Quantitative analysis of multi-factor coupled safety risks for automated driving: Empirical evidence from 232 accident cases.

Abstract

Most automated driving accidents are caused by the coupling of multiple factors rather than individual triggers. Accordingly, clarifying risk coupling mechanisms and reconciling innovation with safety has become an urgent requirement for the industry. Based on 232 automated driving accident cases featuring the interplay of internal and external factors, this study constructs a safety risk factor framework including, personal factors, traditional hardware, automated driving technology, management, and environmental conditions. With the composite N-K coupling model, this study analyzes the interactive effects of safety risks in automated driving, and further quantifies the risk coupling values of secondary risk factors in risk formation. The results show obvious synergistic characteristics of multi-factor risks, with the coupling of automated driving technology, management, and environment ranking as the most frequently observed high-order pattern in the sample. Specifically, the dominant human risks are insufficient driver monitoring and attention failure; the primary technical risks are algorithmic and data deficiencies; the key management weakness lies in gaps in institutional and standard frameworks; and the main environmental factors are severe weather and visual/optical interference. These findings provide empirical evidence for targeted risk prevention and control in automated driving, and offer practical implications for vehicle R&D optimization, policy-making, and road operation management.