Safety-Critical Learning for Long-Tail Events: The TUM Traffic Accident Dataset

TL;DR

This paper introduces the TUM Traffic Accident dataset, a large collection of real-world highway accident data, and proposes Accid3nD, a hybrid accident detection model demonstrating robustness through experiments.

Contribution

The paper provides a new comprehensive traffic accident dataset and a novel hybrid accident detection model combining rule-based and learning-based methods.

Findings

Accid3nD outperforms baseline models in accident detection accuracy.

The dataset includes detailed 2D and 3D annotations for diverse traffic scenarios.

Experiments show the robustness of the proposed method across different accident types.

Abstract

Even though a significant amount of work has been done to increase the safety of transportation networks, accidents still occur regularly. They must be understood as an unavoidable and sporadic outcome of traffic networks. We present the TUM Traffic Accident (TUMTraf-A) dataset, a collection of real-world highway accidents. It contains ten sequences of vehicle crashes at high-speed driving with 294,924 labeled 2D and 93,012 labeled 3D boxes and track IDs within 48,144 labeled frames recorded from four roadside cameras and LiDARs at 10 Hz. The dataset contains ten object classes and is provided in the OpenLABEL format. We propose Accid3nD, an accident detection model that combines a rule-based approach with a learning-based one. Experiments and ablation studies on our dataset show the robustness of our proposed method. The dataset, model, and code are available on our project website: https://tum-traffic-dataset.github.io/tumtraf-a.