CN-122001463-A - Method for identifying rogue ONU, storage medium and program product

CN122001463ACN 122001463 ACN122001463 ACN 122001463ACN-122001463-A

Abstract

The application discloses a method for identifying rogue ONUs, a storage medium and a program product, and belongs to the technical field of optical communication. The method for identifying the rogue ONU comprises the steps of collecting time sequence data to be detected under the condition that an Optical Line Terminal (OLT) receives an abnormal optical signal, inputting the time sequence data to be detected into a preset time sequence classification model, and determining the category of the time sequence data to be detected, wherein the preset time sequence classification model is trained by taking first time sequence data and second time sequence data as model training samples, the first time sequence data is the time sequence data, received by the target OLT, of each ONU to be detected under the condition of service conflict, the second time sequence data is the time sequence data, received by the target OLT, of each ONU to be detected under the condition of normal service, and determining the abnormal ONU from the ONU to be detected based on the category of the time sequence data to be detected.

Inventors

TANG LEI
YANG JIANBO
ZHANG WEILIANG
HE HANDONG
XIE CHENG

Assignees

中兴通讯股份有限公司

Dates

Publication Date: 20260508
Application Date: 20241108

Claims (12)

1. A method for identifying a rogue ONU, comprising: under the condition that an Optical Line Terminal (OLT) receives an abnormal optical signal, collecting time sequence data to be detected, wherein the time sequence data to be detected is the time sequence data sent by an ONU to be detected to a target OLT; Inputting the time sequence data to be detected into a preset time sequence classification model, and determining the category of the time sequence data to be detected, wherein the preset time sequence classification model is trained by taking first time sequence data and second time sequence data as model training samples, the first time sequence data is time sequence data, received by the target OLT, of each ONU to be detected under the condition of service conflict, and the second time sequence data is time sequence data, received by the target OLT, of each ONU to be detected under the condition of normal service; And determining an abnormal ONU from the ONU to be tested based on the category of the time sequence data to be tested.
2. The method of claim 1, wherein the training process of the predetermined time series classification model comprises: preprocessing model training samples to obtain target time sequence data preprocessed by the model training samples, wherein each target time sequence data is time sequence data with a bit stream rule; inputting each target time sequence data into an initial time sequence classification model for training to obtain the preset time sequence classification model.
3. The method of claim 2, wherein the preprocessing of the model training samples to obtain the preprocessed target time series data for each model training sample comprises: Carrying out subsection resampling processing on each model training sample to obtain preset time sequence data, wherein each time sequence data comprises a preset number of bit streams; determining bit streams corresponding to each piece of time sequence data, and dividing each piece of time sequence data into a plurality of categories of time sequence data based on the bit streams corresponding to each piece of time sequence data; And sequencing and concatenating the time sequence data of the plurality of categories according to the categories to obtain the target time sequence data.
4. A method according to claim 3, wherein before sorting and concatenating the time series data of the plurality of categories according to the categories to obtain the target time series data, the method further comprises: determining a time sequence average value of each section of time sequence data; and sequencing each section of time sequence data in the category according to the time sequence average value based on the time sequence average value of each section of time sequence data in the category aiming at the time sequence data in any one of a plurality of categories.
5. A method according to claim 3, wherein prior to said subjecting each model training sample to the piecewise resampling process, the method further comprises: and carrying out normalization processing on each model training sample to obtain normalized time sequence data.
6. The method according to claim 2, wherein the step of inputting each of the target time series data into an initial time series classification model for training to obtain the preset time series classification model includes: Inputting the time sequence data obtained by the i-th ONU to be tested after pretreatment under the condition of normal service and the time sequence data obtained by each ONU to be tested in the i-th ONU to be tested after pretreatment under the condition of service conflict with other ONUs to be tested into an i-th initial classifier for training to obtain an i-th classifier, wherein N is not more than 0<i, and both i and N are integers; inputting the time sequence data obtained by the N pre-processed ONU to be tested under the condition of normal service into an (n+1) th initial classifier for training to obtain an (n+1) th classifier; and taking the N ith classifiers and the (n+1) th classifier as the preset time sequence classification model.
7. The method of claim 6, wherein the inputting the time series data to be measured into a predetermined time series classification model, determining the category of the time series data to be measured, comprises: determining the acquisition time of the time sequence data to be detected, wherein the acquisition time comprises an idle time frequency domain and a non-idle time frequency domain; under the condition that the acquisition time is in an idle time-frequency domain, inputting the time sequence data to be detected into the (n+1) th classifier for classification, and obtaining the category of the time sequence data to be detected, wherein the ONU to be detected corresponding to the category of the time sequence data to be detected is an abnormal ONU; And under the condition that the acquisition time is a non-idle time frequency domain, inputting the time sequence data to be detected into a classifier corresponding to the ONU to be detected corresponding to the time sequence data to be detected, and classifying to obtain the category of the time sequence data to be detected, wherein under the condition that the category of the time sequence data to be detected comprises the ONU to be detected corresponding to the time sequence data to be detected and one other ONU to be detected label, the other ONU to be detected is an abnormal ONU.
8. The method according to claim 2, wherein the step of inputting each of the target time series data into an initial time series classification model for training to obtain the preset time series classification model includes: And inputting the time sequence data obtained by the N pre-processed ONU to be tested under the condition of normal service and the time sequence data obtained by each two ONU to be tested in the N pre-processed ONU to be tested under the condition of service conflict into an initial classifier for training to obtain a target classifier, and taking the target classifier as the preset time sequence classification model.
9. The method of claim 8, wherein the inputting the time series data to be measured into a predetermined time series classification model, determining the category of the time series data to be measured, comprises: determining the acquisition time of the time sequence data to be detected, wherein the acquisition time comprises an idle time frequency domain and a non-idle time frequency domain; Under the condition that the acquisition time is in an idle time-frequency domain, inputting the time sequence data to be detected into the target classifier for classification to obtain the category of the time sequence data to be detected, wherein the ONU to be detected corresponding to the category of the time sequence data to be detected is an abnormal ONU; And under the condition that the acquisition time is a non-idle time frequency domain, determining the ONU to be detected corresponding to the time sequence data to be detected, and inputting the time sequence data to be detected into the target classifier for classification to obtain the category of the time sequence data to be detected, wherein under the condition that the category of the time sequence data to be detected comprises the ONU to be detected corresponding to the time sequence data to be detected and one other ONU to be detected label, the other ONU to be detected is an abnormal ONU.
10. The method of claim 2, wherein prior to said inputting each of said target timing data into an initial timing classification model for training, said method further comprises: marking the target time sequence data corresponding to the first time sequence data to obtain a training sample label of the target time sequence data corresponding to the first time sequence data, wherein the training sample label comprises two ONU to be tested with business conflict; And marking the target time sequence data corresponding to the second time sequence data to obtain a training sample label of the target time sequence data corresponding to the second time sequence data, wherein the training sample label comprises an ONU to be tested with normal service.
11. A storage medium having stored thereon a program or instructions which when executed by a processor perform the steps of the method according to any of claims 1 to 10.
12. A program product, characterized in that it is stored in a storage medium, which program product is executed by at least one processor to carry out the steps of the method according to any one of claims 1 to 10.

