After the initial wave of high expectations, confidence in automated driving has noticeably waned among both industry professionals and the general public. The early years were marked by unrealistic hopes for a swift rollout. However, the market has since stabilized, and we are now observing consistent technological advancements. The reality of self-driving cars is closer than ever, yet significant challenges must be overcome to ensure their safe and responsible introduction. Crucial to this progress are the advancements in quantifying safety risks and the growing importance of AI in safety assessments.
One of the main challenges in automated driving is making safety risks quantifiable. Leading companies in the field are pioneering this area. The US serves as the main testing ground for automated trucking due to its traffic infrastructure and regulatory framework. However, incidents can inflict significant reputational damage, even when safety is not enforced by explicit government rules. This has been evident with the challenges faced by robot taxis.
To ensure trucks can handle virtually all practical situations, manufacturers need to conduct thorough safety assessments in addition to ensure safety. The challenge with real-life road testing is that most kilometres driven are uneventful. To address this, a comprehensive model of public roads must be developed using innovative methodologies. By distilling data from many test-driven kilometres on conceivable driving scenario, including challenging situations, the risk of incidents can be estimated. This approach quantifies safety risks effectively, turning subjective concerns into objective metrics that can be systematically addressed.
Safety is not just about statistics but also about perceived safety. Building confidence in automated vehicles among other road users is crucial. Programs designed to investigate and enhance public trust in automated driving technology play a significant role in this aspect. By demonstrating that OEMs have strong safety processes in place during the design, development and verification can help in gaining public trust.
Various approaches are being employed to demonstrate the safety of automated technology. Some companies focus exclusively on safety assessment based on quantitative analysis. Others are committed to safety by design, i.e., standards based safety processes and implementation. A combination of both strategies provides a comprehensive approach. By understanding the overall system design, safety assessment can target expected weaknesses and ensure that all potential risks are evaluated. The integration of AI in automated driving technology has made safety assessment even more critical, marking another key development in the field.
AI components, often perceived as black boxes, require specific safety assessment methodologies due to their complexity. Research into these methodologies is essential to evaluate the contribution of AI components to the overall safety of the system. This dual focus on system-level and component-level safety assessments ensures a thorough evaluation of automated driving technology.
Automated driving technology is progressing steadily and significant advancements in making safety risks quantifiable are required to gain public trust.
