Search

CN-122027540-A - Real-time discovery and dynamic update method for air-sky-sea heterogeneous network topology

CN122027540ACN 122027540 ACN122027540 ACN 122027540ACN-122027540-A

Abstract

The invention discloses a real-time discovery and dynamic updating method of an aerospace-field-sea heterogeneous network topology, which comprises the steps of acquiring node state and link quality data, designing a cross-domain compatible topology representation model, abstracting different communication protocols, frequency bands and link characteristics into a unified node and edge structure, and establishing a topology mapping table oriented to aerospace-field-sea multi-domain cooperation. Based on a lightweight neighbor discovery and link state diffusion mechanism, each node perceives a neighbor relation in a local range and forms a local topology, the node position and link availability change trend is dynamically calculated, the topology structure is incrementally updated, when node faults or link interruption occur, a redundant link and an alternative path are utilized to trigger a rapid reconstruction algorithm, self-healing recovery of the topology is realized, and robustness and continuity of a network are ensured. The invention supports low-overhead topology maintenance in a high dynamic environment, thereby improving the real-time performance, compatibility, expandability and robustness of the network.

Inventors

  • GENG JIANCHENG
  • LU YI
  • ZHANG YANQIU
  • XU HAIFENG
  • XIAO LIPENG
  • GE NING

Assignees

  • 南通先进通信技术研究院有限公司

Dates

Publication Date
20260512
Application Date
20251205

Claims (9)

