warning to the driver and operate in a degraded mode or disable certain functions in order to maintain occupant safety.” ZF also believes that “the process of repairing a damaged ADS vehicle to the manufacturer’s original design specification could additionally include alignment and recalibration of sensors in a similar way that headlight — or wheel — alignments are done today, after mechanical repairs and before a vehicle is returned to a consumer by an authorized repair facility.”
Maryland Department of Transportation (MDOT): MDOT “encourages NHTSA to work with SAE and entities engaging in testing or deployment to develop a standardized method of returning ADSs to a safe state immediately after being involved in a crash.” MDOT notes that “if each entity establishes a unique safe mode process, public safety will be reduced as emergency and first responders will be challenged with identifying the vehicle and locating the correct safe mode process prior to beginning rescue operations.”
Data Recording
Uber: Uber states that it supports “NHTSA’s encouragement to approach data collection in ways similar to the other safety design elements; that is, for developers to look to relevant guidance, standards, best practice and design principles while continuing to develop the most effective approaches to recording and collecting data.” Uber also notes “the Guidance wisely avoids mandating a particular, prescriptive approach to ADS data collection and instead encourages continual improvement and a timely emergence of standardization.” Uber believes that “given the divergence of technological approaches to ADSs within industry today, the landscape is not yet ready for a singular common approach.” As NHTSA engages further on this topic, Uber recommends “that any such process towards standardization be widened to accommodate participation from a broad spectrum of entities and relevant stakeholders.”
Auto Alliance: Auto Alliance notes that “the SAE ADS Event Data Recorder (EDR) Task Force is currently addressing which data elements and recording parameters should be included for EDRs on ADS-equipped vehicles for the purposes of crash reconstruction. As with EDRs for conventional vehicles, direct root cause analysis and fault assignment are out of scope for ADS EDRs.” Auto Alliance states that “members will continue to contribute to the SAE ADS EDR Task Force to determine the most appropriate data elements and recording time frames to support crash reconstruction for HAVs.” Auto Alliance also believes that for crash data, “the ownership model for EDR data is no different for ADS equipped vehicles than for conventional vehicles. EDR data belongs to the vehicle owner. Any shift in this ownership model would have significant privacy implications.” Auto Alliance also believes “any change would also need federal and state law changes because today’s EDRs provide rights to the vehicle owner.”
Global Automakers: Global Automakers states that it “supports NHTSA’s decision to separate the issues of data recording and sharing as contemplated by the FAVP. While the issue of data sharing is an important element that will need to be addressed as ADS technology continues to evolve, it is not necessary to include it as part of a VSSA.” Global Automakers recommends the following revision: “It should be clarified, however, that crashes may not be attributable to the performance of the automated vehicle, but rather other vehicles or road users. We therefore recommend that NHTSA consider the following revisions to this section of the Guidance: ‘Therefore, entities engaging in testing or deployment are encouraged to establish a documented process for testing, validating, and collecting necessary data related to the occurrence of malfunctions, degradations, or failures in a way that can be used to help establish the cause of any crash that may be attributable to the ADS functionality.’” Global Automakers also recommends the Guidance be amended as follows to acknowledge federal and state law may limit sharing of information with the government: “Entities should, to the extent appropriate and permissible under the circumstances, have the technical and legal capability to share with government authorities the relevant recorded information as necessary for crash reconstruction purposes.”
Bosch: Bosch notes that “during the testing process, large amounts of data ranging up to several petabytes may be recorded depending on the individual function, component and system being tested” and that “making such data readily available for retrieval remains a challenge due to the download, storage, processing and maintenance requirements of such vast amounts of data.”
