Autonomous Driving Safety Control for Quantified SOTIF Risk
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods lack a feasible quantitative analysis for safety of the intended functionality (SOTIF)-related residual risks in autonomous driving assistance systems, making it difficult to implement SOTIF in actual project development processes.
Innovation Solution
A safety control method for autonomous driving assistance systems that involves receiving a driver status to determine reasonably foreseeable indirect misuse (RFIM) events, receiving system events or faults, and calculating a failure rate related to RFIM events based on the severity of these events. This method allows for determining whether SOTIF-related residual risks are acceptable and can adjust human-machine interaction processes or system reliability accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If quantitative analysis methods are applied to SOTIF-related residual risks, then the reliability and safety assessment capability is improved, but the system complexity and implementation difficulty increase
Solution Approach 1:
The patent segments the SOTIF risk assessment into distinct components: driver status monitoring (detecting inattentiveness, drowsiness, absence), RFIM event identification, system event/fault detection, and failure rate calculation. Each component is independently implemented and then integrated, making the complex assessment process manageable and implementable in actual project development
Solution Approach 2:
The patent introduces a failure rate calculation model as an intermediary that bridges the gap between qualitative SOTIF concepts and quantitative risk assessment. The model uses the formula λ = riskfactor / RFIM_TTI to transform qualitative risk factors and time intervals into a quantitative failure rate metric, enabling objective safety assessment without requiring complex direct measurement systems
2Reliability
If driver status monitoring and RFIM event detection are continuously performed, then the safety of the intended functionality is improved, but the computational load and processing time increase
Solution Approach 1:
The system performs preliminary classification of driver status conditions (inattentive, drowsy, absent) and pre-identifies potential RFIM events before actual hazards occur. By detecting and flagging these conditions in advance, the system prepares for potential safety issues without requiring continuous full-scale analysis, reducing real-time computational burden while maintaining high safety monitoring capability
Data Source
AI summary
A safety control method for an autonomous driving assistance system includes: receiving a status signal regarding a driver so as to determine a reasonably foreseeable indirect misuse (RFIM) event; receiving a particular system event and/or system fault; and calculating, with reference to a degree of severity of the particular system event and/or system fault, a failure rate related to the reasonably foreseeable indirect misuse (RFIM) event, wherein it can be determined, on the basis of the failure rate, whether a safety of the intended functionality (SOTIF)-related residual risk in the autonomous driving assistance system is acceptable.