Safety Challenge in Autonomous Driving – SOTIF [ISO/PAS 21448]
May 11, 2021
from 4:00 PM
to 4:45 PM
Today, autonomous driving is the main protagonist in the automotive industry environment. There are many applications, like AI and Machine Learning that have a key role in the development of self-driving vehicle systems. Even if these applications are developing rapidly, they are facing a safety challenge. Verifying AI and Machine Learning can be difficult because of their complexity. Risk avoidance due to system malfunctions is not a challenge anymore. The new challenge is to ensure the intended functionality in all the known and unknown real scenarios, assuming the absence of system malfunctions. The right way to manage these types of hazard events is given by the SOTIF Standard ISO/PAS 21448. These standards provide the guidelines to design (like a requirement for sensor performance), verify (test cases with high coverage of scenarios), and measure (simulation) the systems properly.
Functional safety standards like ISO 26262 are now inconsistent to validate and verify self-driving systems. These systems implement AI and Machine Learning technologies that continuously make decisions to avoid potential safety hazards. Verifying and validating these decision algorithms is difficult because of their complexity and the huge volume of data management. So, we need to understand the SOTIF standard to properly develop these systems and to avoid Foreseeable Misuse or Functional Insufficiencies.
The goals of this webinar are:
Identify the purpose and scope of SOTIF;
Describe the clauses;
Understand the SOTIF role in the Safety Life-Cycle and the key aspect of SOTIF;
Describe the Foreseeable Misuse Analysis;
Understand the Functional Insufficiencies Analysis, the related Safety Goals and the Functional Safety Concept;
Describe the Verification and Validation Strategies.
Reply's Webinars series covers all industries and spans the latest innovation trends.