ZF Group: ZF notes that “a key difference between today’s passive safety EDR data and future ADS system data is that passive safety systems can only collect and store the user’s data, while pre-crash accident data stored by an ADS vehicle may also contain data (e.g. speed, location) collected from other vehicles or road users, giving rise to additional privacy concerns. If pre- and post-crash data is to be stored, guidelines or recommendations could be developed by NHTSA to identify the minimum set of data, and time periods before and after an event that data should be collected and stored — e.g. from ten seconds before a collision to five seconds afterwards.” ZF also believes “additional information that could be captured includes braking or steering interventions, driver hands-on status, environmental sensor (e.g. radar, camera, V2V) tracking data and vehicle position. Depending on the extent of the final requirements for data parameters (e.g. camera images, lidar, radar yaw, lateral acceleration data), required resolution of stored data, and the time window of pre- and post-crash buffered data, this could drive significant hardware requirements and some hardware costs for the supplier of the safety modules that are collecting this data and calculating and distributing system response commands, including non-volatile memory size and energy reserves for post-crash operation of the ECU(s) to allow data collection and storage to be completed.”
State Farm: State Farm states that “in addition to the data elements outlined in 49 CFR Part 563, State Farm’s own research indicates that the following data elements should be considered when creating uniform standards for crash reconstruction purposes:
Message Packaging
• Message ID
• Message Count
• Temporary ID
Location
• Vehicle Position Reference Point
• Longitude and Latitude
• Elevation
Movement
• Heading
Additional Event Based Information
• Path History
• Path Prediction
• Exterior Lights
Vehicle Based Motion I ndicators
• Transmission State
• Steering Wheel Angle
Additional Data
• Front Crash Prevention System state/Action (and number of actions)
• Rear Crash Prevention System State/Action (and number of actions)
• Blind Spot Warning State/A ction (and number of actions)
• Lane Departure Warning/Lane Keep Assist State/Action (and number of actions)
• Driver Messages/Warnings in Instrument Cluster (warnings for Advanced Driver Assistance Systems (ADAS) such as driver notifications or systems malfunctions)
• Software V ersion of each Electronic Control Unit ( ECU)
• Current Stored or Active Diagnostic Trouble Codes (DTC)
• Malfunction Indicator Lamp (MIL) Light(s) Illuminated
This list is not intended to be exhaustive, and instead is intended to highlight the minimum amount of data that is needed for effective crash reconstruction.” State Farm believes “NHTSA should look to expand its section on data recording and sharing to ensure insurer access to this information” so that “insurers will be better able to underwrite risks, determine liability, and settle claims.”
Travelers: Travelers notes that it “supports standardization (via legislation or regulation) of data collection, sharing, storage, and security requirements. While Travelers is agnostic with respect to who develops and imposes data management requirements and what those specific requirements may ultimately be, standardization would clarify obligations and expectations regarding who must report what, when, as well as who is entitled to receive information and for what purposes.” Travelers believes that “broader data sharing with insurers would benefit all ADS stakeholders, including ADS owners and accident victims. Specifically, to facilitate an effective and efficient ADS auto insurance system, the federal government should require timely data sharing (by auto manufacturers and others who obtain data on crashes and ADS performance) with insurance providers.” Travelers contends that “sharing data with insurers would help facilitate proper insurance coverage in several ways; including: Establishing liability/causation in the event of an accident (a function performed by the insurance carrier, not the customer); Assisting with accurate underwriting and pricing of insurance policies; and Contributing to risk mitigation and control measures (e.g., via software updates).”
Insurance Institute for Highway Safety: IIHS notes that it is “disappointed that the guidance gives no indication of a plan for collecting information that would support research evaluating the real-world safety of automated driving systems, assessments of safety claims made by the companies deploying them, or exemptions from Federal Motor Vehicle Safety Standards, as contemplated in pending legislation.” IIHS strongly urges NHTSA “to require that companies designing automated driving systems for use on public roads provide the following information to NHTSA to be made publicly available: 1. A description of each automated driving system, its level of automation, and its operational design domain; 2. Information that identifies vehicles equipped with one or more automated driving systems; and 3. In the event of crashes, information describing the state (engaged or not) and performance (measures describing its motion and the motion of other involved vehicles) of an automated driving system equipped to a vehicle involved in a crash.”
