Fully Autonomous Vehicles for People with Visual Impairment: Policy, Accessibility, and Future Directions
Paul D. S. Fink, Jessica A. Holz, Nicholas A. Giudice · 2021 · ACM Transactions on Accessible Computing · doi:10.1145/3471934
Summary
This policy analysis examines how fully autonomous vehicles (FAVs) could transform mobility for people who are blind or visually impaired (BVI), while arguing that current US legislation fails to address their accessibility needs. With over 25 million Americans reporting travel-limiting disabilities and 26.9 million experiencing visual impairment, FAVs represent a potentially life-changing technology—yet the policy landscape actively discriminates against this population. The authors review both state and federal FAV policy through an accessibility lens. At the state level, 30 states have passed autonomous vehicle legislation, with 25 addressing passenger licensing requirements. Critically, 11 states definitively require a driver's license to operate any autonomous vehicle, effectively barring people with visual impairments from using technology specifically designed to eliminate the need for human driving. The authors argue this requirement is fundamentally illogical: FAVs are designed to remove steering wheels, pedals, and other manual controls, making traditional driving skills irrelevant. At the federal level, promising initiatives have stalled or been weakened. The 2016 Vehicle Performance Guidance for Automated Vehicles (VPGAV) included enforceable accessibility requirements, but was abandoned in favor of voluntary guidance in subsequent Automated Vehicles 2.0-4.0 reports. The SELF DRIVE Act, which includes anti-discrimination provisions, passed the House unanimously but has repeatedly stalled in the Senate. The result is a "laissez-faire" regulatory approach where accessibility remains optional.
Key findings
User research with BVI populations reveals both enthusiasm and skepticism about FAVs. A 2020 survey of 516 BVI respondents found 88.87% view FAVs positively, with over 90% interested in ownership. However, 94.3% expressed concern about discriminatory laws preventing them from operating these vehicles. More than half of focus group participants believed their needs were not being adequately considered in FAV design. BVI users identified specific accessibility requirements across the complete journey: locating the vehicle, verifying correct destination arrival, receiving real-time operational information, and navigating entry/exit processes. 71% wanted voice/dictation input as a primary interaction mode, while opinions on in-vehicle touchscreens were mixed due to current accessibility limitations. Research with rideshare services (as a proxy for FAV experience) revealed that locating the vehicle was the most difficult part of the trip for BVI passengers. The authors synthesize existing user research into three core policy recommendations: (1) eliminate state laws requiring driver's licenses for FAVs by creating a separate "FAV operator's license" not dependent on vision tests; (2) extend ADA Part 37 transit requirements to FAV artificial intelligence, including destination announcements and accessible information formats; and (3) revise and reinstate the VPGAV with enforceable accessibility mandates rather than voluntary guidance.
Relevance
This paper provides essential policy context for anyone working on accessible transportation technology. The analysis reveals a critical disconnect: while researchers and advocacy groups have clearly articulated BVI accessibility needs for FAVs, this knowledge has not translated into enforceable policy. The voluntary nature of current federal guidance means manufacturers face no consequences for ignoring accessibility. For practitioners, the paper highlights smartphone integration as a pragmatic near-term solution. BVI users already have accessible smartphones with screen readers, voice control, and haptic feedback—FAV systems could leverage these existing tools rather than requiring inaccessible in-vehicle interfaces. The vibro-audio interface (VAI) concept, which conveys spatial information through touchscreen vibration patterns, offers a promising approach for route visualization and real-time navigation feedback. The "paradox of FAV information access" is particularly relevant for AI ethics: users want personalized, responsive service but distrust the data collection required to enable it. With over 70% of Americans already distrusting self-driving cars, accessible and transparent design isn't just an accessibility issue—it's essential for mainstream adoption. Organizations developing FAV technology should treat accessibility as a core requirement, not an afterthought, and engage with BVI communities throughout the design process.
Tags: autonomous vehicles · transportation policy · blindness · visual impairment · accessible design · artificial intelligence · mobility
Standards referenced: ADA · ADA Part 37 · VPGAV · SELF DRIVE Act · SAE J3016