Wireless broadband and narrowband network resource visualization and topology automatic discovery construction method

Abstract

The application discloses a wireless wide and narrow band network resource visualization and topology automatic discovery construction method, which is driven by a topology change event, carries out buffer difference and low-dimensional semantic coding on narrow band terminal messages, writes the change vector and time stamp into a digital twin state pool through an asynchronous publishing and subscribing mechanism, and subscribes, classifies and splices the whole network topology state from the wide band side. Further, the link connectivity confidence is inversely calculated according to the state update time, and the transparency of the topology link is controlled by the confidence gradient matrix to render. Thus, the redundant message transmission, terminal energy consumption and polling load can be reduced, the periodic sleep node is avoided from being misjudged as offline, the link discrete on-off is converted into a continuous gradual state, the hop and flicker caused by topology rearrangement are reduced, the automatic discovery, unified aggregation and smooth visualization of the heterogeneous wide and narrow band network topology are finally realized, and the accuracy and continuity of network state interpretation, fault positioning and operation and maintenance management are improved.

Description

Technical Field

[0001] This application relates to the field of network management, and more specifically, to a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topologies. Background Technology

[0002] With the development of new power systems and the Internet of Things for energy, broadband terminals for video surveillance and mobile inspection, as well as low-power narrowband terminals for temperature, partial discharge, and environmental monitoring, are being deployed simultaneously in substations, distribution stations, and other scenarios. These terminals differ significantly in communication standards, operating cycles, and mobility characteristics, and network connectivity is constantly changing. Therefore, it is necessary to construct a unified mechanism for visualizing wireless broadband and narrowband network resources and automatically discovering topologies to accurately present terminal access relationships and link status, providing a basis for fault location, resource scheduling, and operation and maintenance.

[0003] Existing technologies typically involve separate collection of device information by broadband and narrowband network management systems. Terminal online status is obtained through periodic polling, heartbeat detection, or packet sniffing, and a topology map is generated based on device identifiers and connection relationships. When a node is detected to be online, offline, or its parent node changes, the network management platform directly updates the link connectivity status and redraws the topology. While some solutions can aggregate heterogeneous network data to a unified interface, they still rely on original status packets and discrete online identifiers as the basis for topology construction. However, narrowband terminals often employ deep sleep and periodic wake-up mechanisms, while broadband terminals may be in a highly dynamic, mobile state, resulting in significant spatiotemporal asymmetry in status update times. Periodic polling not only increases the energy consumption and network load of narrowband terminals but also easily misjudges normal sleep as offline. Simply relying on the last packet or discrete connectivity status can lead to ghost nodes, broken links, and status lag. Furthermore, sudden changes in link status trigger frequent topology rearrangements and complete map redraws, causing node jumps, screen flickering, and difficulty in continuously reflecting the true evolution of the network.

[0004] Therefore, an optimized method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topologies is needed. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topologies.

[0006] According to one aspect of this application, a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topology is provided, comprising: S1: Strip the header of the physical control message of the narrowband terminal and compare it with the local historical cache. Only when the parent-child relationship of the node changes, perform low-dimensional semantic encoding to obtain the narrowband topology change semantic vector. S2: The narrowband topology change semantic vector is reported to the broadband and narrowband converged access gateway through the asynchronous publish-subscribe protocol. After decoding and verifying the narrowband topology change semantic vector, it is aggregated with the update timestamp and written into the digital twin state pool to obtain the node's last valid state entity. S3: The subscription channel pre-established by the broadband-side network management platform receives the last valid state entity of the node without polling, and performs topology classification and splicing according to the node identifier to obtain the whole network fused topology state set; S4: Based on the time difference between the current system time and the update timestamp, perform connectivity confidence inversion on each link in the whole network fusion topology state set to transform discrete connectivity states into continuous confidence scores and align them in parallel columns to obtain the topology connectivity confidence gradient matrix. S5: Construct a topology base map based on the set of fused topology states of the entire network, and use the gradient transparency of each link in the base map as the mapping parameter to obtain a smooth fused network topology view.

[0007] Compared with existing technologies, this application provides a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topologies. Driven by topology change events, it performs buffered differential encoding and low-dimensional semantic encoding on narrowband terminal packets. Through an asynchronous publish-subscribe mechanism, it writes change vectors and timestamps into a digital twin state pool. The broadband side then subscribes to, categorizes, and assembles the entire network topology state without polling. Furthermore, it inverts link connectivity confidence based on state update times and uses a confidence gradient matrix to control the transparency rendering of topology links. This reduces redundant packet transmission, terminal energy consumption, and polling load, avoids misjudging periodically dormant nodes as offline, and transforms discrete link connectivity into continuous, gradual states, reducing jumps and flickering caused by topology rearrangement. Ultimately, it achieves automatic discovery, unified aggregation, and smooth visualization of heterogeneous broadband and narrowband network topologies, improving the accuracy and continuity of network status judgment, fault location, and operation and maintenance management. Attached Figure Description

[0008] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 This is a flowchart of a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topology according to an embodiment of this application; Figure 2 This is a data flow diagram illustrating the wireless broadband and narrowband network resource visualization and automatic topology discovery construction method according to embodiments of this application; Figure 3 This is a flowchart of step S2 in the wireless broadband and narrowband network resource visualization and topology automatic discovery construction method according to an embodiment of this application; Figure 4 This is a flowchart of step S5 in the wireless broadband and narrowband network resource visualization and topology automatic discovery construction method according to an embodiment of this application. Detailed Implementation

[0010] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0011] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0012] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0013] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0014] The technical solution of this application proposes a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topologies. Figure 1 This is a flowchart of a method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topology according to an embodiment of this application. Figure 2 This is a system architecture diagram of the wireless broadband and narrowband network resource visualization and automatic topology discovery construction method according to an embodiment of this application. Figure 1 and Figure 2As shown, the wireless broadband and narrowband network resource visualization and automatic topology discovery construction method according to an embodiment of this application includes the following steps: S1, stripping the header of the physical control message of the narrowband terminal, and comparing it with the local historical cache, and only performing low-dimensional semantic encoding on it when the parent-child relationship of the node changes to obtain the narrowband topology change semantic vector; S2, reporting the narrowband topology change semantic vector to the broadband and narrowband converged access gateway through an asynchronous publish-subscribe protocol, and after decoding and verifying the narrowband topology change semantic vector, aggregating it with the update timestamp and writing it into the digital twin state pool to obtain the node's last valid state entity; S3, through the broadband side network management platform The pre-established subscription channel receives the last valid state entity of the node without polling, and performs topology classification and splicing according to the node identifier to obtain the whole network fusion topology state set; S4, based on the time difference between the current system time and the update timestamp, the connectivity confidence inversion of each link in the whole network fusion topology state set is performed to transform the discrete connectivity and disconnection states into continuous confidence scores and align them in parallel columns to obtain the topology connection confidence gradient matrix; S5, a topology base map is constructed based on the whole network fusion topology state set, and the gradient transparency rendering of each link in the base map is performed using the topology connection confidence gradient matrix as the mapping parameter to obtain a smooth fusion network topology view.

[0015] Specifically, in S1, the header of the physical control message of the narrowband terminal is stripped, and based on local historical cache comparison, low-dimensional semantic encoding is performed only when the parent-child relationship of the nodes changes to obtain the narrowband topology change semantic vector. It should be understood that in wireless broadband and narrowband converged communication networks, narrowband terminals (such as low-power wireless sensors based on the IEEE 802.15.4 protocol) are limited by battery power and low-power design, adopting an intermittent working mode of deep sleep and periodic wake-up. After each wake-up, the narrowband terminal sends a physical control message to maintain network connectivity. This message contains key information such as the terminal's own source media access control address, the media access control address of its parent node, the received signal strength indicator value, and the remaining battery power. However, during most wake-up cycles, the terminal's topology affiliation does not change. If the complete message generated by each wake-up is reported to the upper-layer gateway for processing, a large amount of redundant data transmission will be generated, which will not only increase the communication load of the narrowband channel but also consume the terminal's already limited battery power. Therefore, in the technical solution of this application, effective information is extracted by stripping the header of the physical control message, and the node parent-child relationship is determined based on the local historical cache comparison mechanism. Only when the topology relationship has changed is the effective information encoded with low-dimensional semantics to generate a compact semantic vector. This achieves data compression and event-driven processing at the source, significantly reducing invalid reporting, reducing terminal power consumption and channel occupancy, and providing efficient and accurate input for asynchronous reporting and digital twin state pool writing in subsequent steps.

[0016] The narrowband terminal physical control message is a low-level control frame sent by the narrowband terminal through a low-power wireless channel during each wake-up cycle. Its payload carries core fields characterizing the terminal's identity and its affiliation within the network topology. In the embodiments of this application, the narrowband terminal physical control message includes four core fields: source media access control address (MIC), which is the unique hardware identifier of the narrowband terminal itself, used to distinguish different terminal nodes in the network; parent node MIC, which is the hardware identifier of the upper-level aggregation node or relay node associated with the current terminal, reflecting the terminal's affiliation in a tree or star topology; received signal strength indicator (RSSI), which is the strength measurement of the signal received by the terminal from the parent node, reflecting the propagation quality of the current wireless link; and remaining battery power, which is the remaining percentage of the terminal's battery energy, used to assess the terminal's continuous operating capability.

