Sensing, communication and computing integrated collaborative passive optical network system

Abstract

The invention discloses a sensing, communication and computing integrated collaborative passive optical network system, and belongs to the technical field of optical communication and optical networks. Comprising an optical fiber sensing unit, an optical fiber communication unit and an optical calculation unit based on artificial intelligence. The optical fiber sensing unit monitors, trains and measures the state of a physical layer in the life cycle of the optical network physical architecture, provides feedback for the optical computing unit, and is used for executing link diagnosis and fault detection tasks; the optical fiber communication unit comprises an optical coherent transceiver taking optical performance monitoring as a main body and an optical fiber transmission channel; the optical calculation unit forms an observation-judgment-execution control loop in a short observation window; according to the invention, the connection of the power divider branches of the ONU is tested and identified by using the OTDR, and the fault monitoring and positioning of the PON system are completed.

Description

Technical Field

[0001] The present invention belongs to the field of optical communication and optical network technology, and specifically relates to an integrated sensing, communication and computing collaborative passive optical network architecture. Background Art

[0002] With the emergence of emerging applications such as the Internet of Things (IoT) cloud computing, mobile X-haul services, virtual reality, smart cities, and 8K / 16K video streaming, massive video, etc., the Internet of Everything is leading the development of future communications. Faced with new demands and challenges, it is urgent to make some important changes to the existing optical information network. In order to meet the growing demand for communication and multimedia service traffic, fiber to the home (FTTH) has become the most widely used broadband fiber access network technology in the world. The growing perception of fiber infrastructure has provided possibilities for various new business applications and provided new added value for fiber access networks. Among them, in the fiber access network, the passive optical network (PON) is the FTTH solution that provides the "last mile" connection between the central office and the user due to its high bandwidth, low cost, flexibility and scalability. The infrastructure of a typical PON topology includes an optical line terminal (OLT), a feeder link, a 1:N power splitter, and an N-branch fiber distribution (DF) link, which connects the optical network unit (ONU) at the user end. Typical systems include Gigabit PON (G-PON), 10 Gigabit / s-PON (XGS-PON), 10Gb / s Ethernet PON (10G-EPON), and Next Generation PON (NG-PON2).

[0003] Due to the risks of NG-PON system optical fiber being cut, broken, cracked, aged, bent, man-made, or caused by earthwork by construction tools such as excavators. In addition, poor fiber installation is also a source of service interruption. Wildlife can also cause damage to underground and aerial optical cables. Service providers report that more than 80% of deployed PON failures occur within the last mile, that is, within the fiber distribution / access segment of the network. When a failure occurs, the traditional network operator's approach is to send technicians to identify, locate and solve the fault. The time, manpower and transportation used for fault identification significantly increase operating costs and user experience. In the fiber access network, the NG-PON status needs to be monitored and checked on demand regardless of whether the ONU is in use or connected. In addition to fault monitoring, inspection and location during normal network operation, PON also needs to be monitored during network installation and final testing. Therefore, quickly finding the cause of the interruption is critical to minimizing the impact of faults and failures.

[0004] The traditional typical NG-PON link monitoring scheme is a point-to-point scheme, that is, using optical time domain reflectometry (OTDR) equipment in the OLT to monitor the fiber infrastructure and quickly detect and locate faults such as fiber cuts. However, when applying OTDR to the PON monitoring system, the OTDR trace is usually identified by experienced field engineers relying on optical reflectometers for signal processing and analysis technology to identify the target reflection event. However, in the OLT, since the backscattered signals from each branch of the NG-PON system are superimposed together, if the NG-PON system branch lengths are the same or close, it is difficult to distinguish the backscattered signals of a single branch, thus limiting the applicability of the traditional OTDR event detection method. In addition, the remote node (RN) optical splitter has high loss, such as a 32-branch ratio optical splitter with a loss of 15dB, which will cause a significant decrease in the backscattered signal power. Therefore, the low signal power and the superposition of reflections from multiple optical fibers make fault detection and location of NG-PON difficult.

