Multi-Station Quality Monitoring for Historical Feature Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing classifiers in manufacturing processes only consider measurements from a single station and do not utilize the historical data from previous stations, limiting their ability to accurately classify articles of manufacture.
Innovation Solution
A classifier is trained using an aggregation of feature vectors from current measurements and encoded time series data representing historical measurements from previous stations, enabling a more comprehensive classification of articles by considering the entire manufacturing process history.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the classifier uses only measurements from a single station, then the device complexity is reduced, but the measurement precision deteriorates
Solution Approach 1:
The patent segments the manufacturing process data into distinct time series components from different stations, processing each station's measurements separately before aggregation. This allows the system to handle complex multi-station data without overwhelming computational burden, while still achieving improved classification accuracy through comprehensive historical context.
Solution Approach 2:
The patent transforms the classification problem from a single-dimensional (single station) approach to a multi-dimensional approach by incorporating time series data from multiple stations. This dimensional expansion adds historical context and process evolution information, significantly improving measurement precision while the systematic aggregation method manages the resulting complexity.
2Measurement precision
If the classifier incorporates historical data from multiple stations, then the measurement precision is improved, but the loss of time for data processing increases
Solution Approach 1:
The patent performs preliminary encoding of time series data from multiple stations before the actual classification process. By pre-processing and structuring the historical data in advance, the system reduces the computational burden during real-time classification, thereby improving accuracy without excessive time loss.
Solution Approach 2:
The patent implements continuous data flow from multiple stations into the classifier, maintaining an ongoing record of article measurements throughout the manufacturing process. This continuous action allows the system to accumulate useful historical data without interrupting the manufacturing flow, balancing processing time with improved precision.
3Reliability
If the classifier uses only current station measurements, then the productivity is maintained, but the reliability of classification deteriorates
Solution Approach 1:
The patent incorporates feedback from multiple stations by feeding historical measurement data back into the classification process. This feedback mechanism allows the classifier to learn from past measurements and improve reliability, while the structured aggregation method prevents excessive system complexity by organizing the feedback data systematically.
Data Source
AI summary
Methods and systems for classifying a article of manufacture are disclosed. A classifier is trained with training data including 1) a feature vector related to the article based on measurements related to the article captured at a particular station of a manufacturing process and 2) encoded time series data representing a history of measurements of articles of the same type as the article of manufacture captured at a sequence of stations of the manufacturing process prior to the particular station.