Information Filter Thresholding for Reliable Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anomaly detection methods face challenges in determining an adequate threshold for magnetic anomaly detection due to the unknown optimal threshold value and its dependence on the sliding window and discretization step size, requiring a laborious trial-and-error process.
Innovation Solution
A method and device for anomaly detection that establishes a threshold based on the convergence limit value of an information filter, which intuitively determines the threshold as the entropy of the probability density plus a small adjustment value, simplifying the detection process and ensuring reliable anomaly identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional anomaly detection methods are used with sliding window and discretization, then anomaly detection capability is achieved, but threshold determination becomes complex and requires laborious trial-and-error processes
Solution Approach 1:
The system performs self-calibration by automatically determining the threshold value through information filter convergence analysis. The threshold is derived from the system's own operational data without requiring external calibration or manual trial-and-error processes, enabling the system to self-optimize its detection parameters
Solution Approach 2:
The method performs preliminary calibration during a normal operational phase before actual anomaly detection begins. By establishing the threshold value in advance through information filter convergence on normal data, the system prepares optimal detection parameters beforehand, eliminating the need for complex real-time threshold adjustment
2Measurement precision
If the sliding window size N is increased to improve detection reliability, then detection accuracy improves, but the threshold determination becomes more complex and tedious
Solution Approach 1:
The system uses feedback from the information filter's convergence behavior to automatically determine the threshold. By monitoring how the filter converges on normal data and using this feedback to set the threshold, the system eliminates manual adjustment regardless of window size, making the process equally simple for any N value
3Productivity
If the discretization step size Δ is increased to simplify processing, then computational complexity decreases, but the filter may react counterintuitively and anomalies may not be detected
Solution Approach 1:
The method changes the parameter used for threshold determination from fixed discretization-based values to dynamic values derived from information filter convergence. This allows the system to adapt to different discretization step sizes automatically, maintaining reliability regardless of the Δ value chosen for processing efficiency
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a device and a method for detecting anomalies within an environment or system, comprising the following steps: - acquiring (E1) a normal signal (NS) reflecting the normal state of the environment or system, - determining from the normal signal (NS) a probability density f that models a normal behavioral state of the environment or system, - establishing (E3) an information filter based on said probability density f, said information filter being configured to converge to a limit value L when applied to samples of a normal signal, while increasing its value in response to the detection of an anomaly, - establishing (E4) a threshold value S based on said convergence limit value, - acquiring (E5) a current signal (CS) reflecting the current behavioral state of the environment or system, - sampling (E6) said current signal (CS) in a current series of N samples,- calculate (E7) a result of applying the information filter to the current series of N samples, - compare (E8) said result with said threshold value S, an anomaly being detected if said result exceeds said threshold value.