ITSC 2025 Paper Abstract

Paper WE-EA-T12.1

Dat, Le (Technical University of Munich), Thomas, Manhardt (Cariad SE), Venator, Moritz (CARIAD SE), Betz, Johannes (Technical University of Munich)

Unsupervised Learning for Detection of Rare Driving Scenarios

Scheduled for presentation during the Regular Session "S12b-Safety and Risk Assessment for Autonomous Driving Systems" (WE-EA-T12), Wednesday, November 19, 2025, 13:30−13:50, Broadbeach 3

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 19, 2025

Keywords Autonomous Vehicle Safety and Performance Testing, Validation of Cooperative Driving and Connected Vehicle Systems, Verification of Autonomous Vehicle Sensor Systems in Real-world Scenarios

Abstract

The detection of rare and hazardous driving scenarios is a critical challenge for ensuring the safety and reliability of autonomous systems. This research explores an unsupervised learning framework for detecting rare and extreme driving scenarios using naturalistic driving data (NDD). We leverage the recently proposed Deep Isolation Forest (DIF), an anomaly detection algorithm that combines neural network-based feature representations with Isolation Forests (IFs), to identify non-linear and complex anomalies. Data from perception modules, capturing vehicle dynamics and environmental conditions, is preprocessed into structured statistical features extracted from sliding windows. The framework incorporates t-distributed stochastic neighbor embedding (t-SNE) for dimensionality reduction and visualization, enabling better interpretability of detected anomalies. Evaluation is conducted using a proxy ground truth, combining quantitative metrics with qualitative video frame inspection. Our results demonstrate that the proposed approach effectively identifies rare and hazardous driving scenarios, providing a scalable solution for anomaly detection in autonomous driving systems. Given the study's methodology, it was unavoidable to depend on proxy ground truth and manually defined feature combinations, which do not encompass the full range of real-world driving anomalies or their nuanced contextual dependencies.