  1. 1. A method for real-time discovery and dynamic update of an aerospace-land-sea heterogeneous network topology is characterized by comprising the following steps: Step 1, a topology sensing module is deployed on heterogeneous nodes to acquire node state and link quality data, the node state data and the link quality data are uniformly processed through a distributed information fusion mechanism, the node state data and the link quality data acquired in real time are synchronously stored in a local data cache library of each node, the sampling frequency is 1Hz-10Hz, and the cache period is more than or equal to 1h; Step 2, designing a cross-domain compatible topology representation model, abstracting different communication protocols, frequency bands and link characteristics into a unified node and edge structure, and establishing a topology mapping table facing the cooperation of space, day, ground and sea multiple domains; Step 3, based on a lightweight neighbor discovery and link state diffusion mechanism, enabling each node to sense a neighbor relation in a local range and form a local topology, and simultaneously adopting a distributed collaborative algorithm to gather local information into a global topology, wherein a node position prediction model adopts a Kalman filtering algorithm, utilizes collected 1h history data of nodes, initializes a state transition matrix and an observation matrix, calibrates noise and observation noise in a process to complete model parameter tuning; Step 4, combining the prediction model with a real-time monitoring result, dynamically calculating node position and link availability change trend, and incrementally updating the topological structure to synchronize the topological state with the actual network environment; And 5, triggering a rapid reconstruction algorithm by using a redundant link and an alternative path when node faults or link breaks occur, so that self-healing recovery of topology is realized, and robustness and continuity of a network are ensured.
  2. 2. The method for real-time discovery and dynamic update of an aerospace-land-sea heterogeneous network topology according to claim 1, wherein the heterogeneous nodes comprise satellites, unmanned aerial vehicles, ground base stations and shipborne terminals, and the node state and link quality data are obtained through means of link detection, position broadcasting and state monitoring.
  3. 3. The method for real-time discovery and dynamic update of an aerospace-earth-sea heterogeneous network topology according to claim 1, wherein the step2 is specifically: The method comprises the steps of constructing a unified topology representation model, namely establishing a link characteristic quantitative calculation-multi-domain topology mapping table, wherein the three-step implementation of cross-domain compatible topology modeling comprises the following steps that the topology representation model adopts a weighted undirected graph G= (V, E, W) structure, wherein a node set V comprises a node V E V corresponding to an air-space-sea heterogeneous node, the node structure is defined as v= (ID, T, P, C, S), each parameter meaning is as follows, ID represents a character string, T represents an enumeration value, P represents a three-dimensional coordinate, C represents a structural body, each side E corresponds to a communication link between two nodes, the side structure is defined as e= (V1, V2, L), V1 and V2 are two nodes connected by a link, L is a link characteristic parameter set, L= (D, B, lr, S1, proto), each parameter meaning is as follows, D represents a one-way time delay of the link, B represents a link actual available bandwidth, lr represents a link packet loss stability coefficient, S1 represents a communication core protocol of the link is adopted; The weight matrix W is a symmetric matrix of |V|×|V|, matrix elements wij represent comprehensive weights of links between the nodes vi and vj and are used for subsequent topology optimization and path selection, and weight values are obtained through calculation of link characteristic parameters; the link stability coefficient S1 and the weight matrix element wij are subjected to quantization calculation, and the specific formula is as follows: calculation formula of link stability coefficient S1 the link stability coefficient is calculated comprehensively based on historical link outage data and real-time link quality, the formula being s1=α× (1-Fh) +β× (1-Lr/100) +γ× (1-l D-D0 l/D0) where: Alpha, beta and gamma are weight coefficients, the condition that alpha+beta+gamma=1 is met, the emergency communication scene alpha=0.4, beta=0.3 and gamma=0.3 is dynamically adjusted according to an application scene, the common monitoring scene alpha=0.3, beta=0.4 and gamma=0.3 is the historical interruption frequency of a link, fh is defined as the number of times of link interruption/total monitoring in the past 1 hour, fh=0.1 is defined as the number of times of link packet loss when no historical data exists, D0 is the reference time delay of the type of link, and the type of the link is determined; Calculation formula of weight matrix element wij the weight matrix element comprehensively reflects the 'availability-transmission efficiency-stability' of the link, wherein the formula is that wij= (B/1000) x S1 x (1-D/1000), wherein B is the actual available bandwidth of the link, S1 is the stability coefficient of the link, D is the unidirectional time delay of the link, the higher the value of wij is in the range of 0-1, the better the comprehensive performance of the link is represented; The topology mapping table is a two-dimensional structured table and is used for realizing the mapping of nodes/links in different domains and unified model parameters, the table is divided into two types, namely a node attribute mapping table and a link attribute mapping table, wherein the node attribute mapping table is used for encoding the original attributes of the nodes in different domains into parameters of a unified node structure v= (ID, T, P, C, S), and the link attribute mapping table is used for encoding the original characteristic parameters of the links in different domains into parameters of a unified edge structure L= (D, B, lr, S1, proto).
  4. 4. The method for real-time discovery and dynamic update of an aerospace-earth-sea heterogeneous network topology according to claim 1, wherein the step 3 is specifically: Each heterogeneous node periodically transmits a broadcast packet through a lightweight neighbor discovery protocol, the broadcast packet comprises node IDs, positions and communication capacity summaries, the broadcast period is dynamically adjusted according to the node movement speed, the broadcast period of a high-speed mobile node is 100ms-500ms, the broadcast period of a static/low-speed node is 1s-3s, after the node receives neighbor broadcast, the node verifies connectivity through a link detection packet, extracts link quality parameters and screens effective neighbors to form a local neighbor list, a local topological graph is built based on the local neighbor list, the distributed hash table is adopted to store local topological information, each node only keeps complete link data of 1-hop-2-hop neighbors, the edge node uploads the local topological summary to regional coordination nodes of a corresponding domain through a hierarchical coordination aggregation mechanism, the regional coordination nodes perform cross-domain topological information interaction with other regional coordination nodes after the local topology in the same domain is de-duplicated and fused, and finally global topology splicing and integration are completed through a multi-domain coordination algorithm, and signaling storm caused by centralized aggregation is avoided.
  5. 5. The method is characterized in that step 4 is specifically implemented by predicting node positions based on a Kalman filtering algorithm, generating a node position prediction interval of 1s-10s in the future by combining node motion track historical data, speed vectors and environmental constraints, predicting link availability by adopting a long-short-term memory network, outputting the availability probability of a link in a plurality of time windows in the future by adopting an input characteristic comprising historical time delay change trend, packet loss rate fluctuation, signal strength attenuation rate and node position relative change quantity, setting a dynamic update trigger threshold by combining a position prediction result and a link availability prediction result, starting incremental update when the node position offset exceeds 10% of a communication coverage radius or the link availability probability is lower than 80%, adjusting the changed node attribute and newly added/failed link in the update process, synchronizing to a local topology and a global topology by a differential update mechanism, and reducing data transmission and calculation cost, and ensuring that the synchronization error of a topology state and an actual network environment does not exceed 200ms.
  6. 6. The method for real-time discovery and dynamic update of an aerospace-earth-sea heterogeneous network topology according to claim 1, wherein the step 5 is specifically: Firstly, defining a fault location and an influence range, based on fault signals reported by a real-time monitoring module, including node heartbeat loss, link delay sudden increase and packet loss rate reaching 100%, calling the node or link associated information in a topology mapping table through a fault node ID or broken link identification, and calculating a fault influence radius, wherein the node fault influence radius is a communication coverage radius of the node, the link broken influence radius is an intersection range of the communication coverage radius of nodes at two ends of the link; Searching redundant links within a 1-2-hop range around a fault node, namely links which are not occupied by current business or have an occupancy rate lower than 30%, extracting characteristic parameters of the redundant links, including time delay, bandwidth and stability coefficients, according to a topology mapping table and a global topology structure, traversing a network subgraph formed by the redundant links by using a depth-first search algorithm by taking an upstream business node and a downstream business node of the fault node as a starting point and an ending point, generating all candidate paths meeting constraint conditions that the time delay of the path is less than or equal to 1.5 times of the time delay of the original link, the bandwidth is more than or equal to the minimum required bandwidth of the influenced business, and the stability coefficient is more than or equal to 0.7, and recording a node sequence, a link sequence and real-time quality parameters of each link contained in each candidate path to form a candidate path set; Then, performing optimal path screening calculation, namely, constructing a multi-objective optimization function to evaluate candidate paths, wherein the minimum total path delay, the highest total stability coefficient and the lowest total bandwidth utilization are taken as optimization targets, calculating a comprehensive Score for each candidate path, wherein the formula is score=0.4× (1-T_total/T_max) +0.3×S_total/n+0.3× (1-B_used/B_total), wherein T_total is the total path delay, T_max is the maximum allowed path delay, S_total is the sum of all link stability coefficients of the paths, n is the number of path links, B_used is the occupied bandwidth of the paths, B_total is the total bandwidth of the paths, sorting the candidate paths according to the comprehensive Score from high to low, selecting the first 2 paths with the highest Score, 1 path is the main path, 1 path is the standby path, expanding the search range of the links to 3 nodes if the candidate paths are less than 2 paths, regenerating the candidate paths and screening; Finally, the sending and verification of the reconfiguration instruction are completed, a link switching instruction is generated according to the node sequence of the optimal path, the link switching instruction comprises a target node ID, a link parameter configuration value and a switching time stamp, and the reconfiguration instruction is sent to all nodes involved in the path through a low-delay control channel; after receiving the feedback of all nodes, the coordination node sends test data packet to the nodes at two ends of the path to verify the connectivity and transmission quality of the path, the test time delay is less than or equal to the service allowable time delay and less than or equal to 1% of the packet loss rate is verification passing, if the verification passes, the link states in the topology mapping table and the global topology are updated, the original fault link is marked as invalid, the new path is marked as active, and if the verification fails, the standby path is immediately started to repeat the verification flow.
  7. 7. A computer device comprising a memory, a processor and a computer program stored on the memory, characterized in that the processor executes the computer program to carry out the steps of the method of claim 1.
  8. 8. A computer readable storage medium having stored thereon a computer program/instruction which when executed by a processor performs the steps of the method of claim 1.
  9. 9. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of claim 1.

