Fault detection and reporting in line monitoring systems
US10404362B1 · kind B1 · utility
Assignee
Inventors
Key dates
| Filing date | Mar 27, 2018 |
| Grant date | Sep 3, 2019 |
| Priority date | — |
| Expiry date | Mar 27, 2038 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04B10/25
- WIPO fieldTelecommunications
- WIPO sectorElectrical engineering
Abstract
In general, a system and method consistent with the present disclosure provides automated line monitoring using a machine learning fault classifier for determining whether a signature associated with the high loss loopback (HLLB) data matches a predetermined fault signature. The fault classifier may be applied to signatures generated in response to line monitoring signals of two different wavelengths. A fault may be reported only if the fault classifier indicates a fault in response to the signature for both wavelengths. A second fault classifier may also be used and a fault may be reported only if both the first and second fault classifiers indicate a fault in response to the signature for both wavelengths. A system consistent with the present disclosure may also, or alternatively, be configured to report the value of a pump degradation, span loss, or repeater failure fault, and may also, or alternatively, report the directionality of a span loss fault or the location of a fiber break fault.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.