[0005] On the other hand, during the deployment of FTTH scenarios, the implementation and deployment of effective monitoring solutions will result in significant additional operating costs. In recent years, artificial intelligence (AI) has made significant progress in assisting monitoring and soft fault detection predictive maintenance, intrusion detection protection, and preventing malicious attacks and eavesdropping using automatic transceivers for network service optimization and configuration. Automated network characterization and monitoring solutions are essential for fast and cost-effective NG-PON system maintenance. Artificial intelligence based on machine learning (ML) refers to the overall set of algorithms learned from network data, the purpose of which is to model specific patterns and behaviors of the system to achieve predefined goals for a given communication service. AI can help NG-PON nodes learn to learn from the conditions of network failures and use this knowledge to optimize network resources. However, at present, neither academia nor industry has conducted research on the collaborative design and performance analysis of an integrated architecture of "sensing, communication and computing" that integrates optical fiber sensing, business communication and artificial intelligence computing capabilities in NG-PON systems. Therefore, the present invention forms a key role in self-configuration, self-optimization, self-protection and self-repair by sensing physical layer parameters through optical fiber sensing technology, transmitting through an optical fiber communication system and artificial intelligence computing. Summary of the invention

[0006] Aiming at the application requirement that NG-PON needs to have the ability to automatically and autonomously detect and locate network faults, the present invention proposes an integrated collaborative passive optical network architecture of "sensing, communication and computing" to realize the integration of sensing, communication and computing of the NG-PON system and solve the problem of cost-effective NG-PON operation.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is:

[0008] A sensing, communication and computing integrated collaborative passive optical network system, comprising an optical fiber sensing unit, an optical fiber communication unit and an optical computing unit based on artificial intelligence; wherein the optical fiber sensing unit and the optical fiber communication unit share a physical architecture entity based on a passive optical network;

[0009] The fiber optic sensing unit monitors, trains and measures the state of the physical layer during the life cycle of the optical network physical architecture and provides feedback to the optical computing unit for performing link diagnosis and fault detection tasks;

[0010] The optical fiber communication unit includes an optical coherent transceiver and an optical fiber transmission channel with optical performance monitoring as the main body;

[0011] The optical computing unit forms an observation-decision-execution control loop within a short observation window;

[0012] During the life cycle of the passive optical network, the data of the physical layer monitoring of the passive optical network is collected, including parameters such as the optical power of the upstream and downstream services sent and received by the optical line terminal and the optical network unit respectively; all the diagnostic data is then stored in the knowledge base in the passive optical network; an abstract mathematical model combined with artificial intelligence is used to access the specific system data of the passive optical network, detect and classify optical time domain reflection trace events with high precision; track, monitor and analyze the deployed passive optical network data, associate the fault events of the optical network with the branch topology of the optical network unit, and realize network monitoring and fault detection of the optical network unit in the passive optical network based on the probability of the associated events and the branch characteristics of the optical network unit of a single passive optical network.

[0013] Furthermore, in the optical fiber sensing unit:

[0014] The coherent optical transceiver is the main software body of the optical performance monitor; the optical time domain reflectometer deployed at the optical line terminal periodically tests the prior messages on the optical fiber and the passive optical network branch in the time dimension, as well as the data related to the optical fiber cut, reflection, open connection and aging effect and the passive optical network;

[0015] For the downstream direction of the passive optical network, the signal is transmitted continuously; the upstream direction is burst mode transmission, that is, each optical network unit has a window allocated by the optical line terminal, and transmits data to the optical line terminal within 125μs; when the distance from the optical network unit to the optical line terminal changes, the burst strength of the received signal in the optical line terminal is also different; therefore, at the hardware level, adjust the power of each burst output of the burst mode transimpedance amplifier; at the software level, update the physical media related to the passive optical network, transmission control and specific transmission information frames in the media access control.

[0016] Furthermore,

[0017] The optical fiber communication unit includes both ends of the transceiver and the optical fiber transmission process, and establishes an interface with the monitoring and positioning function module of the artificial intelligence computing system upward, and establishes an interface with the optical fiber sensing system based on optical time domain reflection downward;

[0018] Taking the digital signal processing composition of the coherent receiver of the optical line terminal of the uplink passive optical network as an example, the optical performance monitoring output of the optical fiber sensing system is integrated into the digital signal processing algorithm of the optical receiver with fiber dispersion, polarization mode dispersion, and nonlinear monitoring; the digital signal processing process of the signal received by the coherent receiver includes: clock recovery, fixed dispersion compensation, adaptive constant modulus algorithm and polarization demultiplexing, frequency offset and phase offset estimation recovery, and statistics of optical signal-to-noise ratio and bit error rate.

