CoChat

Poka Yoke Meets Deep Learning: A Proof of Concept for an Assembly Line Application

Open access
Martinelli, Matteo, Lippi, Marco, Gamberini, Rita · 2022 · Applied Sciences

The DOI matches its Crossref record; 36 of the 36 references it cites were found in a registry or an index; 2 of the 11 sources its claims cite were checked for support.

Abstract

In this paper, we present the re-engineering process of an assembly line... The re-engineering process identified a critical problem in the incorrect assembly of the oil seals... Convolutional neural networks were used to address this issue. The proposed solution resulted to be a Poka Yoke. The whole re-engineering process induced a productivity increase that is estimated from 46% to 80%.

Overview

CoChat Summary

## TL;DR

The paper presents a Lean Manufacturing re-engineering of an assembly line for speed increasers/multipliers and a proof-of-concept CNN-based Poka Yoke for detecting incorrectly mounted oil seals. The re-design is reported to raise line productivity from 46% to 80%, while the CNN achieved 98.8% average micro-accuracy in 5-fold cross-validation.

## Key findings

  • The as-is assembly line had a cycle time of 658.8 s/pc, with 47% non-value added operations, 26% semi-value added operations, and 27% value added operations.
  • The line re-design reduced the estimated cycle time to 354.2 s/pc; the final balanced assembly setup is reported as 186.4 s/pc per operator phase, and the paper states a saving of 286.0 s/pc and a 1,789 h total saving in the current year.
  • For the oil-seal control task, a CNN trained on 500 full-HD images and evaluated with 5-fold cross-validation achieved 98.8% average micro-accuracy, ranging from 97.0% to 99.1%.
  • The confusion matrix shows the CNN never classified a wrongly positioned oil seal as correctly positioned; the reported counts are: Correct 146/0/4, Reverse 0/150/0, Absent 2/0/199.
  • The productivity of the line is reported to increase from 46% to 80%; the paper also reports 864 h saved in the logistics department and programming office.
  • The proposed project’s total expected cost is reported as 27,320 EUR, compared with an average expected crankcase/lid redesign cost of 72,000 EUR.

## Methods

  • The paper uses a case-study re-engineering design of an assembly line, starting from an as-is analysis (layout, material flow, spaghetti chart, Muda analysis, and constraints) and designing a to-be process under Lean Manufacturing principles.
  • For the Poka Yoke proof of concept, it formulates oil-seal verification as a 3-class image classification task: correct position, upside-down, or absent.
  • The model is a CNN with one convolutional layer (32 filters, 3×3), max pooling, a dense layer (64 neurons), dropout (0.5), and a 3-neuron output layer; training used Adam, categorical cross-entropy, mini-batch size 32, early stopping, and 5-fold cross-validation on 500 images.

## Limitations

  • The paper itself notes that the paper’s focus is on the production phase, so the reported savings in the logistics area are not considered in the subsequent discussion; this is a scope limitation stated in the text.
  • The dataset is small, with only 500 images, and the authors explicitly note that this is a “small amount of images,” which makes the result encouraging but indicates limited evidence.
  • Evident from the text: the Poka Yoke is a proof of concept for a specific speed increaser/oil-seal setup, so generalization to other assemblies is suggested but not demonstrated with external validation in the paper.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/app122111071

Cites

36 of 36 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

2 Valid · 6 Not valid · 7 Not checked · 100 Own results and interpretations

52 more claims

Create a free account to see all 115 claims and to verify the 2 references still open.

See all claims — free account
  • Probably not a substantial claim. 42
  • It may state two separate claims as one. 28
  • The quote only partly states this claim. 2

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 36 references

Research works printed but not linked: 2

  1. Shingo, S. (1986). Zero Quality Control: Source Inspection and the Poka-Yoke System, Productivity Press.
  2. Goodfellow, I., Bengio, Y., and Courville, A. (2015). Deep Learning, MIT Press.

Cited by

See all 27

Metadata

Authors
Martinelli, M., Lippi, M., & Gamberini, R.
Affiliations
University of Modena and Reggio Emilia (IT)
Corresponding author
Martinelli, Matteo
Venue
Applied Sciences
Publisher
MDPI AG
Published
2022-11-01
Volume
12
Issue
21
Pages
11071
ISSN
2076-3417 (ISSN-L)
OpenAlex
W4309091299
Cited by
27
References
36
Citations by year
2026: 9 · 2025: 8 · 2024: 4 · 2023: 5 · 2022: 1
FWCI
2.06
Type
journal article
Language
en
Open access
Locations
  • Applied Sciences (publishedVersion, cc-by)
  • Research Square (acceptedVersion, cc-by)
  • IRIS UNIMORE (University of Modena and Reggio Emilia) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Florence Research (University of Florence) (submittedVersion)
Topics
  • Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Digital Transformation in Industry
  • Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Manufacturing Process and Optimization
  • Social Sciences › Business, Management and Accounting › Strategy and Management › Management and Optimization Techniques
Keywords
poka-yoke, lean manufacturing, convolutional neural networks, shop floor layout, human factors, smart manufacturing, productivity improvement

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Poka Yoke Meets Deep Learning: A Proof of Concept for an Assembly Line Application. https://app.cochat.ai/papers/10.3390/app122111071
The paper
Martinelli, M., Lippi, M., & Gamberini, R. (2022). Poka Yoke Meets Deep Learning: A Proof of Concept for an Assembly Line Application. Applied Sciences. https://doi.org/10.3390/app122111071

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.