[0017] In practice, firstly, the protocol parser extracts the effective payload of the narrowband terminal's physical control message to obtain the node state feature tuple. Specifically, when the narrowband terminal wakes up from deep sleep, its radio frequency module constructs and sends a physical control message according to a low-power wireless network communication protocol (such as the low-power wireless network communication protocol for power transmission and transformation equipment IoT based on IEEE 802.15.4). This message contains synchronization and delimitation fields such as a preamble sequence, start-of-frame delimiter, and physical layer header at the physical layer, and encapsulates protocol control fields such as frame control field, sequence number, address field, and frame check sequence at the MAC layer. The protocol parser, as a message processing module deployed locally at the aggregation node or converged access node, performs decapsulation operations sequentially from the outermost layer to the innermost layer of the protocol stack after receiving the complete physical control message. The protocol parser first identifies and strips the physical layer preamble and frame start delimiter to confirm the frame boundary. Then, it parses the frame length field in the physical layer header to determine the byte range of the valid data area. Next, it enters the MAC layer, determining the arrangement of subsequent address fields based on flags such as frame type and address mode in the frame control field, and strips non-service control fields such as the frame control field, sequence number, and frame check sequence. After this header stripping process, the protocol parser extracts four core fields from the remaining payload area, according to predefined field offsets and byte lengths: source media access control address (identifying the narrowband terminal sending the message), parent node media access control address (identifying the current upstream node associated with the terminal), received signal strength indicator (reflecting the signal quality of the wireless link between the terminal and the parent node), and remaining battery power (reflecting the terminal's current power supply status). The protocol parser encapsulates these four fields in a fixed order to obtain a node state feature tuple, formally represented as:

[0018] in, This is the source media access control address. This is the media access control address of the parent node. The received signal strength indication value, This represents the remaining battery percentage. The node's state feature tuple is the output of the message header stripping, serving as the input for subsequent topology state differential decisions.

[0019] Next, based on the historical parent node addresses in the local history cache, topology state differential decision is performed on the node state feature tuple to generate a topology change identifier. The local history cache is a key-value storage structure with the source media access control address as the key and the parent node media access control address recorded in the previous reporting period as the value, deployed in the local memory of the aggregation node or converged access node performing the resolution. Specifically, the topology state differential decision process is as follows: using the source media access control address in the current node state feature tuple... As the query key, the corresponding historical parent node address record for this terminal is retrieved from the local history cache, denoted as . Subsequently, the parent node's media access control address from the current node's state feature tuple is... The retrieved historical parent node address Perform byte-by-byte equality comparison. Topology change identifier. The generation logic is determined by the following decision function:

[0020] In this formula, when When this occurs, it indicates that the terminal's parent-child relationship in the network topology has changed within the current period, meaning the terminal has switched from its original parent node to a new parent node. This change needs to be reported to the upper-layer system to update the overall network topology status. When this occurs, it indicates that the terminal's topology affiliation remains unchanged, and no further semantic encoding and reporting processes are required. After completing the differential decision, regardless of the decision result, the system will store the parent node's media access control address in the current node's state feature tuple. Write to the local history cache to overwrite the original record, ensuring that the differential decision benchmark remains up-to-date for the next reporting period.

[0021] Furthermore, using the topology change identifier as a gating activation factor, low-dimensional semantic encoding is performed on the node state feature tuples to obtain a narrowband topology change semantic vector. Specifically, the topology change identifier, as a gating activation factor, controls whether the low-dimensional semantic encoding process is activated. When the topology change identifier... When the gating condition is not met, the encoding process is not activated, the system directly discards the current node's state feature tuple and does not generate any reported data, thus avoiding invalid encoding and transmission operations when the topology has not changed. When the topology change flag... When the gating condition is met, the encoding process is activated, and the system performs low-dimensional semantic encoding on the current node's state feature tuple.

[0022] In this process, each field in the node state feature tuple is mapped to a fixed-length numerical component according to a predefined encoding rule, and the components are concatenated to form a compact fixed-length vector. Specifically, the source media access control address... The node identifier is compressed into a fixed-width encoding component using a hash mapping function. This mapping maps the original 6-byte (48-bit) MAC address to a fixed-length identifier in a low-dimensional space, ensuring that the identifiers of different terminals are unique and distinguishable.

[0023] in, The capacity parameter of the encoding space is determined based on the maximum size of the terminal nodes in the network; For hash mapping functions, Identify the encoded components for the node. Parent node media access control address. Compressed into parent node identifier encoded components using the same hash mapping rules :

[0024] in, Identify the encoded component for the parent node.

[0025] Received signal strength indication value Link quality coding components are mapped to a closed interval between zero and one using linear normalization. :

[0026] in, For link quality coding components, and These are the minimum and maximum values ​​of the system's preset received signal strength indication, respectively. Remaining battery power. Similarly, it is mapped to terminal state coding components through normalization. :

[0027] in, This is the terminal state encoding component. Finally, these encoded components are concatenated and combined in a fixed order to form a narrowband topology change semantic vector. :

[0028] in, This is a semantic vector for narrowband topology changes.

[0029] This vector is a compact numerical representation with fixed dimensions. While preserving all semantic information about topology changes, such as node identification, topology affiliation, link quality, and terminal status, it significantly reduces data volume compared to the complete frame structure of the original physical control message. This makes it suitable for reporting via narrowband wireless channels with minimal transmission overhead. The gating mechanism ensures that encoding and subsequent reporting are triggered only when the topology actually changes, thus achieving event-driven information compression at the data source.

[0030] Specifically, in step S2, the narrowband topology change semantic vector is reported to the converged broadband and narrowband access gateway via an asynchronous publish-subscribe protocol. After decoding and verifying the narrowband topology change semantic vector, it is aggregated with the update timestamp and written into the digital twin state pool to obtain the node's last valid state entity. It should be understood that in step S1, the narrowband terminal has generated a narrowband topology change semantic vector that only occurs when the topology parent-child relationship changes, through header stripping, local historical cache differential comparison, and low-dimensional semantic encoding. However, this semantic vector still resides in the local storage of the narrowband terminal or its directly connected aggregation node and has not yet been transmitted to the converged broadband and narrowband access gateway responsible for network-wide state aggregation. Due to its low-power design and intermittent operating mode, the narrowband terminal has limited communication capabilities and an extremely short wake-up time window. If a traditional request-response synchronous communication method is used, the narrowband terminal needs to keep its radio frequency module on continuously after sending the semantic vector to wait for the gateway's confirmation reply. This will significantly increase terminal power consumption and prolong the wake-up time window, contradicting the low-power design goal. Meanwhile, the broadband-narrowband converged access gateway needs to simultaneously handle topology change data from a large number of narrowband terminals. If polling or synchronous requests are used one by one, severe concurrency bottlenecks and communication delays will occur. Therefore, in the technical solution of this application, an asynchronous publish-subscribe protocol is used as the transmission mechanism. This allows narrowband terminals to deliver semantic vectors to the message middleware in a "publish and release" manner, quickly returning to a sleep state without waiting for synchronous confirmation from the gateway, thus minimizing terminal power consumption. The broadband-narrowband converged access gateway then asynchronously receives messages as a subscriber, achieving complete decoupling of the sender and receiver in time and space.

[0031] Furthermore, narrowband topology change semantic vectors, after transmission through wireless channels, may experience bit flips or byte loss due to factors such as channel fading, electromagnetic interference, or packet collisions. If these vectors are used directly for topology state updates without integrity verification, erroneous topology information will be introduced, leading to a distorted network topology view. Simultaneously, because narrowband terminals operate in an asynchronous, intermittent mode, the state reporting times of different terminals vary significantly. The broadband / narrowband converged access gateway needs to bind a precise update timestamp to each valid topology change record to establish a temporal baseline for the state data, providing a time difference calculation basis for link connectivity confidence inversion in subsequent steps. Finally, the verified and timestamp-bound topology state data needs to be persistently stored in a structured manner in a digital twin state pool, forming the last valid state entity for each node. This serves as the authoritative data source for the entire network topology state, providing unified, reliable, and time-stamped basic data support for the subscription, consumption, and visualization of the upper-layer network management platform.

[0032] Figure 3 This is a flowchart of step S2 in the wireless broadband and narrowband network resource visualization and automatic topology discovery construction method according to an embodiment of this application. Figure 3 As shown, S2 includes: S21, deserializing and verifying the integrity of the narrowband topology change semantic vector reported to the broadband-narrowband converged access gateway via the asynchronous publish-subscribe protocol to obtain the verified topology semantic payload; S22, performing key-value pair orthogonal concatenation of the verified topology semantic payload and the update timestamp to obtain the time-bound topology state object; S23, using the source media access control address as the index key, overwriting the time-bound topology state object into the digital twin state pool to obtain the node's last valid state entity.