[0019] Furthermore,

[0020] The optical computing unit includes a cognitive loop self-configuring protocol and algorithm for observation, decision, and execution; the protocol and algorithm adjusts its structure, routines, state machines, and conditions according to application requirements and network infrastructure;

[0021] The optical computing unit first forms a basis for AI training based on the reference information about optical fiber bends, cracks, damaged or faulty connections contained in the knowledge base, namely the data of optical time domain reflectometry tracking tests obtained by the optical fiber sensing system; the optical computing unit model analyzes the data collected by the optical time domain reflectometry tracking in the knowledge base and the data of events deployed to specific power splitter branches or optical network units to analyze the probability of specific events; the optical computing unit model outputs the geographical location of the fault connecting the optical network unit in the passive optical network, thereby analyzing the individual optical fiber lines and identifying changes or fault events in the optical fiber infrastructure.

[0022] Compared with the prior art, the present invention has the following advantages:

[0023] 1. The present invention is aimed at the architecture and implementation of the integrated passive optical network of "sensing, communication and computing". First, the topology information of a specific PON system is obtained, that is, the deployment data of the ODN topology, including the feeder fiber segment, the branch ratio of the optical splitter and the length of the fiber distribution segment. Secondly, the network operation supports AI OTDR tracking and monitoring, and generates the probability that the fault event of the optical network is associated with the ODN branch, so as to realize the connection of the power splitter branch of the ONU by OTDR testing and complete the fault monitoring and positioning of the PON system.

[0024] 2. Aiming at the difficulty of fault management of superimposed signal tracking of N power splitter branches in NG-PON system, the present invention adopts AI method to classify the data information of each OTDR, establishes a connection between the reflection or attenuation events tracked by OTDR and event categories such as ODN position, and directly estimates the fiber transmission distance between ONU and OLT according to the corresponding delay changes of each ONU, thereby solving the specific fault monitoring and positioning difficulties in existing PON. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a functional block diagram of the system design of the "sensing, communication and computing" integrated passive optical network system of the present invention. DETAILED DESCRIPTION

[0026] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0027] A sensing, communication and computing integrated collaborative passive optical network architecture is applied to the elastic, flexible, open and reliable next-generation passive optical network. The basic idea is to collect PON physical layer monitoring data during the life cycle of PON, such as the optical power level parameters of the uplink and downlink services sent and received by OLT and ONU respectively; then all diagnostic data are stored in the knowledge base of PON; combined with an abstract mathematical model of artificial intelligence, PON-specific system data is accessed to detect and classify OTDR trace events with high precision; track, monitor and analyze the deployed PON data, associate the fault events of the optical network with the branch topology of the ODN, and realize ODN network monitoring and fault detection in PON according to the probability of associated events and the branch characteristics of the ODN of a single PON.

[0028] like Figure 1As shown. The integrated NG-PON of "Sense, Communication and Computing" includes three parts: optical fiber sensing unit, optical fiber communication unit and optical computing unit based on artificial intelligence. The optical fiber sensing unit of the integrated NG-PON of "Sense, Communication and Computing" monitors, trains and measures the status of the physical layer during the life cycle of the NG-PON physical architecture, and provides feedback to the AI-based intelligent computing system for performing tasks such as link diagnosis and fault detection. The optical fiber communication unit of the integrated NG-PON of "Sense, Communication and Computing" mainly includes an optical coherent transceiver and an optical fiber transmission channel with optical performance monitoring as the main body. The artificial intelligence optical computing unit of the integrated NG-PON of "Sense, Communication and Computing" forms an automated observation-judgment-execution control loop within a short observation window to achieve higher network efficiency and dynamics. The optical fiber sensing unit and optical fiber communication unit of the integrated NG-PON of "Sense, Communication and Computing" share the physical structure entity of NG-PON, that is, a point-to-multipoint topology architecture mainly composed of an OLT located at the central office (CO) connecting multiple ONUs through an ODN configured with different branching ratios. Among them, the branching ratio of the optical splitter used in the ODN is up to 64, the optical fiber is set with varying wavelengths, pulse durations and test time parameters, and the optical fiber transmission distance does not exceed 20 kilometers.

