Dual-Sensor Learning for Early Workpiece Abnormality Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor systems fail to detect abnormalities in workpieces transported on production lines in a timely manner, leading to potential defects in manufacturing.
Innovation Solution
A sensor system comprising a first sensor with a shorter measurement cycle and a second sensor with a longer measurement cycle, where data from the first sensor is used as input data and data from the second sensor as label data to generate learning data for a machine learning model, enabling early detection of abnormalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data from multiple sensors with different measurement cycles is collected in the control device, then communication speed is improved, but the ability to detect abnormalities early is insufficient
Solution Approach 1:
The invention performs preliminary actions by generating learning data in advance that correlates first sensor data (shorter cycle) with second sensor data (longer cycle). The machine learning model is trained beforehand to predict second sensor measurements from first sensor measurements, enabling early abnormality detection before the second sensor would normally measure, thus resolving the contradiction between communication speed and detection timeliness.
2Measurement precision
If a machine learning model is trained to predict workpiece properties using first sensor data, then abnormality detection accuracy is improved, but data processing complexity increases
Solution Approach 1:
The invention introduces learning data as an intermediary that bridges the first sensor data and second sensor data. By pre-processing and correlating these data types into learning data during a training phase, the system creates a intermediary dataset that simplifies the prediction process. The machine learning model uses this pre-correlated learning data to make accurate predictions without requiring complex real-time processing of multiple raw sensor streams, thus resolving the contradiction between prediction accuracy and processing complexity.
Data Source
AI summary
The present invention can detect early an abnormality or signs of abnormality in a workpiece. A sensor system 1 is provided with: a first sensor 30a that measures a workpiece; a second sensor 30b that measures the workpiece in a relatively longer cycle than the first sensor 30a; and a master unit 10. The master unit 10 includes: an acquisition unit 11 that acquires data measured by the first sensor 30a and data measured by the second sensor 30b; and a generation unit 12 that generates learning data which is used for machine learning of a learning model and in which the acquired data of the first sensor 30a is regarded as input data and the acquired data of the second sensor 30b is regarded as label data indicating a property of the input data.