[0033] Specifically, in step S21, the narrowband topology change semantic vector reported to the broadband-narrowband converged access gateway via the asynchronous publish-subscribe protocol is deserialized and its integrity is verified to obtain the verified topology semantic payload. Here, it should be understood that after generating the narrowband topology change semantic vector in step S1, the narrowband terminal or its directly connected aggregation node needs to report this semantic vector to the broadband-narrowband converged access gateway via the asynchronous publish-subscribe protocol. In the embodiments of this application, the system introduces MQTT or Kafka as a message middleware. Instructions issued by the UCI and status reported by the adapter are transmitted through message queues, avoiding delays and packet loss in direct communication; simultaneously, it supports message QoS level settings (e.g., QoS 1: at least once) to ensure that high-priority instructions are not lost. During this process, the narrowband terminal, as the MQTT publisher, serializes the narrowband topology change semantic vector into a compact byte stream, attaches a CRC checksum, and publishes it to a predefined MQTT topic (e.g., "topo / change / {MAC_src}"). It then shuts down the radio frequency module and enters a deep sleep state, without waiting for a synchronous response from the gateway. After receiving the published message, the message broker routes it to the message receiving queue of the broadband and narrowband converged access gateway according to the topic subscription relationship.

[0034] The broadband / narrowband converged access gateway, acting as an MQTT subscriber, pre-subscribes to the wildcard topic "topo / change / #" to receive topology change messages from all narrowband terminals. When the gateway's message receiving thread retrieves a newly arrived message from its local message queue, it first performs a deserialization operation. This involves parsing the received byte stream field by field, restoring it to structured semantic vector data, according to a serialization format agreed upon with the sender (such as TLV format or a fixed-field-length binary encoding format). The deserialization process extracts the node identifier encoding component, parent node identifier encoding component, link quality encoding component, terminal status encoding component, and the CRC checksum field appended to the end of the byte stream, following a predefined field order.

[0035] After deserialization, the system immediately performs an integrity check on the data. Specifically, the CRC checksum is recalculated for each component field of the deserialized semantic vector using the same CRC generator polynomial as the sender, denoted as . The calculated results will then be used to... Compared with the raw CRC checksum extracted from the end of the byte stream Perform a bit-by-bit comparison. The integrity check decision logic is as follows:

[0036] When the verification result is When the narrowband topology change semantic vector remains intact after transmission through the wireless channel and forwarding by the message middleware, without bit flipping or byte loss, the deserialized and restored structured semantic vector data is marked as the verified topology semantic payload and passed to the next sub-step for processing. When the verification result is... When this occurs, it indicates that data corruption has occurred during transmission. The system discards the message and records the error in the error log, preventing it from proceeding to the next processing step. This prevents erroneous data from contaminating the topology state records in the digital twin state pool. The formal representation of the verified topology semantic payload is as follows:

[0037] in, The verified topological semantic payload contains components whose meanings completely correspond to the narrowband topology change semantic vector. The source node identifier component is encoded by a hash map. The parent node identifier component is encoded by a hash map. For the link quality components after normalized mapping, This represents the terminal power component after normalization mapping.

[0038] Specifically, in step S22, the verified topology semantic payload and the update timestamp are orthogonally concatenated using key-value pairs to obtain a time-bound topology state object. That is, after obtaining the verified topology semantic payload, the broadband-narrowband converged access gateway needs to bind a precise update timestamp to this valid topology change record to establish a positioning reference for the state data on the timeline. The update timestamp generation process is as follows: Upon confirming that the narrowband topology change semantic vector has passed integrity verification, the gateway immediately reads the current time value of its local system clock and formats it as a high-precision timestamp record. This timestamp should have millisecond-level or higher precision to ensure sufficient time resolution for connectivity confidence inversion based on the time difference between the current system time and the update timestamp in subsequent step S4. The formal representation of the update timestamp is:

[0039] in, To update the timestamp.

[0040] In this process, the verified topological semantic payload will be As a data component of the "spatial dimension", the timestamp will be updated. As the "time dimension" data component, the two are integrated into a unified structured data object through orthogonal combination. "Orthogonal" means that the two data dimensions are semantically independent and non-redundant—the topological semantic payload describes the position and connectivity of a node in the network topology space, while the update timestamp describes the valid starting point of this state information on the time axis. Together, they depict a complete snapshot of the node's state from different dimensions. The concatenation operation encapsulates the data from these two dimensions into a composite data structure in key-value pair form. Each component in the topological semantic payload is expanded into an independent key-value pair, and the update timestamp is appended as an additional time dimension key-value pair. The final formal representation of the time-bound topological state object is as follows:

[0041] Among them, the time-bound topology state object.

[0042] Specifically, in step S23, the time-bound topology state object is overwritten into the digital twin state pool using the source media access control address as the index key to obtain the node's last valid state entity. The digital twin state pool is a structured key-value database deployed in the local storage of the broadband / narrowband converged access gateway, using the source media access control address (corresponding to the node identifier encoded component in the time-bound topology state object). Its original value is the one extracted in step S1. The state pool maintains a state record for each narrowband terminal node in the network that has ever reported a topology change, and only one record is kept for the same node in the pool—that is, the most recent valid state of the node that has passed the integrity check.

[0043] During this process, the system binds topology state objects in a time sequence. Node identifier encoded component Using the index key, a query operation is performed in the digital twin state pool. If the record corresponding to the index key already exists (indicating that the terminal has previously reported a topology change), the current time series is bound to the topology state object. All fields of the node are overwritten into the record, replacing the original values ​​of parent node identifier, link quality, terminal battery level, and timestamp, ensuring that the node's mapping in the digital twin state pool always reflects its latest known good state. If the record corresponding to the index key does not exist (indicating that this is the first time the terminal has reported a topology change), a new record is created in the digital twin state pool with the index key, and the time sequence is bound to the topology state object. All fields are written to the newly created record. Specifically, the formal representation of the overwrite operation is:

[0044] After the overwrite operation, the digital twin state pool contains... The record corresponding to the index key is the last valid state entity of that terminal node, denoted as . :

[0045] in, This is the last valid state entity of the node.

[0046] Specifically, in step S3, the broadband-side network management platform receives the last valid state entity of the node without polling through a pre-established subscription channel, and performs topology classification and assembly according to the node identifier to obtain a network-wide converged topology state set. It should be understood that in step S2, the broadband-narrowband converged access gateway has persistently stored the topology change data, which has undergone integrity verification and timestamp binding, in the form of the last valid state entity of the node in the digital twin state pool. However, the digital twin state pool is a flat data set stored at the granularity of individual nodes and dispersedly using the source media access control address as the index key; it has not yet established a topology connection relationship map between nodes from a network-wide perspective. As the upper-layer system responsible for network-wide resource visualization and intelligent operation and maintenance, the broadband-side network management platform needs to acquire and integrate the state data of all narrowband terminal nodes to construct a complete topology structure covering all nodes and their link relationships. If the broadband network management platform actively requests the status of each node from the converged broadband and narrowband access gateway using the traditional periodic polling method, it will generate a large amount of invalid query overhead. This is because topology changes of narrowband terminals are low-frequency events, and the node status is not updated in most polling cycles. The polling operation will only increase network load and computing resource consumption, and it cannot guarantee timely acquisition of updated data after a topology change occurs. According to the technical disclosure materials, the project studies network status awareness and visualization technology. Through standardized northbound interfaces (such as RESTful API, SNMP), it outputs structured and visualized evaluation data to the upper-layer network management system to achieve a "one-map presentation" of resource status. At the same time, at the topology dynamic construction level, relying on an event-driven architecture, it monitors events such as terminal access, offline, and roaming handover in real time. When a new MAC binding is detected or no heartbeat is received for three consecutive cycles (heartbeat timeout), it automatically triggers a topology update to add or delete nodes or adjust connection edges. Therefore, in the technical solution of this application, a pre-established subscription channel mechanism is adopted, enabling the broadband-side network management platform to receive the change push of the last valid state entity of the node in real time in an event-driven, polling-free manner. After receiving the data, the platform performs topology classification and merging according to the node identifier, integrating the scattered single-node state data into a set of network-wide fused topology state with complete parent-child link relationships. This provides a unified, complete, and structured network-wide topology data foundation for the link connectivity confidence inversion in step S4 and the visualization rendering in step S5.

[0047] Furthermore, due to the asynchronous and intermittent working mode of narrowband terminals, the reporting times of different terminals vary significantly, and the same terminal may generate multiple state entities within a short period due to multiple topology switches. After receiving multiple state entity pushes from the digital twin state pool, the broadband network management platform needs to group and deduplicate the entities according to the source media access control address to ensure that only the latest valid state is retained for the same terminal. Simultaneously, it needs to sort the state entities of different terminals in ascending order by timestamp to establish an ordered event processing sequence. It also needs to filter out late-arriving old state entities to prevent old data from overwriting new data due to network latency or message out-of-order delivery. Finally, by parsing and aggregating the parent-child link relationships of all terminals, a complete topology state set containing all nodes and their interconnections is formed.

