Phase Space Dissimilarity Forewarning for Critical Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for forewarning critical events such as epileptic seizures and machine failures lack robustness in detecting condition changes, particularly in providing timely and accurate warnings using nonlinear analysis techniques.
Innovation Solution
The method employs phase space dissimilarity measures, computing a composite dissimilarity measure from normalized measures like χt2, χe2, Lt, and Le, and using Shannon entropy to indicate forewarning and failure onset by analyzing time-series data, providing visual or audible signals to observers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional nonlinear methods are used to detect and predict epileptic seizures, then the detection capability is limited, but the method complexity remains low
Solution Approach 1:
The patent combines multiple normalized measures of dissimilarity (U(χC2), U(LC), U(χN2), U(LN)) from both connected and non-connected phase spaces into a single composite measure C. This merging of multiple indicators into one comprehensive metric enhances detection accuracy while managing complexity through integration rather than separate analysis of each measure.
Solution Approach 2:
The composite measure C functions as a composite indicator that integrates information from multiple phase space dissimilarity measures. By creating this composite metric, the system achieves superior detection capability compared to individual measures, analogous to how composite materials combine properties of constituent materials to achieve enhanced performance.
2Measurement precision
If multiple normalized measures of dissimilarity are computed from phase space analysis, then the forewarning accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent divides the analysis into distinct segments: connected phase space measures (U(χC2), U(LC)) and non-connected phase space measures (U(χN2), U(LN)). Each segment is computed and normalized separately, then combined into the composite measure C. This segmentation allows systematic computation of multiple measures while organizing the computational process to manage complexity.
Solution Approach 2:
The patent applies normalization transformations to convert raw dissimilarity measures into standardized forms (U(χC2), U(LC), U(χN2), U(LN)). This parameter transformation enables direct comparison and combination of measures from different phase space analyses, improving forewarning accuracy while maintaining computational tractability through standardized processing.
3Loss of time
If a composite measure of dissimilarity is used to provide forewarning indications, then the timeliness of warning is improved, but the threshold determination becomes more complex
Solution Approach 1:
The patent employs feedback mechanisms where the composite measure C is continuously monitored against predetermined thresholds. When C exceeds the threshold, a forewarning indication is generated. This feedback loop enables timely detection of condition changes while using adaptive thresholding strategies to manage the complexity of determining appropriate warning levels.
Solution Approach 2:
The system performs preliminary analysis by computing the composite measure C from multiple phase space dissimilarity measures before a critical event occurs. By continuously evaluating C and comparing it to thresholds in advance, the system provides timely forewarning indications, allowing preliminary detection of deteriorating conditions before failure or critical events occur.
Data Source
AI summary
This invention teaches further improvements in methods for forewarning of critical events via phase-space dissimilarity analysis of data from biomedical equipment, mechanical devices, and other physical processes. One improvement involves objective determination of a forewarning threshold (UFW), together with a failure-onset threshold (UFAIL) corresponding to a normalized value of a composite measure (C) of dissimilarity; and providing a visual or audible indication to a human observer of failure forewarning and/or failure onset. Another improvement relates to symbolization of the data according the binary numbers representing the slope between adjacent data points. Another improvement relates to adding measures of dissimilarity based on state-to-state dynamical changes of the system. And still another improvement relates to using a Shannon entropy as the measure of condition change in lieu of a connected or unconnected phase space.