CH-LSTTM: A Taxonomy of Traffic Hazards
DOI: 10.1109/ichms59971.2024.10555744
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the lack of a systematic framework for categorizing traffic hazards, which limits the effectiveness of driver training programs. While existing taxonomies classify hazards based on their characteristics (e.g., visibility or imminence), they often overlook the cognitive procedures and information requirements involved in hazard perception. The authors argue that a taxonomy grounded in the hazard perception process can better support the design of training materials that improve drivers’ ability to anticipate and respond to dangerous scenarios, particularly in novel situations. To bridge this gap, the authors propose the CH-LSTTM taxonomy, which categorizes hazards based on four elements derived from the hazard perception procedure: Cues, Hazards, Link Strength (LS), and Time to Materialization (TTM). Cues are classified by the number of road elements required to extract them: single-element-based (S-cues), multiple-element-based (M-cues), or non-element-based (N-cues, relying on the absence of expected elements). Hazards are categorized as visible or invisible. LS is a continuous variable representing the strength of the association between cues and the hazard, ranging from strong positive to strong negative. TTM indicates the time remaining until the hazard materializes, distinguishing between latent and imminent threats. The authors mathematically model the probability of a hazard using a sigmoid function of the summed LS values. The study validates the taxonomy through a case study using approximately 100 hours of naturalistic driving video data collected from autonomous heavy trucks in China. A researcher manually inspected the videos to extract hazardous scenarios. The authors identified 12 logically valid hazard types from the 36 possible combinations in the CH-LSTTM framework, providing specific real-world examples for each type. For instance, an S-cue with strong positive LS and latent TTM was exemplified by a vehicle’s turn signal indicating a likely lane change, while an M-cue with weak LS was illustrated by the relative motion of two vehicles where the outcome was uncertain. The significance of this work lies in its potential to optimize driver education and hazard perception training. By systematically categorizing hazards, the CH-LSTTM framework allows for targeted training that addresses specific cognitive challenges, such as recognizing M-cues or estimating LS, which are often experience-dependent. This approach may help novice drivers achieve hazard perception skills comparable to experienced drivers and reduce the decay of training effects over time. The taxonomy also provides a foundation for developing computational models of hazard perception and supports the generation of diverse, systematic training scenarios rather than relying on limited, ad-hoc collections of hazardous events.
Key finding
The proposed CH-LSTTM taxonomy successfully categorizes real-world traffic hazards into 12 distinct types based on cue structure, hazard visibility, link strength, and time to materialization.
Methodology
naturalistic
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-28 |
| archive | success | canonical_url | — | — | 1 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-07 |
| chunk | success | chunk | — | — | 1 | 2026-06-07 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-07 |
| enrich | success | semantic_scholar | — | — | 4 | 2026-06-15 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- hazard perception
- hazard perception training
- mental model of traffic
- anticipation
- situational awareness
- pre crash contributing factors
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Methodological Resource: dataset resource
- Theoretical Contribution: conceptual framework, theory or model