[0048] In practical implementation, firstly, the last valid state entity of the node is received without polling through a subscription channel pre-established by the broadband-side network management platform. Then, the nodes are grouped, deduplicated, and ordered in ascending order by timestamp according to their source media access control addresses to generate an ordered stream of node state events. Specifically, the broadband-side network management platform establishes a persistent subscription channel with the broadband-narrowband converged access gateway during the system initialization phase. In the embodiments of this application, the backend pushes topology change events in real time via RESTful API or WebSocket to ensure that the view is synchronized with the actual network state. In this scheme, the subscription channel establishment process is as follows: the broadband-side network management platform initiates a WebSocket connection request or MQTT topic subscription request to the northbound interface of the broadband-narrowband converged access gateway through the broadband network, registering to subscribe to the change events of the last valid state entity of the node in the digital twin state pool. After completing the subscription registration, the gateway assigns a unique session identifier to the subscription channel and maintains a long-lived connection. Subsequently, whenever the last valid state entity of any node in the digital twin state pool is overwritten and updated, the event listening module inside the gateway detects the state change event and actively pushes the updated last valid state entity of the node to the broadband-side network management platform through the established subscription channel, without the platform initiating any polling requests.

[0049] The data receiving module of the broadband network management platform continuously monitors the status entity push messages arriving on the subscription channel. Because there are a large number of narrowband terminal nodes across the network, and the topology change events of each node are unevenly distributed over time, the platform may receive multiple status entity pushes from multiple different nodes within a certain time window, or it may receive multiple status entities from the same node due to multiple topology switches within a short period. Therefore, the platform needs to perform grouping, deduplication, and ordering processing on the received status entity batches.

[0050] Based on this, the platform uses the source media access control address (MACC) in the last valid state entity of each received node as the grouping key to group all state entities received within the same time window according to their source MACC addresses. Within each group, if multiple state entities correspond to the same source MACC address (indicating that the node has undergone multiple topology changes within a short period), the update timestamp field values ​​of each entity are compared. Only the entity with the largest timestamp (i.e., the latest) is retained as the valid representative of that node, while the other older entities are discarded. The group deduplication decision logic is as follows: for the same source MACC address... The corresponding set of state entities The valid entities retained are:

[0051] in, For the first The update timestamp field value in the status entity. Valid entities to be retained.

[0052] After deduplication, each source media access control address corresponds to only one valid state entity. The platform sorts all deduplicated valid state entities in ascending order according to their update timestamp field values, forming an ordered sequence of events organized chronologically. Specifically, for any two valid state entities... and ,like ,but Arranged in Previously, after grouping, deduplication, and timestamp ascending ordering, an ordered node state event stream was formed. :

[0053] in, For the first A state entity, This represents the total number of valid state entities after deduplication within the current time window, with each entity arranged in order from earliest to latest according to its update timestamp.

[0054] Next, the ordered node state event stream is parsed for parent-child link relationships and late-arriving entities are filtered to obtain the node link state mapping table. That is, after obtaining the ordered node state event stream, the broadband network management platform needs to extract and establish parent-child link connection relationships between nodes from each state entity, while identifying and removing old state entities that are late due to network latency, to ensure that the link mapping table records the latest valid link states of each node.

[0055] In this process, firstly, the platform processes each valid state entity one by one according to the ordered sequence of node state events. For the state entity currently being processed... The system extracts the node identifier encoding component from it. (Corresponding source media access control address) (representing the identity of child nodes) and parent node identifier encoded components (Corresponding parent node media access control address) (representing the parent node's identity), forming a directed edge between the two. , indicating from child node Point to parent node The system defines the topological connectivity of the entity. Simultaneously, it extracts the link quality coded components from this entity. Terminal status coding components and update timestamp This information serves as an adjunct attribute of the directed edge. The formal representation of a single link relationship is as follows:

[0056] Based on this, during the process of parsing the link relationships one by one, the system maintains a temporary mapping table with the source media access control address as the key, recording the update timestamp of the latest status entity accepted by each node. When a new status entity is processed, the system first checks whether the source media access control address corresponding to the entity already exists in the temporary mapping table. If a record already exists, the update timestamp of the current entity is compared. Timestamps already recorded in the temporary mapping table :like This indicates that the current entity is in an updated state; accept the entity and update the record in the temporary mapping table; if This indicates that the current entity is old state data that arrived late, and it is identified as a late entity and filtered out and discarded.

[0057] In the embodiments of this application, dynamic updates rely on an event-driven architecture, which monitors events such as terminal access, offline, and roaming switching in real time. When a new MAC binding is detected or no heartbeat is received for three consecutive cycles (heartbeat timeout), a topology update is automatically triggered to add or delete nodes or adjust connection edges. An aging time is also set to mark inactive terminals for a long time as "offline" and gray them out to avoid topology redundancy. In the process of filtering late entities, the system also introduces an aging time mechanism: if the difference between the timestamp of a node's record in the temporary mapping table and the current system time exceeds a preset aging threshold, the record is marked as expired. Subsequent arrivals of the node entity are accepted regardless of the timestamp, to avoid the permanent stagnation of the node state due to a long period without updates.

[0058] After resolving parent-child link relationships and filtering late entities, the system aggregates all valid link relationship records that pass the filtering decision into a node link state mapping table. This mapping table uses the source media access control address as the key and the corresponding directed link relationship and associated attributes as the values:

[0059] in, This represents the number of valid nodes after filtering by late entities. Each record represents the latest valid parent-child link status of an end node.

[0060] Furthermore, the node link state mapping table is aggregated across the entire network to obtain a unified network topology state set. In other words, after obtaining the node link state mapping table, the broadband-side network management platform needs to aggregate the scattered single-node link records into a complete topology structure with a network-wide perspective. In the embodiments of this application, the topology modeling adopts a graph model, with APs and terminals (cameras, sensors, etc.) as nodes and wireless connection relationships as edges, constructing a two-level "AP-terminal" hierarchical structure, while also supporting recursive modeling of ZigBee / LoRa multi-hop networks.

[0061] In this process, firstly, all unique node identifiers are extracted from the node link state mapping table, including all source media access control addresses appearing as keys (child node set) and all media access control addresses appearing as parent node identifiers as values ​​(parent node set). Then, a union operation is performed on the two sets to obtain the complete network node set. :

[0062] The first set is the set of all source node identifiers in the node link state mapping table, and the second set is the set of all parent node identifiers. The union operation ensures that even if some parent nodes have not reported their own status (such as converged access devices acting as aggregation nodes), they are included in the entire network node set. Subsequently, all directed edge relationships are extracted from the node link state mapping table to construct the entire network edge set. For each record in the mapping table, the system generates a directed edge from the child node to the parent node, and uses link quality, terminal battery level, and update timestamp as auxiliary attributes of the edge:

[0063] During the construction of the entire network node set and the entire network edge set, the system also needs to perform topology classification, that is, classifying nodes into the corresponding topology level according to their position in the hierarchical structure. In the embodiments of this application, for data acquisition, the system obtains the associated logs of terminal MAC address binding / unbinding from WAPI broadband APs and low-power narrowband gateways (such as LoRa gateways), while continuously monitoring the beacon frames or heartbeat messages periodically sent by the terminals to confirm their online status. The system determines the hierarchy based on the role of nodes in the directed edges: nodes that appear only as parent nodes and not as child nodes are classified as top-level aggregation nodes (such as the converged access gateway itself or a first-level AP); nodes that appear as both parent and child nodes are classified as intermediate relay nodes; nodes that appear only as child nodes and not as parent nodes are classified as terminal leaf nodes (i.e., narrowband sensor terminals). The topology classification determination function is as follows:

[0064] The splicing and merging operation splices all child node links under the same parent node according to their affiliation, forming a local star-shaped topology subgraph centered on the parent node. Then, these local subgraphs are recursively spliced ​​according to their hierarchical relationships, ultimately forming a complete converged topology structure covering all nodes and links in the entire network. (Full network converged topology state set) Represented in graphical form as follows:

[0065] in, For the entire network's integrated topology state set, For the entire network of nodes, For the set of directed edges in the entire network, It is a set of node attributes (including information such as the hierarchical classification of each node and the terminal power level). It is a set of edge attributes (including the quality score, update timestamp, and other information for each link).

