Automatic Generation of PLC Control Code from Natural Language Requirement Specifications
Bibliographic record
Abstract
Developing control programs for manufacturing systems is time-consuming and requires expert control designers.While manual programming is common, it becomes complex as systems grow, leading to long development times, frequent errors, and difficult maintenance.To address these issues, researchers have introduced formal methods like Supervisory Control Theory (SCT) and model checking to improve precision and verification.Although these are some of the most advanced approaches, they are difficult to use in practice because they are time-consuming, require high mathematical expertise, and face scalability problems such as combinatorial explosion in large systems.This study aims to overcome these limitations by presenting an AI-based system that automatically generates programmable logic controller (PLC) code from natural language requirement specifications.The approach uses AutoFactory, a dataset of annotated specifications, and fine-tunes two Bidirectional Encoder Representations from Transformers (BERT)-based models to extract actuators, pre-actuators, and sensors before generating International Electrotechnical Commission (IEC) 61131-3 Structured Text (ST) code.BERT-Base achieved an F1 score of 0.9711, showing reliable component extraction.The study proves that transformer models can accurately detect control components and initiate logic generation.These results confirm that AI can assist and augment control designers by automating extraction and initial coding.Future work will complete the pipeline to deliver verified IEC 61131-3 code ready for industrial deployment.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".