AASHTO: AASHTO “continues to have a number of concerns regarding Data Recording. These concerns are reflected in questions that include: Who is this information intended to be shared with? Is it limited to NHTSA, or will state and local law enforcement agencies, state DOTs, and insurance companies have access? Will data sharing be the prerogative of the individual manufacturers, or will there be regulation about which entities have access? Who owns and controls this data: the vehicle owner, the manufacturer, or a government agency?” AASHTO notes that “state and local agencies would be able to make productive use of this data to evaluate hazardous road conditions and other things, and will be interested in this data. However, most state agencies are subject to government records requests, which can become very burdensome if the data can be tied to specific instances.” AASHTO also recommends “that the data for which events are shared includes noncrash data such as since ‘near miss’ and disengagement events which can be as important as crash scenarios when assessing road conditions. Currently, the data recording is suggested to be limited to fatal crashes, personal injury crashes, and crashes involving towed vehicles.”
Automotive Safety Council: ASC suggests that “further work may be required by the Agency to establish standards and requirements for the types and amounts of data necessary to adequately perform post-crash analysis. The guidelines also reference the need to ensure all applicable considerations and protections are taken with respect to ensuring privacy relative to the data collected.” ASC notes that “it will be important that future guidance be very clear regarding the collection and sharing of such information that may be critical for autonomous vehicle operation.”
Connecticut DOT: CDOT encourages “the following data elements to be considered part of the uniform data elements collected as part of the crash investigation: Specific details of the crash, including information as to the possible cause or causes of the crash;Whether the ADS was activated immediately prior to and/or during the crash; Whether or not the ADS was functioning properly within its ODD immediately prior to and/or during the crash; If the ADS was inactive or operating outside its ODD, information about how, when and why the system deactivated or disengaged from autonomous vehicle mode and/or from its ODD; Whether the ADS suffered any additional technical failures; and any other details pertinent to the crash not already captured.” CDOT also suggests that “broad sharing of information associated with crashes and near-misses should become a best practice standard so that collective learning can happen while still protecting proprietary information of connected and automated vehicles. Additionally, there may need to be some federal requirements developed that prevent reasonable data information from being labeled proprietary.”
Transportation for America (TFA): TFA recommends “that the guidance be expanded to encourage the collection and reporting of data on disengagements in a similar manner to the guidance on crashes. Understanding why an AV has to disengage will allow NHTSA to work with industry, cities, states and law enforcement to improve the safety of AV’s before they get involved in crashes and potentially seriously hurt persons or property.” TFA recommends that “NHTSA include in any voluntary guidance the following disengagement data that cities and states have identified as useful in their efforts: 1) The outcome of the disengagement; 2) the post crash behavior of the vehicle, including whether the vehicle tried to continue driving after the disengagement; 3) location information data such as GPS coordinates, cross streets or road mile markers where the disengagement occurred; 4) date and time of the disengagement; 5) infrastructure considerations including disengagements due to confusing or unusual traffic signals, roadway signs, lane markings, railroad crossings, road designs, etc. 6) the vehicle speed and acceleration or deceleration at disengagement; 7) weather conditions at the time of the disengagement and 8) other parties involved in the disengagement such as other human driven vehicles, other AV’s, pedestrians, bicyclists and motorcyclists.”
TFA also recommends “that NHTSA initiate a rulemaking requiring manufacturers to share data with relevant entities like NHTSA and states, local governments and law enforcement that have jurisdiction over areas AV’s are deployed. Such a rulemaking would enhance safety and the ability for governments to work with manufacturers, operators and testers to identify and solve problems with AV’s before they kill or injure humans and damage property.”
Maryland Department of Transportation (MDOT): MDOT suggests that“NHTSA should consider including as part of data collection, data not only safety focused, but also·data that would help state and federal agencies implement efficient mobility strategies, including information linked to vehicle operations, interaction with infrastructure (e.g. speed ofdata transfers), and route information.”
Consumer Education and Training
Global Automakers: Global Automakers “agrees that education and training is a key consideration in the deployment of ADS.” Global Automakers believes “the Guidance should make clear, however, that this section is applicable or need only be considered when testing or deployment involves consumers that are not employees or designees of a manufacturer or entity.”
