Transition Path Analysis for Abnormal Sign Detection in Conveyor Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conveyor apparatuses face challenges in detecting abnormal signs, particularly in distinguishing complex processing issues from actual abnormalities, as existing methods either focus on limited causes or use black-box machine learning models, making it difficult to present the grounds for detection results.
Innovation Solution
An information processing system that generates and corrects transition paths to build an abnormal sign detection model, using a candidate generating unit and correcting unit to identify deviating paths from normal patterns, and calculates a sign score to quantify abnormality, providing clear grounds for detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used for abnormal sign detection, then detection accuracy is improved, but interpretability of detection results deteriorates
Solution Approach 1:
The patent introduces transition paths as an intermediary representation that bridges the gap between complex machine learning detection processes and human-understandable explanations. The system generates transition paths that show the sequence of state changes leading to abnormal detection, making the black-box model's reasoning visible and interpretable while maintaining high detection accuracy
Solution Approach 2:
The patent replaces traditional rule-based detection mechanisms with machine learning models that analyze operation logs and generate transition paths. This substitution enables more accurate detection of complex abnormalities while the generated transition paths provide the interpretability that rule-based systems naturally offer
2Device complexity
If abnormal sign detection focuses on limited causes, then detection simplicity is improved, but detection comprehensiveness deteriorates
Solution Approach 1:
The patent creates a universal detection framework that can handle multiple types of abnormalities through a single system. The transition path generation mechanism works across different abnormal causes (roller degradation, conveyance interruptions, processing complexities) without requiring separate detection methods for each cause, thereby achieving both simplicity and comprehensiveness
3Measurement precision
If complex processing is not distinguished from actual abnormalities, then false positive rate is improved, but abnormality detection accuracy deteriorates
Solution Approach 1:
The patent segments the analysis into distinct transition paths that separate normal complex processing from actual abnormalities. By breaking down operation sequences into discrete state transitions, the system can identify and distinguish between expected complex operations and genuine abnormal conditions, reducing false positives while maintaining detection accuracy
Data Source
AI summary
An information processing apparatus includes a candidate generating unit and a correcting unit. The candidate generating unit generates, using first log information indicating states of a monitoring target and acquired in a certain time period, a frequently appearing transition path of states of the monitoring target, as a candidate of a first path that is a transition path presumed to represent a first specific condition. The correcting unit obtains the first path by correcting the candidate using a predetermined reference path.