Скачать или смотреть Human Factors and Final Evaluation - Misuse Analysis and Acceptability Determination

This video provides a comprehensive technical overview of Clause 7 of the ISO 21448 (SOTIF) standard, focusing on the identification and evaluation of functional insufficiencies and triggering conditions.

In autonomous systems, safety isn't just about preventing malfunctions (ISO 26262); it's about managing the inherent limitations of sensors, algorithms, and AI when faced with complex real-world environments. We explore how to transition from "Unknown Unsafe" scenarios to "Known Safe" operations through rigorous analysis and validation.

Key Topics Covered:
Understanding Functional Insufficiency vs. System Failure
Physical limitations of Radar, LiDAR, and Camera sensors
AI/Machine Learning edge cases and perception challenges
The SOTIF risk evolution process (The 4-Quadrant Model)
Defining and determining "Acceptable Risk" for autonomous vehicles

Time Stamp:
00:52 Core concepts of SOTIF (ISO 21448)
01:28 Definition of functional insufficiency of autonomous driving systems
02:10 Physical limitations of sensors: Cases of radar ghost targets and false braking
02:33 Limitations of perception systems: Camera misreading and false detection phenomena
02:57 Limitations of LiDAR: Light-absorbing materials and detection gaps
03:21 Challenges of AI and machine learning: Unpredictable edge cases
04:02 Risk management process: Reducing the unknown unsafe area
04:41 Criteria for determining acceptable risk
05:32 Setting target failure rates and establishing statistical safety goals
06:14 Risk mitigation strategies and system validation loop
07:21 Future prospects for social acceptance of autonomous driving and safety standards

#SOTIF #ISO21448 #AutonomousDriving #SelfDrivingCars #FunctionalSafety #ADAS #Clause7 #TriggeringConditions #FunctionalInsufficiency #EdgeCases #AISafety #SystemsEngineering #AutomotiveEngineering #LiDAR #Radar #ComputerVision #AutoSysEng