Description

Method for identifying rogue ONU, storage medium and program product Technical Field The application belongs to the technical field of optical communication, and particularly relates to a rogue ONU identification method, a storage medium and a program product. Background A passive Optical network (Passive Optical Network, PON) system comprises an Optical line terminal (Optical LINE TERMINAL, OLT), an Optical distribution network (Optical Distribution Network, ODN) and an Optical network unit (Optical Network Unit, ONU). A plurality of ONUs are connected under one PON, and the plurality of ONUs share the same upstream receiving port. The rogue ONU refers to that, due to software defect or hardware aging failure of the ONU, an optical network unit is not started or closed according to a time slot allocated by an uplink bandwidth, or an uplink signal is transmitted on an incorrect wavelength, so that a time-frequency domain resource of a normal ONU for transmitting uplink data is occupied, and uplink transmission error codes of the normal ONU are caused. In the related art, the positioning of the rogue ONU generally adopts a manual approach, i.e. the rogue ONU is suspected to exist at a certain PON port, then the ONUs under the PON are turned off one by one through a manual plug/control command, and the received optical signal condition and the error condition of the OLT are observed, so as to complete the investigation of the rogue ONU. But this approach is time consuming and inefficient. Disclosure of Invention The application aims to provide a rogue ONU identification method, a storage medium and a program product, which at least solve the problems of long time consumption and low efficiency of manual investigation in the related technology. In a first aspect, an embodiment of the present application provides a method for identifying a rogue ONU, where the method includes collecting to-be-detected time sequence data when an OLT receives an abnormal optical signal, where the to-be-detected time sequence data is time sequence data sent by an ONU of an OLT, inputting the to-be-detected time sequence data into a preset time sequence classification model, and determining a class of the to-be-detected time sequence data, where the preset time sequence classification model is trained by using first time sequence data and second time sequence data as model training samples, the first time sequence data is time sequence data, received by the OLT, of each ONU to be detected under a service collision condition, the second time sequence data is time sequence data, received by the OLT, of each ONU to be detected under a normal service condition, and determining, based on the class of the to-be-detected time sequence data, an abnormal ONU from the ONUs to be detected. In a second aspect, an embodiment of the present application proposes a storage medium having stored thereon a program or instructions which, when executed by a processor, implement the steps of the method according to the first aspect. In a third aspect, embodiments of the present application provide a program product stored in a storage medium, the program product being executed by at least one processor to carry out the steps of the method according to the first aspect. In the embodiment of the application, firstly, under the condition that an Optical Line Terminal (OLT) receives an abnormal optical signal, time sequence data to be detected is collected, the time sequence data to be detected is time sequence data sent by an ONU of an optical network unit to be detected to a target OLT, then the time sequence data to be detected is input into a preset time sequence classification model, and the category of the time sequence data to be detected is determined, wherein the preset time sequence classification model is trained by taking first time sequence data and second time sequence data as model training samples, the first time sequence data is time sequence data, received by the target OLT, of each ONU to be detected under the condition of traffic collision, the second time sequence data is time sequence data, received by the target OLT, of each ONU to be detected under the condition of normal traffic, and finally, the abnormal ONU is determined from the ONU to be detected based on the category of the time sequence data to be detected. According to the embodiment of the application, the first time sequence data obtained by each ONU to be tested under the condition of service conflict and the second time sequence data obtained by each ONU to be tested under the condition of normal service are used as model training samples, the training sample categories comprise more complete categories, so that the precision of the rogue ONU is higher, and the time sequence to be tested, which is sent by the ONU to be tested to the target OLT, is input into a trained time sequence classification model for category identification, namely, the automatic identification is a