Prediction of destination entry and retrieval times using keystroke-level models
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This study investigates the application of keystroke-level models (KLM), a variant of the GOMS framework, to predict the time required for drivers to enter and retrieve destinations in the Ali-Scout navigation system. The research was motivated by the need to evaluate the usability and safety of in-vehicle navigation interfaces without the high cost and time associated with full-scale experimental testing. Specifically, the authors sought to determine typical keystroke and mental operation times for a compact alphanumeric keyboard with poor tactile feedback and to assess how accurately KLMs could predict actual performance times compared to standard model parameters. The experimental design involved 36 licensed drivers divided into three age groups: young (18–30), middle-aged (40–55), and older (>65). Participants performed destination retrieval and entry tasks using both a real Siemens Ali-Scout unit and a touchscreen simulation. Retrieval tasks involved typing, scrolling, or a hybrid method to find stored destinations, while entry tasks required typing the destination name and coordinates. Data collection focused on interkeystroke intervals and mental operation times. To ensure validity for KLM assumptions, the analysis excluded trials with errors, initial keystrokes (which included mental planning time), and data from older subjects who exhibited inconsistent pausing, focusing primarily on error-free sequences from young subjects. The results revealed that standard KLM parameters significantly underestimated performance times. For young subjects, mean interkeystroke intervals ranged from 0.6 seconds for space keys to 1.7 seconds for initial cursor actions, whereas standard models assume much faster typing. Mental operation times averaged 2.2 seconds, substantially higher than the standard 1.35 seconds. Age had a profound impact, with middle-aged drivers taking 40% longer and older drivers taking 120% longer than young drivers. When linear equations based on pure standard KLMs were applied, they accounted for 41–78% of the variance in retrieval times and 12–39% in entry times. However, tailored models using experimentally derived keystroke values and adjustments for age and lighting improved predictive accuracy, accounting for 58–83% of retrieval variance and 47–49% of entry variance. The study concludes that while standard keystroke-level models are insufficient for predicting performance on complex automotive interfaces, tailored models significantly reduce prediction errors. These calibrated models provide a viable, efficient tool for engineering evaluations of alternative interface designs, allowing designers to estimate usability metrics during the development phase rather than relying solely on costly post-hoc experimental testing. The findings underscore the importance of context-specific parameter calibration, particularly regarding mental operation times and demographic factors like age, in human-computer interaction modeling for automotive applications.
Key finding
Tailored keystroke-level models accounted for 83 percent of the variance in retrieval times for young subjects and 58 percent for all subjects, significantly reducing prediction errors compared to standard models.
Methodology
lab_experiment
Sample size: 36
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 bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 24 | 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).
- Empirical Findings: behavioral performance data
- Theoretical Contribution: computational model