Description

Real-time discovery and dynamic update method for air-sky-sea heterogeneous network topology Technical Field The invention belongs to the technical field of communication networks, and particularly relates to a method for discovering and dynamically updating an air-sky-sea heterogeneous network topology in real time. Background With the rapid development of space-based, foundation and sea-based multidimensional communication means, an air-space-ground-sea integrated heterogeneous network gradually becomes an important component of national strategic communication and information infrastructure. The network is composed of a satellite communication system, a high-altitude balloon, an unmanned aerial vehicle platform, a ground cellular base station, a private network node, a shipborne communication terminal and other heterogeneous units, has the characteristics of wide coverage range, cross-domain interconnection, strong dynamic property and the like, and can be widely applied to the fields of emergency communication, national defense security, intelligent traffic, ocean monitoring and the like. The prior art mainly focuses on the following topology discovery and updating methods: Static configuration and a preset topology table, namely in a part of satellite network and ground private network system, the connection relation between nodes is maintained through static routing and the preset topology table. The method can better operate in a fixed network, but under the scenes of satellites and unmanned airports moving at high speed, the topology changes frequently, static configuration is quickly invalid, and network connectivity is reduced. In cellular communication and self-organizing network (MANET), the periodic topology detection method is to periodically detect through Hello message or neighbor discovery protocol, and update topology. However, in the air-space-earth-sea integrated network, due to the huge number of nodes and frequent link switching, the update period of periodic detection can not meet the real-time requirement, so that topology perception is lagged. Centralized control and topology management in recent years, SDN (software defined network) controllers have been introduced into satellite-ground converged networks for centralized state collection and unified construction of global topologies. This approach is effective in small and medium scale scenarios, but in multi-domain large scale networks, the centralized controller needs to handle a large amount of signaling and computation, causing control plane bottlenecks and excessive delay problems. The prior researches focus on a certain type of network, such as dynamic link management among satellites, cellular base station switching management, shipboard ad hoc network discovery and the like. These methods are feasible in a single domain environment, but lack unified topology discovery and updating capability of cross-domain and cross-protocol, and are difficult to support the requirement of air-to-ground Taiwan Strait Exchange Association for communication. However, the method has obvious defects in the space-sky-sea heterogeneous network scene: The real-time performance is insufficient, namely, nodes such as satellites, unmanned aerial vehicles, ships and the like move at high speed, so that the topological structure changes frequently, and the existing method generally adopts a fixed-period updating mechanism and cannot reflect the actual network state in time. The isomerism is difficult to be compatible, different domains adopt different communication protocols, frequency bands and link characteristics, most of the existing methods are designed for a single network type, and unified topology management is difficult to realize. The control and calculation costs are overlarge, and the centralized method generates huge control costs in a large-scale network scene, so that the expandability of the network is limited. The robustness is insufficient, when a link is suddenly interrupted or a node fails (such as satellite out-of-range, unmanned aerial vehicle loss and ship drift), the existing topology discovery mechanism is slow in response, and the topology connectivity cannot be recovered in time. The root cause of the above problem is that the prior art is mostly based on a static or single domain topology management concept, and lacks the capability of adapting to high-speed dynamic and strong heterogeneous environments. In attempting to solve this problem, researchers have mainly encountered the following difficulties: 1. how to realize fast state sensing and topology reconstruction of large-scale heterogeneous nodes while ensuring low overhead; 2. How to build a cross-domain unified topology representation model, and is compatible with different protocols and link conditions; 3. How to realize dynamic update with high real-time performance and stability under the condition of frequent topology change; 4. how to maintain the robustness of