Pattern-Section State Detection in Noisy Measurement Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting device states from measurement data are prone to errors due to noise and inaccuracies, leading to incorrect identification of patterns, which can result in missed or false alarms, particularly in critical applications like air travel where precise automatic state detection is essential.
Innovation Solution
A system that acquires flawed measurement data and uses an analysis unit to compare independently two sections of a predetermined pattern, determining their positions and sequence within the data to accurately detect the pattern, thereby reducing incorrect detections and ensuring precise device state assessment, utilizing techniques like dynamic time warping and multiple searches with varying framework conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pattern recognition is performed on flawed measurement data, then device state detection is achieved, but detection accuracy deteriorates due to noise and measurement inaccuracies
Solution Approach 1:
The predetermined pattern is divided into multiple pattern sections that are searched for independently in the measurement data. This segmentation allows each section to be evaluated separately, reducing the impact of noise on the overall pattern detection and improving reliability in flawed measurement data.
2Device complexity
If the entire predetermined pattern is searched for as a single unit, then pattern detection is simplified, but detection accuracy deteriorates due to noise causing false positives or missed detections
Solution Approach 1:
The pattern is segmented into multiple sections that are searched independently. While this increases computational steps, it significantly improves detection reliability by allowing the system to tolerate some noise in individual sections while still identifying the complete pattern when all sections are found in the correct sequence.
Solution Approach 2:
The system uses the positions and sequences of detected pattern sections as feedback to confirm the presence of the complete predetermined pattern. By verifying that multiple sections appear in the expected sequence, the system reduces false positives and improves detection reliability.
3Measurement precision
If multiple pattern sections are searched independently, then pattern detection accuracy improves, but computational complexity increases
Solution Approach 1:
The measurement data is processed by dividing the pattern into searchable sections, which enables more precise position detection. The analysis unit manages this complexity through systematic comparison of sections against the predetermined pattern, achieving higher precision while maintaining manageable computational requirements.
Data Source
AI summary
The invention relates to a system (1) for the automatic detection of a state of a device (2), comprising: a signal acquisition means (10) for acquiring defective measurement data (D) of a physical variable characterising the device (2); and an analysis unit (11) for identifying a specified pattern (M) in the measurement data (D) acquired by the signal acquisition means (10). The analysis unit (11) is designed:— to compare at least two different pattern sections (M1-M14) of the specified pattern (M) separately from each other with the measurement data (D):— on the basis of the respective comparison, to determine at least one position of each of the pattern sections (M1-M14) in the measurement data (D):— on the basis of the positions determined and the order of the positions of the pattern sections (M1-M14), to detect the specified pattern (M) at one or more positions in the measurement data (D) and,— on the basis of the one or more positions of the specified pattern (M), to determine the state of the device (2).