Operation Analysis Clustering for Reference Motion Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for analyzing worker operations require significant time and skilled human resources to extract reference operation information, making it difficult to efficiently identify appropriate worker operations.

Innovation Solution

An operation analysis device that classifies time series information into clusters based on similarity, selects clusters meeting specific conditions, and extracts time series information indicating a reference operation, using hierarchical clustering and dynamic time warping to facilitate efficient extraction of standard operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If reference operation information is extracted by an expert from accumulated operation information, then the accuracy of reference operation information is improved, but the time required and human resource requirements increase significantly

Engineering Contradiction:
Improveaccuracy of reference operation informationVSAvoidtime required for extraction
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-service by automatically extracting reference operation information from accumulated operation data through clustering algorithms, eliminating the need for expert manual extraction while maintaining high accuracy through data-driven identification of standard operation patterns

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical process of expert manual extraction with an automated computational system using clustering algorithms and similarity calculations, substituting human expertise with algorithmic processing to reduce time and resource requirements

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If reference operation information is extracted by an expert from accumulated operation information, then the quality of reference operation information is improved, but the complexity of the extraction process increases

Engineering Contradiction:
Improvequality of reference operation informationVSAvoidcomplexity of extraction process
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system automatically performs the extraction process through self-service mechanisms using clustering algorithms that identify standard operation patterns from accumulated data, eliminating the need for complex expert-driven processes while maintaining information quality

Inventive Principle:
Principle #25Self-service

3Measurement precision

If hierarchical clustering is used to classify time series information, then the accuracy of clustering is improved, but the computational time increases

Engineering Contradiction:
Improveaccuracy of clusteringVSAvoidcomputational time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-calculating similarity between time series information before clustering, and uses predetermined thresholds to guide the hierarchical clustering process, reducing computational time while maintaining clustering accuracy through pre-prepared data structures

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3916654B1Operation analysis device, operation analysis method, operation analysis program, and operation analysis system
Publication Date: 2025.12.03 OMRON CORP
  • EP3916654B1 patent drawingFigure 1
  • EP3916654B1 patent drawingFigure 2
  • EP3916654B1 patent drawingFigure 3

AI summary

Provided is an operation analysis device or the like which enables the extraction of information that indicates a reference operation from among pieces of accumulated information for indicating operations with ease. This operation analysis device is provided with: an acquisition unit 11 which acquires a plurality of pieces of time series information that indicate a prescribed operation recorded when the operation is performed one or a plurality of times by one or a plurality of workers; a calculation unit 14 which calculates similarities between the plurality of pieces of time series information; a classification unit 15 which classifies, on the basis of the similarities, the plurality of times series information into a plurality of clusters; a selection unit 16 which selects one or a plurality of clusters that satisfy a prescribed condition among the plurality of clusters; and a first extraction unit 17 which extracts time series information that indicates a reference operation from the selected one or plurality of clusters.