[0066] Specifically, in step S4, based on the time difference between the current system time and the update timestamp, connectivity confidence inversion is performed on each link in the network-wide converged topology state set to transform discrete connectivity states into continuous confidence scores, which are then aligned in parallel to obtain the topology connection confidence gradient matrix. It should be understood that in step S3, the broadband-side network management platform has constructed a network-wide converged topology state set containing all nodes and their directed link relationships. However, the state information of each link in this set is still expressed in a discrete binary form—the existence of a node's last valid state entity only indicates that the node was indeed in a connected state at the time indicated by the update timestamp, but it cannot directly reflect the actual connectivity probability of the link over time. Narrowband terminals adopt an intermittent working mode of deep sleep and periodic wake-up; their links are not truly disconnected during sleep, but are in an uncertain state of "known last good but currently unconfirmed." If terminals that fail to report within the expected time are simply marked as offline, normally dormant nodes may be misjudged as disconnected. Conversely, if all nodes are considered connected without considering time decay, nodes that are actually disconnected may be misjudged as online. Therefore, in the technical solution of this application, a connectivity confidence inversion mechanism is introduced. Based on the time difference between the current system time and the update timestamps of each link, the discrete connectivity states are transformed into continuous confidence scores within a closed interval of zero to one using an adaptive decay function. The confidence scores of all links are then aligned by the node index rows and columns to construct a topology connectivity confidence gradient matrix, providing link-by-link visual mapping parameters for gradient transparency rendering in step S5.

[0067] In practical implementation, firstly, based on the time difference between the current system time and the updated timestamp, link timing parameters are extracted and adaptive attenuation calibration is performed on the entire network converged topology state set to obtain a link timing attenuation parameter set. During this process, timing-related parameters for each link are extracted from the entire network converged topology state set, and an attenuation rate coefficient adapted to its own characteristics is calibrated for each link. Specifically, firstly, the current time value of the local system clock on the broadband side network management platform is read, denoted as... Subsequently, the system traverses the entire network edge set in the fused topology state set, and for each directed edge, extracts its update timestamp. And calculate the time difference between the current system time and the update timestamp. :

[0068] in, This indicates the time span since the link was last confirmed to be connected. The smaller the value, the more recent the status information of the link. A higher value indicates that the link has not received a status update for a long time. Furthermore, since different narrowband terminals have different wake-up cycles, operating modes, and communication environments, using a uniform fixed attenuation rate to calculate the confidence level for all links will lead to unreasonable assessments—terminals with longer wake-up cycles will inevitably have a greater time difference during normal sleep periods than terminals with shorter wake-up cycles. If the same attenuation rate is used, long-cycle terminals will be assigned excessively low confidence levels even under normal operating conditions. Therefore, an attenuation rate parameter adapted to the characteristics of each link is calibrated, and this parameter is determined based on the historical reporting cycle statistics of the corresponding terminal. Specifically, the system obtains the average reporting cycle of the terminal from historical records. (That is, the average time interval between two consecutive topology status reports from the terminal), and the decay half-life is set based on this. This ensures that when the time difference is exactly equal to one reporting period, the confidence level decays to a preset baseline level. The calibration formula for the decay half-life is:

[0069] in, The value of the natural logarithm 2 is used to ensure that when At this point, the output of the subsequent exponential decay function is exactly 0.5, meaning that when a normal reporting cycle just times out, the confidence level drops to the intermediate level of 50%. For newly joined terminals without historical reporting cycle statistics, the system uses the default reporting cycle as the benchmark for calibration. After time-series parameter extraction and adaptive decay calibration, a time-series decay parameter record is generated for each link in the entire network edge set, including the source node identifier, target parent node identifier, time difference, and decay half-life of the link. The time-series decay parameter records of all links are collected to form a link time-series decay parameter set. :

[0070] in, This represents the total number of links in the entire network edge set.

[0071] Next, connectivity confidence inversion is performed on the link timing attenuation parameter set to transform discrete connectivity / disconnection states into continuous confidence scores, resulting in a continuous confidence record set for the link. During this process, for each record in the link timing attenuation parameter set... The system calculates the connectivity confidence score of the link at the current moment using an exponential decay function. :

[0072] in, It is a natural constant. This is the time difference between the current system time and the update timestamp of this link. This is the adaptive decay half-life of the link. The exponential decay function has the following properties: when... At that time (i.e., just after receiving the status update). This indicates complete confidence that the link is connected; when Time (i.e., exactly exceeding one normal reporting cycle). This indicates that the confidence level has dropped to the median level of 50%; when Time (i.e., a very long period without updates), This indicates that the confidence level is close to zero, and there is a high suspicion that the link has been broken. The output value of the exponential decay function is always at... Within the range, the boundedness of the confidence score is ensured.

[0073] In addition, combined with link quality coding components The confidence score is adjusted. Since link quality itself reflects the physical reliability of the link—a link with higher signal strength is more likely to remain connected during sleep, while a link with lower signal strength is more likely to disconnect due to minor environmental changes—link quality is incorporated as a modulation factor into the confidence score calculation. The adjusted connectivity confidence score is shown below. for:

[0074] in, This is the baseline confidence retention factor, used to ensure that even links with lower quality are not excessively undervalued due to signal factors. When When the link quality is optimal, the correction factor is 1, and the confidence level is not reduced; when When the link quality is at its worst, the correction factor is: The confidence level was reduced to its original value. times.

[0075] For each link in the link time-series attenuation parameter set, the above connectivity confidence inversion calculation is performed, generating a record for each link containing the source node identifier, the target parent node identifier, and the corrected continuous confidence score. The confidence records of all links are then compiled to form a link continuous confidence record set. :

[0076] in, The continuous confidence score for each link is adjusted for both time decay and link quality factors. This is the set of link continuity confidence records.

[0077] Then, the link continuity confidence record set is aligned in rows and columns and padded with the largest possible value to obtain the topology connectivity confidence gradient matrix. In this process, the dimension of the matrix is ​​first determined. This is done from the entire network node set of the fused network topology state set. Get the number of all unique nodes , build a A square matrix. The system assigns each node in the entire network node set a number from 0 to... Given a unique integer index, establish a bidirectional mapping relationship between node identifiers and matrix row and column positions. Let the mapping function from node identifier to index be... Then the first in the matrix Line 1 The elements of the column correspond to the nodes. To the node The connectivity confidence value of the link between them. The system initializes this. All elements of the square matrix are zero, and then the continuous confidence record set of the links is traversed. Each record in , and the confidence score Fill in the matrix line, number Column position. The row and column alignment fill operation is as follows:

[0078] in, For to be built Topological connectivity confidence gradient matrix.

[0079] Specifically, in step S5, a topology base map is constructed based on the entire network's fused topology state set, and gradient transparency rendering is applied to each link in the base map using the topology connection confidence gradient matrix as a mapping parameter to obtain a smooth fused network topology view. It should be understood that in step S4, the system has already converted the discrete connectivity states of each link in the entire network into continuous confidence scores through connectivity confidence inversion, and constructed a topology connection confidence gradient matrix in a row-column aligned manner. However, this matrix, as a purely numerical two-dimensional data structure, cannot be directly perceived and understood by operations and maintenance personnel. Therefore, it is necessary to transform the numerical confidence matrix into a visualized topology view, enabling operations and maintenance personnel to intuitively and quickly grasp the true status of the entire network's topology connections through a graphical interface. Traditional topology visualization schemes directly perform full map redrawing and node rearrangement when link states change, resulting in node jumping and screen flickering in the topology view, severely interfering with operations and maintenance personnel's interpretation of the actual network structure. In the technical solution of this application, gradient transparency rendering of each link is performed by using the confidence gradient matrix as the mapping parameter, and a double buffering and frame synchronization replacement mechanism is adopted to make the link status in the topology view smoothly presented in a way of gradual transparency change - high confidence links are displayed as completely opaque, and low confidence links are displayed as semi-transparent or nearly fully transparent - thereby avoiding abrupt jumps between "showing / disappearing" links, and realizing a smooth, continuous, and flicker-free fused network topology view output.

[0080] Figure 4 This is a flowchart of step S5 in the wireless broadband and narrowband network resource visualization and automatic topology discovery construction method according to an embodiment of this application. Figure 4 As shown, in the first embodiment of this application, step S5 includes: S51, using the connection confidence in the topology connection confidence gradient matrix as the link constraint weight, and combining it with historical display coordinate anchor points, performing hierarchical layout solution on the entire network fusion topology state set to generate fusion topology layout data; S52, performing continuous mapping between connection confidence and visual attributes on the fusion topology layout data to obtain a confidence visual primitive set; S53, performing transparency synthesis and full-frame rendering on the confidence visual primitive set in the off-screen buffer to obtain a background topology rendering frame; S54, based on the verification results of integrity identifier and topology version number, performing atomic pointer swapping and frame synchronization replacement on the background topology rendering frame during the vertical synchronization gap to output a smooth fusion network topology view.

[0081] Specifically, in step S51, the connection confidence in the topology connection confidence gradient matrix is ​​used as the link constraint weight, and combined with historical display coordinate anchor points, a hierarchical layout solution is performed on the entire network fusion topology state set to generate fusion topology layout data. During this process, the instantaneous confidence scalar value in the topology connection confidence gradient matrix is ​​used as the link constraint weight. For slave nodes in the entire network fusion topology state set... To the node The link, its constraint weight Directly retrieve the confidence value at the corresponding position in the matrix:

[0082] in, For the topological connectivity confidence gradient matrix, the first... Line 1 The element value of the column, i.e., from the node To the node The connectivity confidence score.

