The Expert Locomotive Engineer’s Mental Model
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 mental models of expert locomotive engineers to inform the design of automated control systems for freight rail. Researchers from General Electric (GE) Research and the Massachusetts Institute of Technology (MIT) aimed to create a man-machine collaborative interface that leverages expert knowledge while utilizing machine precision for improved safety and efficiency. The research was motivated by the need for automated systems to be comprehensible and intuitive, requiring a shared understanding between the operator and the automation. Specifically, the study sought to identify external environmental cues and control strategies that constitute an engineer’s mental model, ensuring that future automation reflects human intentions and goals. Experiments were conducted from October 2019 to July 2020 at the Federal Railroad Administration’s Cab Technology Integration Lab (CTIL). The study paired five expert freight engineers with novice subjects who had minimal rail knowledge. Participants drove two routes in a simulator under two conditions: novice-at-the-controls (NAC) and expert-at-the-controls (EAC). To force reliance on communication and reveal cognitive processes, the driver’s view of the external environment was blocked, requiring oral interaction with the non-driving participant. Interactions were recorded and coded using linguistic markup, distinguishing between syntactic (Level 1) and semantic (Level 2) information. Sub-codes identified specific domain content, such as signals, grades, and speed restrictions. The analysis revealed that expert engineers primarily function as situation assessors rather than continuous planners. In the EAC scenario, "check precondition" (CKP) interactions accounted for 68% of semantic codes, indicating that experts follow predetermined plans triggered by specific environmental cues. The most critical real-time updates involved mileposts, signals, grades, and speed restrictions. In the NAC scenario, CKP interactions comprised 28% of semantic codes, while "execute command" (EXC) accounted for 55%, reflecting the novice’s reliance on expert instruction. The study also found significant variability in individual driving strategies, particularly regarding braking profiles and the use of dynamic versus air brakes, suggesting that engineers internalize different goals and preferences. The findings imply that effective shared control systems must facilitate re-planning based on preconditions and allow engineers to adjust driving profiles to match their preferences. The researchers concluded that train handling relies on identifying preconditions that trigger action sequences, a process fundamental to maintaining situational awareness. These insights were used to prototype a shared speed control system that conveys information, supports re-planning, and summarizes plan goals (e.g., "save fuel"). Future work will test how these transparency and flexibility features affect trust, usage, and performance in both expert and novice operators.
Key finding
Expert locomotive engineers primarily operate as situation assessors who rely on checking preconditions to trigger predefined action plans, with high-level decisions generally planned in advance.
Methodology
simulator
Sample size: 5
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 | — | — | 19 | 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).
- Theoretical Contribution: computational model, theory or model