Scenario-Based ADAS & AV Safety Simulation

SelfDrivingSim automates the discovery of unknown-unsafe operational conditions (Area 3 to Area 4 transitions) in autonomous vehicles. Powered by an OpenSCENARIO 2 constraint-solver engine, it fuzzes edge cases across millions of parameter variations to certify autonomous driving stacks under ISO 21448 guidelines.

Tackling the Unknown-Unsafe Scenarios in Autonomous Mobility

Unlike ISO 26262 functional safety which addresses random hardware faults, ISO 21448 SOTIF targets hazards resulting from functional insufficiencies or environmental conditions in fault-free systems. SelfDrivingSim systematic boundary hunting mutates lighting, trajectory geometries, and occlusions to discover algorithmic failure envelopes.

Core Engineering Areas

• ISO 21448 SOTIF Verification — Executing automated test matrices that deliberately probe trigger conditions for performance limitations in perception sensors and motion planners.
• Combinatorial Boundary Fuzzing — Applying genetic algorithms and Bayesian optimization to discover the exact combinations of cut-in distances and decelerations that defeat AEB controllers.
• OpenSCENARIO 2.0 Execution Runtime — Executing declarative, intent-based driving scenarios with dynamic actors capable of reactive behavior based on the autonomous vehicle's actions.

Technical Articles

• ISO 21448 SOTIF Simulation Frameworks: Beyond ISO 26262
• OpenSCENARIO 2: Declarative Safety DSL for Edge Case Generation

Related Cluster Architecture Specifications

• AutomotiveSDV — Full SDV Stack & Reference Architecture
• A2HCard — Agent-to-Human (A2H) Interface Cards & In-Cabin Autonomous Interaction Manifests
• A2HumanCard — Standardized In-Vehicle AI Agent Capability Discovery, Trust & Authorization Cards
• AIDrivenSynBio — AI-Driven Synthetic Biology for Automotive In-Cabin Air Purification, Bio-Engineered Pathogen Scavengers & Molecular Biosensors