Comment from Anonymous

Several public comments argue that physical brake pedals must be retained as a mandatory failsafe across all vehicles. However, forcing a physical brake pedal into a fully automated vehicle rests on an unstated assumption that a pedal-free design cannot be made as safe, an assumption the record directly refutes. Waymo's June 2026 safety data, 82% fewer injury-causing crashes and a 94% reduction in serious crashes across 220 million fully driverless miles, show that this safety performance was achieved without reliance on manual intervention. Waymo's vehicles retain unused steering wheels and pedals throughout these miles, which sat idle while the software drove. Separately, Zoox's purpose-built robotaxi, which NHTSA has already exempted from FMVSS requirements after finding it provides an overall safety level at least equal to nonexempt vehicles, confirms that a vehicle with no manual controls at all can meet that same bar. Together, these show that a pedal-free architecture is not hypothetical: it is already operating under NHTSA-approved exemptions on public roads. While these figures do not dictate that every ADS design is universally safer, they establish that a pedal-free architecture is achievable, which is precisely the threshold a performance-based, design-neutral standard requires. Mandating manual controls in proven ADS-equipped vehicles restricts engineering freedom in exchange for a perceived safety benefit we do not require elsewhere in the vehicle code. Because both architectures are demonstrably achievable, both should remain legally available, with the choice between them left to the manufacturer's design approach. Furthermore, demanding a manual brake pedal as a failsafe for catastrophic ADS failure applies an inconsistent safety standard. In a human-driven car, the human driver is already a single point of failure, and when a driver suffers a medical emergency, the agency accepts total loss of human control by not requiring any passenger-operated backup control. Purpose-built Level 4 vehicles, by contrast, are commonly engineered with redundant compute, steering and braking systems, and independent power sources: if the primary ADS computer fails, the secondary system is designed to execute a Minimal Risk Condition (MRC) stop. A 2025 connectivity-dependent stop during a citywide power outage and a 2026 recall for delayed stopping on flooded roads show genuine software and edge-case limitations, not hardware failure. Critically, a physical brake pedal at the driver seat would not have addressed the documented causes of these events. While safety engineering requires mitigations against common-cause software failures, a traditional floor pedal is not the only out-of-band solution. A manual pedal places emergency vehicle control in the hands of an occupant who may have little or no training and may be under significant stress, whereas a pedal-free architecture allows manufacturers to implement out-of-band, hardwired emergency-stop circuits backed by isolated, independent power supplies that bypass the primary compute to trigger the vehicle's MRC stability controls. Retaining the prescriptive FMVSS 135 foot-pedal mandate introduces an additional human-factor risk while not addressing the documented software, connectivity, and edge-case limitations discussed above. Finally, public concerns regarding the behavioral testing and safety of the ADS software are valid, but they fall entirely outside the scope of FMVSS No. 135. Physical hardware rules should not be held hostage while the agency develops separate behavioral driving standards. The hardware standard must allow for pedalless designs now, so that when the software standards are finalized, the vehicles are structurally optimized for true autonomy.