COTDR Event Statistic Generation for Intrusion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Intrusion detection systems using Coherent Optical Time Domain Reflectometer (COTDR) technology face challenges in distinguishing between nuisance alarms caused by environmental disturbances and actual intrusion events, particularly in outdoor environments with heavy wind, rain, or nearby traffic, as these systems often generate false alarms due to noise interference.
Innovation Solution
A computer-implemented method and apparatus that processes return signals from a COTDR to transform time-domain signals into frequency-domain signals, calculates signal power areas for noise and event-related frequencies, and generates event statistics by dividing the event signal power area by the noise signal power area, enabling effective separation of intrusion signals from system noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If COTDR technology is used to detect signals from buried optical fiber in outdoor environments, then the system can detect intrusion events, but it generates false alarms due to environmental noise such as wind, rain, and traffic
Solution Approach 1:
The patent segments the frequency spectrum into multiple frequency bins and identifies specific frequency bands characteristic of intrusion events versus environmental noise. By dividing the spectral analysis into discrete segments, the system can selectively process and compare different frequency components to distinguish genuine intrusions from nuisance alarms caused by wind, rain, or traffic.
Solution Approach 2:
The patent changes the parameter of frequency domain analysis by transforming time-domain signals into frequency-domain signals and calculating signal power areas in specific frequency bands. This parameter transformation enables the system to identify characteristic frequency signatures of intrusion events and differentiate them from environmental noise through ratio-based event statistics.
2Measurement precision
If the system processes all return signals to detect any events, then it increases detection sensitivity, but it also increases false alarm rate from nuisance events
Solution Approach 1:
The patent applies local quality by assigning different processing weights to different frequency bands based on their characteristic properties. Frequency bands known to contain intrusion-related energy are enhanced, while bands dominated by environmental noise are suppressed. This selective processing improves detection sensitivity for genuine intrusions while maintaining reliability by filtering out nuisance event frequencies.
Solution Approach 2:
The patent implements feedback through event statistic calculation that compares signal power in intrusion-characteristic frequency bands against noise baseline levels. The system uses this feedback ratio to dynamically assess whether detected events represent genuine intrusions or environmental disturbances, thereby reducing false alarms while maintaining high detection sensitivity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The method enhances the ability to discriminate between nuisance and intrusion events, improving the sensitivity and accuracy of intrusion detection systems by providing a robust means to identify genuine intrusions even when they are masked by noise, thereby reducing false alarms and enhancing detection capabilities.
Implementation Method 1
processing a plurality of return signals from a coherent optical time domain reflectometer into time domain signals
Implementation Method 2
transforming the respective time-domain signal into a corresponding frequency-domain signal
Data Source
AI summary
A computer-implemented event statistic generation method for intrusion detection comprises processing a plurality of return signals from a coherent optical time domain reflectometer into time domain signals for each of a plurality of sensor bins, the plurality of return signals corresponding to a plurality of stimulation pulses injected into an optical sensor fiber during a time period. For each sensor bin, the method comprises transforming the respective time-domain signal into a corresponding frequency-domain signal, calculating, from the respective frequency-domain signal, a first signal power area of a first frequency band expected to contain system noise, calculating, from the respective frequency-domain signal, a second signal power area of a second frequency band expected to contain any energy related to at least a first event; and generating an event statistic proportional to the ratio of the second signal power area to the first signal power area at least in part by dividing the second signal power area by the first signal power area.