[0083] The layout optimization objective function consists of three penalty terms: link length deviation, historical position deviation, and inter-node exclusion. The link length deviation term constrains the deviation between the displayed distance and the desired distance between connected nodes. It is weighted by link constraint weights, resulting in a stronger penalty for the length deviation of high-confidence links (i.e., a greater tendency to maintain tight connections) and a weaker penalty for the length deviation of low-confidence links (i.e., allowing their connected nodes to gradually loosen). The historical position deviation term constrains the displacement amplitude between the current frame coordinates and the previous frame coordinates of each node, preventing large jumps between adjacent frames. The inter-node exclusion term prevents excessive overlap of non-connected nodes in the display space, ensuring the readability of the topology graph. Specifically, the formal expression of the layout optimization objective function is as follows:

[0084] in, The total penalty value of the layout optimization objective function is to minimize the optimization objective. ; For the node To the node The link constraint weights are the instantaneous confidence scalar values ​​at the corresponding positions in the topology connection confidence gradient matrix. For nodes With nodes Display the current Euclidean distance on the canvas; The expected display length of the link (preset constant) represents the standard distance between connected nodes under ideal conditions; For nodes The two-dimensional coordinate vector to be solved in the current frame; For nodes The display coordinate position determined in the previous frame, i.e., the historical display coordinate anchor point; The weighting coefficient for the historical position deviation term controls the strength of the node's "anchoring" to the previous frame's position. The larger the value, the more likely the node is to remain near the position of the previous frame; The weighting coefficients for the repulsion term between nodes control the intensity of the repulsion between non-connected nodes; It is the set of directed edges in the entire network; It is a collection of all nodes in the network.

[0085] Regarding hierarchical constraints in the layered layout, hierarchical constraints are applied to the vertical coordinates of nodes at each level based on node hierarchy: the root node is placed at the top layer of the canvas, relay nodes are placed in the middle layer, and leaf nodes are placed at the bottom layer. Hierarchical constraints are added to the objective function as additional penalty terms to ensure that the layout result conforms to the hierarchical display logic of tree or star topology.

[0086] The objective function is minimized using gradient descent or force-directed iterative algorithms, iteratively calculating the optimal 2D coordinates of each node on the display canvas. After solving, the system outputs fused topology layout data, including the display coordinates of every node in the entire network and the start and end coordinates of each link. Simultaneously, the system stores the node coordinates obtained in the current frame into a historical display coordinate anchor point cache, serving as a reference for the layout solution in the next frame.

[0087] Specifically, in step S52, the fused topology layout data is continuously mapped between connection confidence and visual attributes to obtain a set of confidence visual primitives. In this process, gradient transparency is used as the core visual mapping dimension, supplemented by line width and color as auxiliary mapping dimensions. Specifically, firstly, the confidence score of each link is obtained from the topology connection confidence gradient matrix. This is linearly mapped to the alpha channel value (transparency value) of the linked graphic element. The alpha channel value ranges from 0 to 255 (8-bit color depth), where 255 represents complete opacity and 0 represents complete transparency. The mapping formula is:

[0088] in, For the node To the node The Alpha channel value of the link in the display screen; This represents the continuous confidence score for the link, with a value range of [range missing]. ; This is a rounding function that quantizes continuous values ​​into 8-bit integers. hour, The link is displayed in a completely opaque manner, indicating a high degree of certainty that it is connected; when hour, The link is displayed semi-transparently, indicating a medium confidence level; when When it approaches 0, A value close to 0 indicates that the link is almost invisible, representing extremely low confidence.

[0089] A continuous mapping connecting confidence level and linewidth serves as an auxiliary visual dimension, making high-confidence links not only more opaque but also thicker (more noticeable), while low-confidence links are not only more transparent but also thinner (less noticeable). The linewidth mapping formula is:

[0090] in, This is the display linewidth of the link, in pixels; This is the minimum line width, corresponding to the line width when the confidence level is 0; This represents the maximum line width, corresponding to the line width with a confidence level of 1.

[0091] The continuous mapping between confidence level and color achieves a gradual color gradation from green (high confidence) to yellow (medium confidence) and then to red (low confidence) through hue interpolation. The color mapping uses linear interpolation of the hue channel in the HSV color space.

[0092] in, The hue value of this link, with a range of values. Degree. When hour, (Green); when hour, (Yellow); when hour, (red).

[0093] The system generates a complete visual primitive description for each link in the entire network, including the link's start and end coordinates (from the fused topology layout data), alpha channel value, line width, color, and other visual attributes. Simultaneously, the system generates a corresponding node primitive description for each node, including the node's display coordinates, icon type (distinguished by device type), and annotation information. All link and node primitives are aggregated to form a confidence-based visual primitive set.

[0094] Specifically, in step S53, the confidence visual primitive set is composited with transparency and rendered as a whole frame within the off-screen buffer to obtain the background topology rendering frame. In the embodiments of this application, the front end uses high-performance graphics libraries such as WebGL or D3.js to support rendering of thousands of nodes. To avoid screen tearing caused by users seeing unfinished images during the rendering process, all rendering operations are performed in the off-screen buffer (background buffer). Specifically, the initialization process of the off-screen buffer is as follows: according to the rendering order from bottom to top, each primitive in the confidence visual primitive set is rendered to the off-screen buffer in sequence. The rendering order is as follows: first, the background mesh and coordinate reference lines (bottom layer) are drawn; then, all link primitives are drawn, each link according to its Alpha channel value. Line width and color The drawing process involves blending links with the background using alpha values ​​while drawing them; finally, all node primitives (top layer) are drawn, with nodes drawn on top of the links in a completely opaque manner.

[0095] Alpha blending is the core calculation process for transparency composition. When a link primitive with a semi-transparent alpha value is drawn on an existing background, the final displayed pixel color is calculated by the following blending formula:

[0096] in, These are the final pixel color values ​​after synthesis. The normalized Alpha mixing factor has a range of values. ; The color value of the link element itself (determined by hue mapping); The existing background color value for the location covered by the link element. When When the link color completely covers the background (completely opaque); when At the initial stage, the link color is completely invisible, only the background color is displayed; the intermediate value presents a blending effect of the link color and the background color, forming a semi-transparent visual gradient. After all primitives are drawn, the off-screen buffer contains a complete topology view frame that has been composited with transparency. The system sets an integrity flag (marking that the drawing is complete and the data is intact) and a topology version number (marking the topology data version corresponding to this frame) for this frame, forming a background topology rendering frame.

[0097] Specifically, in step S54, based on the verification results of the integrity identifier and the topology version number, atomic pointer swapping and frame synchronization replacement are performed on the background topology rendering frame during the vertical synchronization gap to output a smoothly merged network topology view. After the background topology rendering frame is drawn, the system needs to replace it in the foreground display buffer to present it to the maintenance personnel. However, if the replacement is performed directly while the display device is reading data from the foreground buffer, it may cause a "screen tearing" phenomenon where the upper half of the screen displays the old frame and the lower half displays the new frame. Therefore, a vertical synchronization (VSync) mechanism is adopted to perform buffer swapping during the brief gap (vertical synchronization gap) between when the display device finishes reading one frame and before it starts reading the next frame.

[0098] Specifically, first, check if the integrity flag of the background topology rendering frame is in the "complete" state to confirm that the entire frame rendering has been completed without any abnormal interruption. Second, compare the topology version number of the background frame with the topology version number of the foreground frame to confirm that the data version corresponding to the background frame is indeed newer than the currently displayed frame. The verification logic is as follows:

[0099] in, This serves as an integrity indicator for the background frame (1 indicates that drawing is complete). This is the topology version number of the background frame; This refers to the topology version number of the foreground frame. When the verification condition is met, the system performs an atomic pointer swap operation within the vertical synchronization interval—atomically swapping the foreground buffer pointer with the background buffer pointer. This instantly makes the original background buffer the foreground buffer, which is then read by the display device, while the original foreground buffer becomes the background buffer, waiting to be overwritten by the next frame. The execution time of the atomic pointer swap is in the nanosecond range, far less than the display cycle of a single pixel. Therefore, the user cannot perceive the switching process at all, and the image achieves a seamless and tear-free smooth transition between frames. When the verification condition is not met (such as the background frame not being fully drawn or the version number being mismatched), the system does not perform the swap operation. The foreground continues to display the current frame, and the background continues to complete its drawing, waiting for the next vertical synchronization interval to attempt verification and swapping again.

[0100] After atomic pointer swapping and frame synchronization replacement, the foreground display buffer presents the latest smoothly merged network topology view. The transparency of each link in this view reflects its connectivity confidence in a continuous, gradual change—recently updated links are clearly visible, links that haven't been updated in a long time gradually fade, and completely disconnected links become almost invisible. Maintenance personnel can intuitively judge the connectivity health of different areas of the network by observing the transparency gradient distribution of the links, and quickly locate potentially faulty weak links. Simultaneously, the dual-buffered frame synchronization replacement mechanism ensures that frame-to-frame transitions do not produce tearing or flickering. Combined with the constraints of historical coordinate anchor points on node positions in layout optimization, the topology view maintains a stable, continuous, and smooth visual presentation even in dynamic scenarios where link states are constantly changing.