ZF Group: ZF suggests that in addition to consumer training, “training be extended to first responders so that they understand how to interface with ADS vehicles in emergency and crash situations. NHTSA could also update the safercar.gov website or conduct other outreach activities to provide information to and educate consumers on ADS technologies and vehicle capabilities.”
American Trucking Associations: ATA notes that “the education and training programs for automated driving systems for commercial vehicles will be different from those developed for passenger vehicles; however, we believe that the guidance provides sufficient flexibility to allow for those differences. The commercial vehicle purchasing process is a business-to-business transaction, with the purchaser generally more informed and knowledgeable about the product than an average consumer purchasing a passenger vehicle, who may require more information at the point of sale and training upon delivery of the vehicle. Fleets purchasing vehicles for their drivers can be expected to be a partner with the manufacturer of an automated driving system or highly automated commercial vehicle when it comes to training their drivers.”
Disability Rights Education and Defense Fund: DREDF makes the following suggestion: “Consider recommending disability sensitivity training for entity staff, marketers, dealers and distributors. Recommend that materials be available in accessible formats, that braille and captioned videos are provided.”
Federal, State and Local Laws
Global Automakers: Global Automakers supports the Guidance “clarifying that for testing purposes, an entity may rely on an ADS test driver or other mechanism to manage compliance with applicable laws.”
ZF Group: ZF notes that “overall, to promote widespread deployment of ADS vehicles, ZF suggests that safety and regulatory requirements are made as consistent as possible across the U.S. to avoid a patchwork of regulation that could hinder market penetration of ADS technology.”
Guidance to States
Some comments address the guidance given to the states regarding the division of responsibility between the federal and state governments.
General Motors: GM agrees that “states should ‘review others’ draft ADS policies and legislation and work toward consistency’ and that ‘[t]he goal of State policies in this realm need not be uniformity or identical laws and regulations,’ but rather, ‘sufficient consistency of laws and policies to promote innovation and the swift, widespread, and safe integration of ADSs.’” GM notes that “in the past year, seven states have achieved this goal: Michigan, Georgia, Colorado, Texas, Tennessee, Nevada, and North Carolina. Each of these seven states enacted HAV legislation that promotes innovation and swift, safe introduction of HAVs to consumers — and that is consistent with the HAV legislation enacted this year in the other six states.” GM notes that these state HAV laws “include the following elements:
• States removed legal barriers to testing and deployment of HAVs, including HAVs without human drivers behind the wheel;
• States adopted a self-certification model, not an application-based, pre-approval approach, for both testing and deployment of HAVs;
• States allow commercial deployment of HAVs when entity satisfies financial responsibility requirements and certifies that the vehicle is capable of complying with applicable traffic and motor vehicle laws and meets other statutory requirements; and
• States rely on existing tort laws and principles to protect consumers and adapt to new technologies, as those laws have successfully adapted in the past, and allow real-world experience to guide any changes that may be desirable in the future.”
In addition, GM “appreciates that the Guidance encourages state regulation of HAVs to maintain statewide uniformity.” GM notes however that “the Guidance indicates that local authorities may be separately empowered to regulate HAV activity.” Instead, GM “encourages DOT/NHTSA to clarify that, to preserve uniformity, any state HAV regulation should remain exclusively at the state level.”
Ford Motor Company: Ford “shares NHTSA’s views about the delineation of federal and state roles in ADS 2.0, in particular that states should remove statutory and regulatory barriers to testing and deployment.” Ford also appreciates “the clarification that the VSSA should not be codified” by states.
Volkswagen Group: VWGoA objects to the statement in the Guidance that “states can take steps to monitor safe ADS operation through reporting and communications mechanisms so that entities can coordinate with public safety agencies.” VWGoA believes this recommendation “is too generic and could be interpreted to mean that States should collect whatever data they deem necessary to effectively regulate ADS operation within their jurisdiction.” VWGoA notes that States “can interpret the word ‘data’ to mean a variety of different things” which becomes “problematic for testing because States may request certain data that provides little to no value to be shared, but places significant burden on the entity conducting the testing.” VWGoA notes this “becomes even more problematic for deployed ADSs as the data collected in those vehicles will be controlled by the owner, not the manufacturer.” VWGoA recommends that “the Agency remove this section from the guidance.”
