On May 28, the National Highway Traffic Safety Administration (NHTSA) released an advance notice of proposed rulemaking (ANPRM) soliciting public comments on proposed testing and verification rules for automated driving system-dedicated vehicles (ADS-DVs) with Federal Motor Vehicle Safety Standards (FMVSS). FMVSS are designed to (1) reduce the likelihood of a crash occurring and reduce the severity of crashes when they do happen (in the 100-series FMVSS), and (2) reduce the risk of injury to vehicle occupants in a crash (in the 200-series FMVSS). The document addresses only 100-series FMVSS, with a stated intention to release further documentation to address 200-series FMVSS.

The nature of FMVSS being developed before the emergence of ADS technologies has resulted in narrow definitions of operational and safety features that pertain specifically to in-vehicle human drivers. Currently, manufacturers of highly automated vehicles skirt the FMVSS requirements in a limited and temporary manor. Manufacturers can apply to NHTSA for an exemption from FMVSS for up to 2,500 vehicles sold after proving to NHTSA on a specific case basis that the vehicle is as safe or safer than a compliant design by meeting performance standards.

The ANPRM is NHTSA’s first step toward creating comprehensive safety standards designed for vehicles with ADS. The document pertains only to ADS-DVs, indicating that any changes initiated under the ANPRM would not affect FMVSS for vehicles that at any time would operate under manual control, including remote control (except during testing). Given that current FMVSS are not prohibitive of ADS technologies, the changes would be purely to allow for further innovation in design and technologies in ADS-DVs. There are two ways that NHTSA plans to update FMVSS for ADS-DVs:

1. to add safety standards for AV technologies; and
2. to remove requirements that pertain only to manual driving tasks.

Developing new standards for technologies that would perform the driving task will entail considerably more research, according to NHTSA, and thus the current ANPRM addresses only amendments to existing FMVSS.

Removing requirements for elements such as the presence of a steering wheel and positioning of a driver can be examined in a fairly straightforward way from a performance-based perspective. While including manual driving components in a dedicated automated vehicle does not limit safety performance of the vehicle, they can take up space which can decrease capacity, they can add cost and add weight which decreases energy efficiency. NHTSA identifies these requirements as “barriers” in the FMVSS for ADS innovation. Steering performance of ADS does not need to depend on a wheel location and size, and performance standards for reaction time and precision could be developed for automated steering systems.

In order to initiate adjustments the existing standards, NHTSA identified three types of barriers to ADS-DVs complying with existing FMVSS:

1. If the standard requires manual control (such as a “foot activated brake pedal”).
2. If the standard specifies how the agency will use manual controls in the regulatory description of how it will be tested.
3. If the definition or use of the terms (e.g. ‘driver’) in the FMVSS that assume human control of the vehicles.

USDOT began to address the third type of barrier with their October 2018 guidance document, Preparing for the Future of Transportation 3.0, in which they commit to adjusting the definitions of “driver” and “operator” to include instances when an automated system may be performing driving functions. The other two types of barriers must still be addressed. The ANPMR includes a table with specific examples of 100-series FMVSS that face barrier types 1 and 2.

Table 1. Examples of FMVSS with barriers, by barrier type

Source:NHTSA, ANPRM 84 FR 24433

NHTSA has worked extensively with research organization including the Volpe National Transportation Center and the Virginia Tech Transportation Institute (VTTI), whose research they reference directly in the ANPRM.

The report from Volpe identified few barriers for testing ADS, but notes the difficulties for certification for ADS-DVs with unconventional designs. The set of proceedings from a convening between USDOT and VTTI demonstrates the need for, methodology behind, and considerations for FMVSS for ADS. The stakeholder group convened by the VTTI researchers focus on testing methodologies that include simulation for testing both within and outside of the operational design domain of the vehicles, and stress that research and testing will have to occur in close collaboration with the OEMs. The VTTI research provides information for NHTSA “technical translations” for FMVSS to reduce barriers for ADS-DVs.

Prior to concluding the ANPRM addressing public commenting, NHTSA walking through six potential methodologies for testing ADS-DVs to ensure compliance with safety standards:

1. Normal ADS-DV operation;
2. Test mode with pre-programmed execution;
3. Test mode with external control;
4. Simulation;
5. Technical documentation for system design and/or performance approach; and
6. Use of surrogate vehicle with human controls.

Each of these methods have benefits and challenges, and effective testing is likely to encompass a combination of the options. NHTSA is asking for detailed comment on advantages and disadvantages, safety standard applicability, redundancies and uniqueness, performance standard viability, terminology considerations, and research considerations, of each approach.

Addressing changes for the 100- and 200-series FMVSS is a step towards allowing for innovative designs in ADS-DVs, and while developing testing methodologies and final performance standards will take more research and development, the technologies are still years away from being market-ready in a context that would not require a safety operator either internal or external to the vehicle. NHTSA’s iterative approach with frequency and direct feedback from stakeholders allow for time to consider comments and develop a framework that encourages safety while allowing for innovation in design. The testing and standard development could also apply to other types of vehicles, such as vehicles that has SAE level 4 technologies, furthering the benefit of the efforts behind this ANPRM.

Eno’s report Beyond Speculation 2.0 further discusses recent and future regulation and legislation pertaining to FMVSS for automated driving systems.