[0101] Specifically, research shows that in the topology visualization application scenario of converged broadband and narrowband wireless communication networks, the converged access node carries the unified access of WAPI broadband terminals and low-power narrowband terminals. Due to limitations in battery power supply and low-power design, narrowband terminals adopt an intermittent working mode of deep sleep-periodic wake-up. Their topology survival status relies on an asynchronous publish-subscribe mechanism to report the last known good state entity to the digital twin state pool. After the visualization engine of the upper-layer network management platform performs link confidence assessment based on this state entity, it needs to input each element in the topology connection confidence gradient matrix as link constraint weights into the layout optimization objective function to jointly calculate link length deviation, historical position deviation, and inter-node exclusion terms, ultimately solving for the displayed coordinates of each node.

[0102] In the first embodiment, sub-step S51, when constructing the layout optimization objective function, directly uses the instantaneous confidence scalar value in the topology connection confidence gradient matrix as the link constraint weight, without considering the temporal change rate attribute of the confidence value between adjacent evaluation periods.

[0103] In typical deployment environments such as substations and distribution stations, the asynchronous sleep / wake-up mechanism of narrowband sensor nodes determines that their link confidence does not decay or recover at a uniform rate. When a narrowband terminal is unexpectedly disconnected from the network due to fault conditions such as abnormal power loss, radio frequency shielding, or continuous occupation of the communication channel by broadband equipment, the confidence of the corresponding link of that node will drop sharply from a high value close to 1 to a low value close to 0 within a single evaluation period (on the order of several seconds to tens of seconds). Conversely, when the narrowband terminal completes wake-up and successfully reports the topology change semantic vector through the asynchronous publish-subscribe protocol, the corresponding entity in the digital twin state pool is refreshed immediately, and its link confidence will rise sharply from a low value to a high value. The original mechanism only perceives the current position of the confidence and ignores its movement trend, resulting in an undamped rigid coupling between the link constraint weights in the layout objective function and the confidence—a sharp drop in confidence will instantly reduce the weights to zero, and a sharp rise in confidence will instantly overload the weights. This rigid coupling causes the gradient field faced by the optimization solver to change drastically between adjacent display frames, resulting in a large jump in node position, i.e., layout shock.

[0104] Specifically, when a narrowband node suddenly goes offline, the constraint weights of the links connected to it instantly drop to zero. The node loses spatial traction from its parent node and, under the influence of the repulsion term, is rapidly ejected to a coordinate region far from its original position. When the node returns online, the link weights instantly return to full value, and the node is violently pulled back to the vicinity of its parent node by a strong link traction force, resulting in a significant positional rebound. This layout shock phenomenon, in the spatial dimension, negates the flicker-free effect achieved by subsequent double-buffered frame synchronous replacement in the temporal dimension. This causes the topology view to still exhibit momentary jumps in localized areas when the narrowband node's state changes drastically, severely interfering with maintenance personnel's interpretation of the actual network topology and fault location.

[0105] In view of the above-mentioned technical defects, this application further proposes a second embodiment.

[0106] Specifically, firstly, based on the previous period's confidence retention snapshot stored in the rolling cache as the differential benchmark, element-wise temporal differencing is performed on the topology connection confidence gradient matrix of the current evaluation period to obtain the link confidence temporal change rate matrix. In the context of asynchronous state updates in converged broadband and narrowband wireless networks, the confidence decay or recovery of different links is not synchronous. Some links may be in a rapid decline phase during the current evaluation period, while others may be recovering from a trough, and some links may remain in a steady state for a long time. Without identifying and distinguishing the changing trends of each link, the deployment system can only passively respond after the confidence jump, unable to prepare for spatial transition before or during the jump. Therefore, a quantitative description of the changing trend for each link needs to be established from a temporal perspective.

[0107] After calculating the topology connectivity confidence gradient matrix in each evaluation cycle, the system writes a complete copy of the matrix to the rolling cache, forming a previous cycle confidence retention snapshot. When the topology connectivity confidence gradient matrix for a new evaluation cycle is generated, the previous cycle confidence retention snapshot is used as the differential reference. Element-wise subtraction is performed on the two matrices of the same dimension to calculate the net change in confidence for each link between two adjacent evaluation cycles, as shown below:

[0108] in, Let be the time-series rate of change of confidence level for the link from node i to node j. This represents the element value at the corresponding position in the topological connectivity confidence gradient matrix for the current nth evaluation period. The element value at the corresponding position in the snapshot is retained for the previous period's confidence level.

[0109] It should be noted that the polarity of the differential results directly reflects the direction of link state movement—positive values ​​indicate that the link is recovering, corresponding to the narrowband node having just completed wake-up reporting and refreshed the last known good state entity in the digital twin state pool; negative values ​​indicate that the link is decaying, corresponding to the narrowband node having entered a deep sleep cycle and failing to complete reporting within the expected time window; zero values ​​indicate that the link remains in a steady state between two evaluation cycles. Aligning and encapsulating the differential results of all links according to the same row and column node indices as the topology connection confidence gradient matrix yields a link confidence temporal change rate matrix of the same dimension. By introducing a temporal differential mechanism, the placement system gains the ability to perceive the direction of link state movement, enabling it to distinguish, at the numerical level, two physically completely different conditions: a rapidly decreasing 0.6 and a long-term stable 0.6. This provides a directional basis for subsequent constraint pre-adjustment, rather than treating them as equals during placement optimization.

[0110] Next, a continuous compression mapping of each element in the link confidence temporal rate of change matrix is ​​performed within a bounded interval from zero to one to obtain the layout constraint pre-adaptation factor matrix. After obtaining the link confidence temporal rate of change matrix, the unbounded rate values ​​need to be converted into bounded factors that can directly participate in constraint force modulation. Directly using the original rate values ​​has two problems: first, the absolute value range of the rate of change is affected by the evaluation period length and the amplitude of network state fluctuations, lacking a normalization benchmark; second, without saturation compression, extreme rate of change will produce modulation coefficients that exceed physical meaning, leading to negative constraint forces or abnormal amplification far exceeding the original confidence. Therefore, a nonlinear mapping needs to be designed that can both preserve polarity information and limit the output to a bounded interval.

[0111] A saturated nonlinear mapping function with a temperature parameter is applied to each element of the link confidence time-series rate of change matrix to achieve continuous compression from the unbounded real domain to a closed interval from zero to one, where the mapping relationship is:

[0112] in, The pre-adaptation factor for the layout constraints of the link from node i to node j. This represents the element value at the corresponding position in the link confidence time-series rate of change matrix. To control the temperature parameter of the mapping curve sensitivity, It is an exponential function with the natural constant as its base.

[0113] It should be noted that when the rate of change is a large negative value (the link confidence is rapidly decreasing, corresponding to an abnormal disconnection of a narrowband node or continuous interference with the channel), the mapping output approaches zero. This means that the constraint traction of the link in subsequent layout optimization will be significantly relaxed in advance, allowing the displayed coordinates of the relevant nodes to smoothly transition to a relaxed state before the link completely fails, rather than experiencing a sudden positional change the instant the confidence reaches zero. When the rate of change is zero (the link is in a steady state, corresponding to a narrowband node within a normal asynchronous reporting cycle), the mapping output is the baseline level corresponding to 0.5, without applying a directional bias to the original constraint characteristics. When the rate of change is a large positive value (the link confidence is rapidly recovering, corresponding to a narrowband node that has just completed wake-up and successfully reported), the mapping output approaches a value that is limited by the saturation curve and will not overshoot. This allows the traction of the newly recovered link on the node's position to recover gradually, preventing the recovered node from bouncing back due to full traction and dragging neighboring nodes into a chain reaction of displacement. The introduction of temperature parameters provides on-site maintenance personnel with flexible control methods. In large substations with a large number of nodes and frequent state changes, the temperature parameter can be appropriately reduced to smooth the response curve. In small-scale distribution substations, the temperature parameter can be appropriately increased to accelerate the layout's tracking speed of real topology changes. The scalars of each element after mapping are encapsulated according to the original matrix row and column structure to form a layout constraint pre-adaptation factor matrix.

[0114] Furthermore, the instantaneous state and changing trend of the topology connection confidence gradient matrix and the layout constraint pre-adaptation factor matrix are jointly modulated to obtain fused topology layout data. After the first two steps, the system has obtained the topology connection confidence gradient matrix representing the current position of the link and the layout constraint pre-adaptation factor matrix representing the movement trend of the link, respectively. The link constraint weights in the original layout optimization objective function are determined only by the single dimension of instantaneous confidence, which cannot perceive the differences in layout control strategies that should exist between links whose states are changing rapidly and links whose states are stable in the long term. Therefore, it is necessary to fuse the information of the two dimensions into a unified modulated weight to replace the original single-dimensional weight in the optimization solution.