Uber: Uber supports “NHTSA’s efforts to emphasize primary federal oversight of vehicle design and performance, while carefully delineating federal and state responsibilities over motor vehicles and their operation.” Uber notes that “in favoring the rapid development of safe ADS technology, we agree with NHTSA’s direction that states avoid imposing parallel and conflicting requirements on the design or performance of ADS, as well as take steps to remove unnecessary barriers in existing state laws to ADS deployment.” Uber also agrees that “states should provide a ‘technology-neutral’ environment that does not limit testing or deployment opportunities — whether by type of technology, by type of vehicle platform, or by type of developer (e.g., no restriction favoring traditional vehicle manufacturers versus other entities that might be involved in responsibly developing ADS technology).”
Auto Alliance: Auto Alliance “appreciates the clarification in the Guidance regarding the Federal and State roles, but we are concerned that the focus of Section 2: Technical Assistance to States seems to be on testing in a pre-approval structure, and in some cases may even suggest pre-approval requirements for deployment.” Auto Alliance believes “this guidance could lead to numerous state-by-state application processes that would inadvertently create unnecessary delays to HAV testing and deployment.” Auto Alliance recommends “that the Agency and States avoid pre-approval processes for testing and deployment. For example, the legislatures of Colorado, Georgia, Michigan, Nevada, North Carolina, Tennessee and Texas have each passed laws that allow for testing and deployment of HAVs on public roads, with and without human drivers. These laws rely on self-certification and do not require an application or pre-approval prior to deployment. Legislation of this kind paves the way for driverless deployment, while allowing NHTSA to fulfill its role as regulator of vehicle safety performance.”
Global Automakers: Global Automakers “supports the agency’s approach in providing specific best practices and recommendations for consideration by state legislatures as they relate to the development of potential future legislation.” Global Automakers believes “it is important to ensure a technology-neutral environment for the testing and deployment of ADS.” Global Automakers believes “the Guidance could be improved further by providing additional clarification in the Best Practices for Legislatures, urging states not to regulate the performance and design of ADS — consistent with the appropriate roles of federal and state governments. This should extend to system performance indirectly related to the ADS, such as a crash data recording devices, for example.” Global Automakers notes that “the United States has long recognized that automobiles are sold in a national market, and that manufacturers’ success rests on the ability to produce vehicles that can be sold and operated in all fifty states. Additionally, we urge the agency to include a recommendation that states consider the impact of policy decisions on the ability of a vehicle to travel across state lines.”
Maryland DOT: MDOT suggests that “NHTSA should consider an update to the Uniform Vehicle Code (UVC) that considers and addresses issues unique to ADS and is coordinated with the provisions of the Manual of Uniform Traffic Control Devices (MUTCD). These updates or other guidance or model policy should include standardized and clarifying definitions and any infrastructure guidance for safe operation of ADS.”
NYC Mayor’s Office: NYC Mayor’s Office notes that “our greatest concern continues to be the involvement of cities and urban transportation officials in the development of federal HAV policy.” NYC Mayor’s Office “hopes that moving forward, USDOT and NHTSA will engage with cities more directly, and create more opportunities to share best practices across levels of government. To that end, we advocate that USDOT expand the ACAT to include more city representatives, and commit more resources to the Automated Vehicle Proving Grounds program, including more urban areas and sharing testing and deployment best practices with cities not directly involved in the program.” NYC Mayor’s Office also “encourages NHTSA to work with industry and governmental stakeholders to identify important data sets and develop a system to share them with transportation agencies, while taking the steps necessary to protect consumer privacy and essential proprietary information.” In addition, NYC Mayor’s Office “feels strongly that it is critical for NHTSA to promulgate HAV performance standards, as the necessary first step toward either rigorous manufacturer self-certification or direct NHTSA pre-market certification. In that vein, NYC encourages NHTSA to continue its evaluation of the appropriateness of new, more active certification authority, particularly pre-market certification, as explored in the original Federal Automated Vehicles Policy.”