[0029] For the downstream direction of NG-PON, the signal is transmitted continuously, and the upstream direction is burst mode (BM) transmission, that is, each ONU has a window allocated by the OLT to transmit data to the OLT within 125s. In the optical fiber sensing unit, optical performance monitoring provides effective sensing for the physical layer of the "sensing, communication and computing" integrated passive optical network, enhancing the observation capability of the network. The coherent optical transceiver is the main body of the optical performance monitor. The OTDR deployed in the OLT periodically tests the prior information on the optical fiber and the PON splitter in the time dimension, as well as the data related to the PON such as optical fiber truncation, reflection, open connection and aging effect. When the distance from the ONU to the OLT changes, the burst strength of the received signal in the OLT is also different. Therefore, at the hardware level, it is necessary to adjust the power of each burst output of the burst mode (BM) transimpedance amplifier (BM-TIA). At the same time, at the software level, update the physical medium related (PMD), transmission control (TC) and media access control (MAC) of the NG-PON. Specific transmission information frames. The updated OTDR test data provides actionable intelligence for operational decisions, thereby optimizing real-world operational results and supporting closed-loop automation.

[0030] The optical fiber communication unit includes the operations at both ends of the transceiver and the optical fiber transmission process. It establishes an interface with the monitoring and positioning function modules of the artificial intelligence computing system, and establishes an interface with the optical fiber sensing system based on OTDR. Taking the OLT coherent receiver DSP composition of the upstream PON as an example, the optical performance monitoring output integrated in the optical fiber sensing unit is output to the receiver DSP algorithm with fiber dispersion, polarization mode dispersion, nonlinear monitoring, etc. The DSP process that the signal received by the coherent receiver passes through mainly includes: clock recovery, fixed dispersion compensation, adaptive constant modulus algorithm and polarization demultiplexing, frequency offset and phase offset estimation recovery, and statistics of optical signal-to-noise ratio (OSNR) and bit error rate.

[0031] The artificial intelligence optical computing unit mainly includes cognitive loop self-configuration protocols and algorithms for observation, judgment, and execution. These protocols and algorithms can adjust their structures, routines, state machines, and conditions according to application requirements and network infrastructure. The optical computing unit first forms the basis for AI training based on the reference information contained in the knowledge base about fiber bends, cracks, damaged or faulty connections, that is, the data of OTDR tracking tests obtained by the fiber optic sensing system. The artificial intelligence computing optical system model analyzes the data collected by the knowledge base OTDR tracking and the data of events deployed to specific power splitter branches or ONUs to analyze the probability of specific events. The model output locates the geographical location of the fault connecting the ONUs in the NG-PON, thereby analyzing the individual fiber lines and identifying changes or fault events in the fiber optic infrastructure.

[0032] The present invention targets the needs of an integrated passive optical network architecture of "sensing, communication and computing". This information is carried through a layered monitoring architecture, and the AI ​​engine decides how to act and re-execute the optimized network and arranges manual intervention when re-optimization cannot solve the problem. Short-term predictions can be used for active recovery to improve recovery efficiency. The invention is able to adapt, self-configure and self-optimize its structure and related intelligent hardware / software modules. Modulation format, bit rate, number of wavelengths used, transmission power, amplification gain, compensation and many others can be adjusted, reused and optimized according to the needs of infrastructure and services.

[0033] In summary, this embodiment implements an integrated NG-PON system of "sensing, communicating and calculating", and realizes a fault monitoring and positioning method based on different channel characteristics for different optical fiber branch links through a simple and easy-to-implement method. The "sensing, communicating and calculating" integrated NG-PON fault monitoring and positioning architecture can program the optical physical layer performance monitoring, select monitoring parameters, monitoring algorithms, cycles, etc., and can also efficiently transmit and analyze / correlate data from multiple optical performance detections. Therefore, the present invention plays the role of a global coordination entity and builds a bridge between monitoring and control functions.

