Optical Communication Device Anomaly Detection and Data Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optical communication systems lack flexibility in monitoring the transmission path, leading to resource wastage due to the output of all waveform data for analysis, and are prone to false-positive alerts from path fluctuations without operational problems.
Innovation Solution
A communication device with a mode changeover device for learning and monitoring modes, using machine learning to determine anomaly criteria, extracting only relevant data for external analysis, and suppressing excessive alerts by learning the normal state of the network including fluctuations within an allowable range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all waveform data is output for external analysis, then comprehensive monitoring is achieved, but resource wastage occurs
Solution Approach 1:
The patent extracts only the necessary waveform data for analysis by using an anomaly detector to identify abnormal segments in the transmission path. The data writer then outputs only these specific anomaly-related waveform data to external devices, rather than outputting all waveform data. This extraction approach maintains monitoring comprehensiveness while significantly reducing resource wastage.
2Reliability
If traditional monitoring methods are used, then all path fluctuations are detected, but false-positive alerts increase
Solution Approach 1:
The patent applies preliminary action by performing a learning mode before actual monitoring, where the system learns the normal state of the transmission path and stores it as reference data. During operation, this pre-acquired knowledge is used to compare against current waveform data, enabling the anomaly detector to distinguish between normal fluctuations and actual anomalies, thereby reducing false-positive alerts.
Solution Approach 2:
The system implements feedback by continuously comparing current waveform data against the learned normal state and using the anomaly detector to identify deviations. The data writer provides feedback output of anomaly-related data to external devices, which can further analyze and refine the monitoring process, improving the reliability of anomaly detection while filtering out false positives.
3Loss of information
If detailed analysis of all waveform data is performed, then comprehensive insights are obtained, but processing time increases
Solution Approach 1:
The patent extracts only the relevant waveform data segments that contain anomalies using the anomaly detector. The data writer then outputs only these extracted anomaly-related data portions to external devices for detailed analysis. This approach maintains analysis comprehensiveness for actual problems while dramatically reducing processing time by excluding normal, non-anomalous data from the analysis pipeline.
Data Source
AI summary
A communication device used in an optical communication system, the communication device includes a mode change over device configured to switch between a learning mode for learning a normal state of an optical transmission path before operation and a monitoring mode for monitoring a state of the optical transmission path during operation, an anomaly detector configured to detect an anomaly of the optical transmission path using a prediction model determined by the learning mode when the monitoring mode is selected, and a data writer configured to extract waveform data including information related to the anomaly to output the extracted waveform data to an outside when the anomaly is detected.