Transportation for America: TFA states that “we are disappointed that NHTSA only includes guidance for state legislatures, but not for the local governments and law enforcement agencies that do the bulk of roadway safety enforcement. These stakeholders are already on the front lines dealing with the impacts of AV’s being tested and deployed in the communities they serve.” TFA notes that “with AV legislation currently being considered in Congress that would preempt cities and states from a number of their current regulatory authorities, we urge NHTSA to include relevant guidance in future updates to clarify the legal uncertainty on how state and local governments should update and enforce their laws and regulations.”
American Trucking Associations: ATA states that “with regard to federal and state roles, the regulation of performance and technical specifications of automated and connected truck technology should be solely the responsibility of the federal government. States should maintain their existing responsibilities that do not interfere with the flow of interstate commerce. In the absence of federal regulation, states should support operations of commercial motor vehicle automated and connected technologies within their rights of intrastate jurisdiction.” ATA notes that “conflicting or duplicative requirements among federal and state agencies will create roadblocks to deployment of automated technology, delaying the safety benefits, fuel savings, emissions reductions, and potential efficiency improvements to our country’s transportation system.” ATA also “concurs with NHTSA’s recommendation that states should identify and change traffic laws and regulations that may serve as barriers to operation of automated driving systems.” ATA believes that states “should commit to ensuring a unified national framework to facilitate the development, testing, and deployment of commercialized automated and connected truck technology, including further harmonization of state-level traffic and vehicle rules affecting the operation of such technology.”
AAMVA: AAMVA is “concerned that in the consideration of respective state and federal roles, the responsibility for crash investigation was not included under ‘states responsibilities’” and believes that “states are uniquely equipped to handle this important safety function.” In addition, AAMVA believes that “state governments also have a responsibility to ensure their constituency is informed of the progress of ADS technologies. AAMVA members understand that whether persons choose to engage with or utilize these technologies, the general public needs to be informed of how the government is working to ensure their safety as the foreign concept of driverless cars pervades society on a more dramatic level. This includes educating the public on how to prepare for vehicles with no steering wheel, additional distractions between traditional and driverless vehicles, the age of passengers being transported in vehicles, and providing a sense of comfort as society transitions to a new norm.”
Travelers: Travelers encourages “the development of a model state law that will help avoid inconsistencies in state regulatory approaches, which could undermine the development of sound policies for ADS development and deployment.” Travelers is also “particularly supportive and interested in further development of the following elements within the Guidance: Recognition of auto insurance as a primary risk transfer mechanism and compensation system in the ADS world and development of appropriate insurance-specific policies and regulations; Standardization of ADS data collection/management — especially surrounding data sharing with public and private stakeholders, including insurers — and cybersecurity requirements for ADSs with input from private industry experts; Coordination and collaboration among policymakers and private stakeholders to maximize uniformity in the legal/regulatory frameworks for ADSs; and Effective communication of safety-related information to consumers.”
Travelers notes that “among the primary legal/regulatory issues states are encouraged to address (in both the legislative and regulatory spheres) under the Guidance is motor vehicle insurance and liability, including: How to allocate liability among ADS owners, operators, passengers, manufacturers and others when an accident occurs; and Who should be responsible for maintaining insurance coverage (e.g., owner, operator, passenger, manufacturer, others, etc.).” Travelers “agrees with the Guidance’s recommendation that financial responsibility be required for all ADS-equipped vehicles and that insurance be required at the testing stage. We also agree with the recommendation that states begin the process of setting rules to determine how to allocate responsibility for crashes and which persons/entities must carry insurance.”
Travelers supports “early adoption of appropriate rules and regulations to address these and other insurance-specific issues. For instance, insurance regulators must ensure that there are sufficiently high coverage limits, including adequate limits for property damage to account for more expensive technology in ADSs.” Travelers believes that “auto insurance is best suited among risk transfer regimes — versus, for instance, a product liability-based system — to address compensation issues (e.g., bodily injury, property damage, cyber incidents, etc.) arising out of the emerging ADS world.”