[0034] The above description is only a specific implementation of the present invention in the embodiment, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with the field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A sensor-communication-computing integrated collaborative passive optical network system, characterized in that: It includes three parts: optical fiber sensing unit, optical fiber communication unit and optical computing unit based on artificial intelligence. Among them, the optical fiber sensing unit and optical fiber communication unit share the physical architecture entity based on passive optical network. The fiber optic sensing unit monitors, trains and measures the state of the physical layer during the life cycle of the optical network physical architecture and provides feedback to the optical computing unit for performing link diagnosis and fault detection tasks; The optical fiber communication unit includes an optical coherent transceiver and an optical fiber transmission channel with optical performance monitoring as the main body; The optical computing unit forms an observation-decision-execution control loop within a short observation window; During the life cycle of the passive optical network, the data of the physical layer monitoring of the passive optical network is collected, including parameters such as the optical power of the upstream and downstream services sent and received by the optical line terminal and the optical network unit respectively; all the diagnostic data is then stored in the knowledge base in the passive optical network; an abstract mathematical model combined with artificial intelligence is used to access the specific system data of the passive optical network, detect and classify optical time domain reflection trace events with high precision; track, monitor and analyze the deployed passive optical network data, associate the fault events of the optical network with the branch topology of the optical network unit, and realize network monitoring and fault detection of the optical network unit in the passive optical network based on the probability of the associated events and the branch characteristics of the optical network unit of a single passive optical network.

2. The sensing, communication and computing integrated collaborative passive optical network system according to claim 1 is characterized in that: In the fiber optic sensing unit: The coherent optical transceiver is the main software body of the optical performance monitor; the optical time domain reflectometer deployed at the optical line terminal periodically tests the prior messages on the optical fiber and the passive optical network branch in the time dimension, as well as the data related to the optical fiber cut, reflection, open connection and aging effect and the passive optical network; For the downstream direction of the passive optical network, the signal is transmitted continuously; the upstream direction is burst mode transmission, that is, each optical network unit has a window allocated by the optical line terminal to transmit data to the optical line terminal within 125μs; When the distance between the optical network unit and the optical line terminal changes, the burst strength of the received signal in the optical line terminal also varies; Therefore, at the hardware level, the power output of each burst of the burst mode transimpedance amplifier is adjusted; At the software level, update the physical media related to the passive optical network, transmission control and specific transmission information frames in medium access control.

3. The sensing, communication and computing integrated collaborative passive optical network system according to claim 1 is characterized in that: The optical fiber communication unit includes both ends of the transceiver and the optical fiber transmission process, and establishes an interface with the monitoring and positioning function module of the artificial intelligence computing system upward, and establishes an interface with the optical fiber sensing system based on optical time domain reflection downward; Taking the digital signal processing composition of the coherent receiver of the optical line terminal of the uplink passive optical network as an example, the optical performance monitoring output is integrated into the digital signal processing algorithm of the optical receiver with fiber dispersion, polarization mode dispersion and nonlinear monitoring in the optical fiber sensing system; The digital signal processing process of the signal received by the coherent receiver includes: clock recovery, fixed dispersion compensation, adaptive constant modulus algorithm and polarization demultiplexing, frequency offset and phase offset estimation and recovery, and statistics of optical signal-to-noise ratio and bit error rate.

4. The sensing, communication and computing integrated collaborative passive optical network system according to claim 1 is characterized in that: The optical computing unit includes a cognitive loop self-configuring protocol and algorithm for observation, decision, and execution; the protocol and algorithm adjusts its structure, routines, state machines, and conditions according to application requirements and network infrastructure; The optical computing unit first forms a basis for AI training based on the reference information about optical fiber bends, cracks, damaged or faulty connections contained in the knowledge base, namely the data of optical time domain reflectometry tracking tests obtained by the optical fiber sensing system; the optical computing unit model analyzes the data collected by the optical time domain reflectometry tracking in the knowledge base and the data of events deployed to specific power splitter branches or optical network units to analyze the probability of specific events; the optical computing unit model outputs the geographical location of the fault connecting the optical network unit in the passive optical network, thereby analyzing the individual optical fiber lines and identifying changes or fault events in the optical fiber infrastructure.

Images