News

From "Compliance Check" to "Safety Case Argumentation": Setting the New Global Standard for Physical AI Safety at HKU
Date:2026-09-18 15:10:25 | Page view:

At the landmark HKU LAE x SCC Joint Symposium, Digauto CEO Li Xiaoying delivered an authoritative keynote address entitled "Hardware-in-the-Loop (HIL) Simulation & Physical AI Testing," unveiling Digauto's cutting-edge validation framework before officiating leaders from the HKSAR Civil Aviation Department (CAD), academia, and global industry partners.

IMG_4882_副本.jpg

Addressing the convergence of the world’s three premier regulatory architectures—the UN WP.29 (2026), the US SELF DRIVE Act (2026), and China’s GB 44721 mandatory standard—Ms. Li emphasized that automated driving systems (ADS) and low-altitude economy (LAE) platforms are entering an era where static compliance is no longer sufficient. The future requires full lifecycle Safety Case Assessment.

Key insights from Ms. Li’s keynote:

  • Tackling the SOTIF Long-Tail Dilemma: Physical road testing cannot exhaust mathematically infinite, unknown unsafe edge cases. Through the SOTIF Index quantitative evaluation framework ($SOTIF_{index} = F_K \times F_{Uk}$), Digauto is systematically converting unknown hazardous scenarios into verified safe operational domains (ODD).

  • DigGenX Multimodal World Model & Closed Loop: Showcased the proprietary DigGenX multimodal platform—integrating DigGen-World (interactive world model) and DigGen-Scene (physics-anchored scenario generation) with high-fidelity Vehicle-Traffic-Environment Hardware-in-the-Loop (VTEHIL) to generate verified, high-value training and testing datasets.

  • Tricycle Safety Integration: Bridging ADS, LAE, and Embodied AI robotics. Quoting IDC’s user study highlighting that 61.5% of industrial users rank performance stability and safety reliability as their #1 priority, Ms. Li demonstrated how high-precision testing is the foundational gateway to commercialization.

IMG_4895_副本.jpg

Digauto is committed to serving as the Master Lab Architect and Land Application Lead for the proposed Physical AI Safety Lab (PAIS-Lab) in Hung Shui Kiu, Northern Metropolis, turning safety validation from a compliance burden into an exportable competitive advantage!


 
 Prev:The exhibition marks its 20th anniversary this year, and Digauto has been an exhibitor for nine consecutive years.
 Next:Forging a World-Class Physical AI Safety Consortium: Digauto Signs Strategic MOU at The University of Hong Kong!