Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CRASH: Cognitive Reasoning Agent for Safety Hazards in Autonomous Driving

Erick Silva, Rehana Yasmin, Ali Shoker

Abstract

As AVs grow in complexity and diversity, identifying the root causes of operational failures has become increasingly complex. The heterogeneity of system architectures across manufacturers, ranging from end-to-end to modular designs, together with variations in algorithms and integration strategies, limits the standardization of incident investigations and hinders systematic safety analysis. This work examines real-world AV incidents reported in the NHTSA database. We curate a dataset of 2,168 cases reported between 2021 and 2025, representing more than 80 million miles driven. To process this data, we introduce CRASH, Cognitive Reasoning Agent for Safety Hazards, an LLM-based agent that automates reasoning over crash reports by leveraging both standardized fields and unstructured narrative descriptions. CRASH operates on a unified representation of each incident to generate concise summaries, attribute a primary cause, and assess whether the AV materially contributed to the event. Our findings show that (1) CRASH attributes 64% of incidents to perception or planning failures, underscoring the importance of reasoning-based analysis for accurate fault attribution; and (2) approximately 50% of reported incidents involve rear-end collisions, highlighting a persistent and unresolved challenge in autonomous driving deployment. We further validate CRASH with five domain experts, achieving 86% accuracy in attributing AV system failures. Overall, CRASH demonstrates strong potential as a scalable and interpretable tool for automated crash analysis, providing actionable insights to support safety research and the continued development of autonomous driving systems.

CRASH: Cognitive Reasoning Agent for Safety Hazards in Autonomous Driving

Abstract

As AVs grow in complexity and diversity, identifying the root causes of operational failures has become increasingly complex. The heterogeneity of system architectures across manufacturers, ranging from end-to-end to modular designs, together with variations in algorithms and integration strategies, limits the standardization of incident investigations and hinders systematic safety analysis. This work examines real-world AV incidents reported in the NHTSA database. We curate a dataset of 2,168 cases reported between 2021 and 2025, representing more than 80 million miles driven. To process this data, we introduce CRASH, Cognitive Reasoning Agent for Safety Hazards, an LLM-based agent that automates reasoning over crash reports by leveraging both standardized fields and unstructured narrative descriptions. CRASH operates on a unified representation of each incident to generate concise summaries, attribute a primary cause, and assess whether the AV materially contributed to the event. Our findings show that (1) CRASH attributes 64% of incidents to perception or planning failures, underscoring the importance of reasoning-based analysis for accurate fault attribution; and (2) approximately 50% of reported incidents involve rear-end collisions, highlighting a persistent and unresolved challenge in autonomous driving deployment. We further validate CRASH with five domain experts, achieving 86% accuracy in attributing AV system failures. Overall, CRASH demonstrates strong potential as a scalable and interpretable tool for automated crash analysis, providing actionable insights to support safety research and the continued development of autonomous driving systems.
Paper Structure (21 sections, 5 figures, 4 tables)

This paper contains 21 sections, 5 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Early AV crash analysis
  4. Qualitative analysis of reports
  5. Text classification and summarization using NLP
  6. CRASH: Cognitive Reasoning Agent for Safety Hazards
  7. Preprocessing
  8. Dataset and Filtering
  9. Processing
  10. LLM prompt construction
  11. CRASH Taxonomy of AV Incident Causes
  12. Choosing the LLM model
  13. Postprocessing and Human-in-the-Loop Validation
  14. Evaluation
  15. System Reliability
  16. ...and 6 more sections

Figures (5)

  • Figure 1: CRASH Architecture. During preprocessing, the database is filtered for incomplete entries and unified into four columns. Then, each row is sent to Processing through the LLM. We finally aggregate all information into a CSV file and send it to Postprocessing, where the data is used to generate our analysis and simulation descriptions.
  • Figure 2: System prompt for the CRASH Agent, incorporating heuristic rules and one-shot examples.
  • Figure 3: Compact taxonomy of AV incident causes.
  • Figure 4: Primary cause distribution and subsystem breakdown across 2,168 AV incidents.
  • Figure 5: Overlap between late AI behavior, AV failures, and rear-end collisions ($N{=}2{,}168$).