Sensor Output Change Detection via Sparse Error Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In complex systems with hundreds of sensors, such as automobiles, it is challenging to accurately identify abnormal sensors due to dynamic changes in sensor correlations and output values, leading to difficulties in distinguishing normal from abnormal sensors, especially when sensors have strong correlations.
Innovation Solution
A detection device that acquires data from sensors, generates a model using regularization to make errors sparse, and identifies abnormal sensors by reconstructing data in a low-dimensional latent space, thereby isolating changes and scoring sensors effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensors are increased to monitor complex systems, then measurement coverage is improved, but difficulty of detecting and measuring abnormalities increases
Solution Approach 1:
The patent segments the problem of detecting abnormalities in hundreds of sensors by dividing sensors into groups based on their correlation relationships. Instead of analyzing all sensors simultaneously, the system creates multiple sensor groups where sensors within each group have similar characteristics. This segmentation reduces the complexity of analysis and enables more effective abnormality detection in each subgroup.
Solution Approach 2:
The patent introduces a sensor group as an intermediary concept between individual sensors and the overall system. By analyzing sensor groups rather than individual sensors directly, the system creates an intermediate layer of analysis that simplifies the detection process. The sensor group acts as a mediator that aggregates sensor information while maintaining the ability to identify specific abnormal sensors within the group.
2Measurement precision
If regularization is applied to make errors sparse, then identification accuracy of abnormal sensors is improved, but device complexity increases
Solution Approach 1:
The patent applies parameter changes by modifying the error representation through regularization. Specifically, it uses L1 regularization (lasso regularization) to transform the error distribution into a sparse form where most errors are zero or near-zero, and only a few significant errors correspond to abnormal sensors. This parameter transformation enables accurate identification of abnormal sensors while the patent manages the complexity burden through efficient computational approaches.
Data Source
AI summary
A method includes acquiring a first data column output from a plurality of sensors, generating a model for estimating data from the plurality of sensors on the basis of the first data column, acquiring a second data column output from the plurality of sensors, obtaining an estimated data column corresponding to the second data column based on the model by using regularization for making an error between the second data column and the estimated data column sparse, and identifying a sensor in which a change occurred between the first data column and the second data column on the basis of the error between the second data column and the estimated data column. A corresponding computer program product and apparatus are also disclosed herein.