Towards a Human-Centred Cognitive Model of Visuospatial Complexity in Everyday Driving
DOI: 10.48550/arxiv.2006.00059
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 human-centred benchmarks and standardisation in autonomous driving research, which currently focuses primarily on vehicle control rather than complex human-machine interaction and visual sensemaking. Motivated by ethical and legal requirements for transparent, safe autonomous systems, the authors propose a cognitive model of visuospatial complexity in everyday, naturalistic driving conditions. The goal is to establish a framework for evaluating how environmental factors affect human visual attention and cognitive load, thereby guiding the creation of realistic benchmark datasets and explainable computational models for autonomous vehicles. The methodology combines insights from AI, cognitive psychology, and human-computer interaction to define a taxonomy of visuospatial complexity attributes. These attributes are categorised into three groups: quantitative (e.g., clutter, luminance, object density), structural (e.g., symmetry, order, regularity), and dynamic (e.g., motion, flicker, speed). To empirically evaluate this model, the researchers conducted a behavioural study using virtual reality (VR) simulations. They created a matrix of VR scenes varying in complexity levels by manipulating these attributes. Participants performed visual search tasks in four scenarios involving multimodal interactions, such as pedestrians crossing streets or cyclists moving. Data collection included eye-tracking metrics (e.g., fixation duration, saccade latency, scanpath length) and performance measures (e.g., reaction time, accuracy) to assess how different complexity levels influenced visual search efficiency and attentional patterns. The study outlines a framework for analysing the correlation between visuospatial attributes and human behavioural metrics. The authors hypothesise that quantitative attributes, particularly clutter, will negatively correlate with visual search performance, while structural attributes like order and symmetry may mitigate complexity effects. The VR experiments allow for the isolation of specific environmental factors to determine their individual and combined impact on cognitive processes. The paper reports preliminary steps in applying this model, including the development of a dynamic bubble diagram to visualise the relationships between complexity attributes and behavioural outcomes. It also details the specific metrics used to quantify search inefficiency, such as the number of fixations on distractors and the latency of the first saccade, which indicate confusion and uncertainty. The significance of this work lies in its contribution to human-centred AI and autonomous driving standardisation. By providing a empirically grounded model of visuospatial complexity, the authors offer a semantic template for computational analysis and a guideline for constructing benchmark datasets that reflect real-world cognitive challenges. This approach supports the development of autonomous systems capable of understanding complex, multimodal human interactions and varying environmental conditions. The model aims to bridge the gap between technical performance metrics and human-centred expectations, facilitating the ethical deployment and validation of autonomous vehicles in diverse, unstructured urban environments.
Key finding
The study proposes a cognitive model of visuospatial complexity based on quantitative, structural, and dynamic attributes, validated through virtual reality experiments to benchmark human visual search performance in driving scenarios.
Methodology
lab_experiment
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. Discovered via author_sweep_intake on 2026-05-29.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-29 |
| archive | success | canonical_url | — | — | 1 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-29 |
| 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.
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: tool software
- Theoretical Contribution: theory or model, computational model