[0115] The topology connectivity confidence gradient matrix and the layout constraint prefitting factor matrix are multiplied element-wise (Hadamard product) according to the node indices of the same row and column to generate the modulated topology layout constraint weight matrix, where the modulation relationship is:

[0116] in, Let be the element values ​​at positions corresponding to nodes i to j in the modulated topology layout constraint weight matrix. This represents the instantaneous confidence at the corresponding position in the topological connectivity confidence gradient matrix. This refers to the prefitting factor at the corresponding position in the layout constraint prefitting factor matrix.

[0117] Specifically, each element in the modulated topology layout constraint weight matrix simultaneously encodes the instantaneous connectivity state and state change trend of the link. For links with rapidly decaying confidence, their prefitting factor approaches zero, resulting in modulated weights significantly lower than the original confidence value. The spatial traction exerted on adjacent nodes by this link during layout optimization is weakened in advance, and the node display coordinates complete a gradual spatial relaxation transition before the link completely fails. For links with rapidly recovering confidence, their prefitting factor, although approaching one, is always subject to saturation constraints. The modulated weights are slightly lower than the original confidence value and increase with each evaluation cycle. The traction exerted on nodes by this link recovers gradually rather than being applied instantaneously within a single frame.

[0118] For links with a long-term steady-state confidence level, their prefitting factor stabilizes at an intermediate baseline level, and the modulated weight is approximately half of the original confidence level. This link maintains a stable and moderate constraint contribution in the layout. Each element in the modulated topology layout constraint weight matrix directly replaces the constraint weight coefficients in the link length deviation term of the original layout optimization objective function, participating in the joint optimization solution of the link length deviation term, historical position deviation term, and node exclusion term. This significantly narrows the gradient field variation amplitude faced by the optimizer between adjacent display frames, fundamentally suppressing sudden node position jumps caused by abrupt changes in confidence.

[0119] In summary, the second embodiment introduces a pre-adaptive modulation stage based on the time-series change rate before the topology connection confidence gradient matrix enters the layout optimization objective function, enabling the layout system to have the ability to proactively perceive and respond to changes in link status.

[0120] In typical scenarios where narrowband terminals experience sudden changes in confidence due to abnormal disconnections or recovery, the modulated constraint weights undergo a smooth transition before the confidence jumps. This significantly reduces the magnitude of gradient field changes between adjacent frames faced by the optimization solver, effectively constraining the inter-frame displacement of node display coordinates to a pixel-level range imperceptible to the human eye, thus completely eliminating layout impact from a spatial perspective. This gradual transition effect in the spatial dimension complements the anti-flicker effect of subsequent double-buffered frame synchronous replacement in the temporal dimension—the former ensures that the node coordinates themselves do not jump within each frame, while the latter ensures that there is no tearing during frame-to-frame transitions. Together, they enable the front-end visualization screen to consistently present a stable, continuous, and seamless fused network topology view even in complex dynamic scenarios such as frequent sleep / wake-up and abnormal disconnection recovery of narrowband nodes. This effectively ensures that maintenance personnel can accurately interpret the actual network connectivity status and quickly locate faults.

[0121] Meanwhile, the boundedness of the saturated nonlinear mapping function ensures that the modulation weights will not overshoot or become negative under extreme rate of change conditions, thus preventing the layout optimization objective function from falling into numerical divergence or non-convex oscillation due to abnormal weights. This ensures the convergence stability and computational determinism of the layout solver under all conditions and reduces the ineffective iteration overhead under the condition of limited computing power of embedded fusion access nodes.

[0122] In summary, the wireless broadband and narrowband network resource visualization and automatic topology discovery construction method according to the embodiments of this application is explained. Driven by topology change events, it performs cached differential and low-dimensional semantic encoding on narrowband terminal packets. Through an asynchronous publish-subscribe mechanism, it writes the change vector and timestamp into a digital twin state pool, and the broadband side subscribes, categorizes, and splices the entire network topology state without polling. Furthermore, it inverts link connectivity confidence based on state update time and controls the transparency rendering of topology links using a confidence gradient matrix. This reduces redundant packet transmission, terminal energy consumption, and polling load, avoids misjudging periodically dormant nodes as offline, and transforms discrete link connectivity into continuous, gradual states, reducing jumps and flickering caused by topology rearrangement. Ultimately, it achieves automatic discovery, unified aggregation, and smooth visualization of heterogeneous broadband and narrowband network topologies, improving the accuracy and continuity of network status judgment, fault location, and operation and maintenance management.

[0123] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for visualizing wireless broadband and narrowband network resources and automatically discovering and constructing topology, characterized in that, include: S1: Strip the header of the physical control message of the narrowband terminal and compare it with the local historical cache. Only when the parent-child relationship of the node changes, perform low-dimensional semantic encoding to obtain the narrowband topology change semantic vector. S2: The narrowband topology change semantic vector is reported to the broadband and narrowband converged access gateway through the asynchronous publish-subscribe protocol. After decoding and verifying the narrowband topology change semantic vector, it is aggregated with the update timestamp and written into the digital twin state pool to obtain the node's last valid state entity. S3: The subscription channel pre-established by the broadband-side network management platform receives the last valid state entity of the node without polling, and performs topology classification and splicing according to the node identifier to obtain the whole network fused topology state set; S4: Based on the time difference between the current system time and the update timestamp, perform connectivity confidence inversion on each link in the whole network fusion topology state set to transform discrete connectivity states into continuous confidence scores and align them in parallel columns to obtain the topology connectivity confidence gradient matrix. S5: Construct a topology base map based on the set of fused topology states of the entire network, and use the gradient transparency of each link in the base map as the mapping parameter to obtain a smooth fused network topology view.

2. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 1, characterized in that, The physical control message of a narrowband terminal includes the source media access control address, the parent node media access control address, the received signal strength indicator value, and the remaining battery power.

3. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 2, characterized in that, Step S1 includes: The node state feature tuple is obtained by extracting the effective payload of the physical control message of the narrowband terminal through the protocol parser. Based on the historical parent node addresses in the local historical cache, topology state differential decision is performed on the node state feature tuple to generate topology change identifiers. Using topology change identifiers as gating activation factors, low-dimensional semantic encoding is performed on node state feature tuples to obtain narrowband topology change semantic vectors.

4. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 1, characterized in that, Step S2 includes: The narrowband topology change semantic vector reported to the broadband-narrowband converged access gateway via the asynchronous publish-subscribe protocol is deserialized and its integrity is verified to obtain the verified topology semantic payload. The verified topological semantic payload and the update timestamp are orthogonally concatenated using key-value pairs to obtain a time-bound topological state object; Using the source media access control address as the index key, the time-bound topology state object is overwritten into the digital twin state pool to obtain the last valid state entity of the node.

5. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 1, characterized in that, Step S3 includes: The last valid state entity of the node is received without polling through the subscription channel pre-established by the broadband network management platform, and the nodes are grouped and deduplicated according to the source media access control address and ordered in ascending order of timestamp to generate an ordered node state event stream. Perform parent-child link relationship parsing and late entity filtering on the ordered node state event stream to obtain the node link state mapping table; The node link status mapping table is aggregated across the entire network to obtain the set of network-wide converged topology statuses.

6. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 1, characterized in that, Step S4 includes: Based on the time difference between the current system time and the update timestamp, link timing parameters are extracted and adaptive attenuation calibration is performed on the entire network converged topology state set to obtain the link timing attenuation parameter set. Connectivity confidence inversion is performed on the link timing attenuation parameter set to transform discrete on / off states into continuous confidence scores, resulting in a continuous confidence record set for the link. The link continuity confidence record set is aligned in rows and columns and padded with the largest possible value to obtain the topology connection confidence gradient matrix.

7. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 1, characterized in that, Step S5 includes: Using the connection confidence in the topology connection confidence gradient matrix as the link constraint weight, and combining it with historical display coordinate anchor points, the hierarchical layout solution of the entire network fusion topology state set is performed to generate fusion topology layout data. A continuous mapping between connection confidence and visual attributes is performed on the fused topology layout data to obtain a set of visual primitives with confidence. Within the off-screen buffer, the confidence visual primitive set is composited with transparency and rendered in a whole frame to obtain the background topology rendering frame. Based on the verification results of integrity identifier and topology version number, atomic pointer swapping and frame synchronization replacement are performed on the background topology rendering frames during the vertical synchronization interval to output a smooth and fused network topology view.

8. The method for visualizing and automatically discovering topology in wireless broadband and narrowband networks according to claim 7, characterized in that, Using the connection confidence in the topology connection confidence gradient matrix as the link constraint weights, and combining historical display coordinate anchor points, a hierarchical layout solution is performed on the entire network fusion topology state set to generate fusion topology layout data, including: Based on the previous period confidence retention snapshot stored in the rolling cache as the differential benchmark, the element-wise time-series difference of the topology connection confidence gradient matrix of the current evaluation period is performed to obtain the link confidence time-series change rate matrix. The elements in the link confidence time-series rate of change matrix are continuously compressed and mapped to a bounded interval from zero to one to obtain the layout constraint pre-adaptation factor matrix. The instantaneous state and change trend of the topology connectivity confidence gradient matrix and the layout constraint pre-adaptation factor matrix are jointly modulated to obtain fused topology layout data.

Images