A Scalable Framework for Safety Assurance of Self-Driving Vehicles based on Assurance 2.0

2025-09-30

AI Summary

Traditional CAE models exhibit limitations in confidence quantification, residual doubt management, automation support, and mitigation of confirmation bias. Method: This paper proposes a scalable safety assurance framework for end-to-end AI-based autonomous driving systems, grounded in the Assurance 2.0 paradigm. It integrates multidimensional decomposition, a three-tier argument structure, and structured templates; incorporates reusable assurance theories and explicit defeaters; and extends the 5M1E model to enable fine-grained traceability and continuous validation of safety claims, evidence, and risks. Contribution/Results: It is the first work to systematically embed defeater mechanisms into AI autonomous driving safety arguments, enabling automated, transparent safety assurance across requirements engineering, verification & validation (V&V), and post-deployment phases. Empirical evaluation demonstrates significant improvements in evidence completeness, traceability, and residual risk control.

Technology Category

• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Knowledge Representation and Reasoning: Argumentation

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Assurance 2.0 is a modern framework developed to address the assurance challenges of increasingly complex, adaptive, and autonomous systems. Building on the traditional Claims-Argument-Evidence (CAE) model, it introduces reusable assurance theories and explicit counterarguments (defeaters) to enhance rigor, transparency, and adaptability. It supports continuous, incremental assurance, enabling innovation without compromising safety. However, limitations persist in confidence measurement, residual doubt management, automation support, and the practical handling of defeaters and confirmation bias. This paper presents extcolor{black}{a set of decomposition frameworks to identify a complete set of safety arguments and measure their corresponding evidence.} Grounded in the Assurance 2.0 paradigm, the framework is instantiated through a structured template and employs a three-tiered decomposition strategy. extcolor{black}{A case study regarding the application of the decomposition framework in the end-to-end (E2E) AI-based Self-Driving Vehicle (SDV) development is also presented in this paper.} At the top level, the SDV development is divided into three critical phases: Requirements Engineering (RE), Verification and Validation (VnV), and Post-Deployment (PD). Each phase is further decomposed according to its Product Development Lifecycle (PDLC). To ensure comprehensive coverage, each PDLC is analyzed using an adapted 5M1E model (Man, Machine, Method, Material, Measurement, and Environment). Originally developed for manufacturing quality control, the 5M1E model is reinterpreted and contextually mapped to the assurance domain. This enables a multi-dimensional decomposition that supports fine-grained traceability of safety claims, evidence, and potential defeaters.

Problem

• Developing scalable safety assurance frameworks for autonomous vehicles
• Addressing limitations in confidence measurement and residual doubt management
• Providing structured decomposition for safety arguments and evidence traceability

Innovation

• Decomposition frameworks for safety arguments and evidence
• Three-tiered strategy for self-driving vehicle development phases
• Adapted 5M1E model for multi-dimensional safety assurance