Dual-mode carrier communication aggregation interference identification system and device based on graph neural network

Abstract

The application discloses a dual-mode carrier communication aggregation interference identification system and device based on a graph neural network, which comprises the following modules: a communication graph construction module, which is used for constructing a communication topology graph and configuring node dynamic attributes and disturbance direction labels; a graph neural network modeling module, which is used for constructing a multi-layer rotation perception graph neural network, perceiving structure connection relationships and direction changes, and generating an interference heat map; an interference path search module, which is used for performing jump search by using an improved fruit fly optimization algorithm and generating an interference path candidate set; a path evaluation and source point positioning module, which is used for determining an optimal interference path and an interference source position; a linkage control module, which is used for dynamically implementing interference suppression and path switching operations; and a self-learning update module, which is used for collecting communication link signal quality data, updating an interference heat map generation mechanism and a fruit fly jump strategy. The application fuses a graph neural network and an improved fruit fly optimization algorithm and accurately identifies a communication interference path.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication interference identification and control, and particularly relates to a dual-mode carrier communication aggregated interference identification system and device based on a graph neural network. BACKGROUND

[0002] Under the background that current dual-mode carrier communication is widely applied to industrial control, power transmission and intelligent sensing network, the problem of communication interference is increasingly prominent, especially in complex environments, communication links are affected by multiple nonlinear and burst electromagnetic interference sources, which seriously threatens the stability of signal transmission and the reliable reception of information. Traditional interference identification and suppression methods are mostly based on frequency domain filtering, level detection, channel equalization and other means. Such methods usually rely on preset thresholds or static rules for interference identification, and cannot adapt to the dynamic changes of interference characteristics and the structural complexity caused by multi-path signal propagation. In addition, existing methods have limited effect in dealing with typical crosstalk, harmonic coupling and frequency hopping interference in power line communication systems. Especially when the interference signal propagates in the path with sudden direction or multiple nodes in series attenuation accumulation, traditional methods are difficult to effectively depict the evolution process of the interference path, thereby affecting the subsequent suppression and compensation strategy deployment.

[0003] In recent years, the development of artificial intelligence technology, especially graph neural network (GNN), has provided a new idea for communication modeling and interference identification under structured data. Some research attempts to model communication networks as graph structures, and extracts the topological dependency between nodes through graph convolution or graph attention mechanism. However, most graph neural network models do not combine physical properties such as channel disturbance direction and path angle change, and cannot accurately model the propagation trajectory of interference signals in the graph structure. In addition, existing methods based on graph neural networks mostly focus on link quality assessment, node classification or local interference detection, lack the ability to track and identify interference paths from the overall network structure, and also fail to effectively solve the problem of interference source positioning.

[0004] In the aspect of interference path search, the current mainstream methods are mostly based on heuristic search algorithms such as genetic algorithm, ant colony optimization and simulated annealing. Although these algorithms have certain global search ability, they lack the ability to perceive communication semantics such as disturbance directionality and node heat evolution trend in the graph structure, and often can only obtain local optimal paths, and the search efficiency significantly decreases in the presence of multiple source interference and high-dimensional path space. At the same time, most of these methods do not design a strategy updating mechanism linked with neural network models, lack the ability of closed-loop feedback at the system level, and cannot dynamically adjust the search behavior according to the system running results.

[0005] In addition, the prior art still mainly adopts a fixed strategy in communication hardware control, and typical schemes are fixed bandwidth filtering and amplifier gain control based on a level threshold. Such a control strategy has poor flexibility, and it is difficult to automatically adjust a compensation loop and filter configuration according to dynamic changes of an actual interference path, resulting in disconnection between a control means and identification results, overall response lag of the system, and lack of end-to-end capability for path-level interference accurate positioning, adaptive suppression and stable recovery.

[0006] In the face of the above problems, the traditional technology has the following outstanding defects: 1. Lack of a unified modeling method that can integrate communication graph structure, channel state and interference propagation characteristics; 2. The graph neural network model cannot effectively perceive changes in interference propagation direction and multi-scale adjacency characteristics, and cannot form a deep understanding of the interference path; 3. The interference path search lacks a dynamic perception mechanism and a feedback-driven strategy evolution capability; 4. The identification results and hardware control cannot form close linkage, and the system lacks closed-loop self-learning capability.

[0007] Therefore, how to provide a dual-mode carrier communication aggregated interference identification system and device based on a graph neural network is a problem that those skilled in the art need to solve. SUMMARY

[0008] An object of the present application is to provide a dual-mode carrier communication aggregated interference identification system and device based on a graph neural network. The present application combines a multi-layer rotation perception graph neural network and an improved fruit fly optimization algorithm to construct a communication graph with rich dynamic structural features, and realizes accurate identification of an interference path and positioning of an interference source in a dual-mode carrier communication system. The system has the advantages of fine interference modeling, efficient path search, accurate source point judgment, fast control response and strong strategy adaptation capability, and can effectively improve the stability and robustness of the communication link, and is suitable for power line communication, industrial field bus and intelligent communication system interference suppression application scenarios in complex environments, and has good engineering application prospect and promotion value.

[0009] The dual-mode carrier communication aggregated interference identification system based on a graph neural network according to the embodiment of the present application comprises the following modules:

[0010] A communication graph construction module is configured to identify various nodes and communication links in a dual-mode communication system through a channel detection device, construct a communication topology graph, and configure node dynamic attributes and disturbance direction labels.

[0011] A graph neural network modeling module is configured to construct a multi-layer rotation perception graph neural network, perceive structural connection relationships and direction changes, output node interference heat values, and generate an interference heat map.

[0012] The interference path searching module is configured to initialize a plurality of fruit fly individuals in the high-heat area of the heat map, perform a jump search by using the improved fruit fly optimization algorithm, and generate an interference path candidate set;

[0013] The path evaluation and source point positioning module is configured to evaluate the heat coverage and propagation rationality of the interference path candidate set, and determine the optimal interference path and the interference source position.

[0014] The linkage control module is configured to control the filter, the compensation loop and the buffer amplifier hardware device, and dynamically implement the interference suppression and path switching operation.

[0015] The self-learning updating module is configured to collect the communication link signal quality data, and feed back to the multi-layer rotating perception graph neural network and the improved fruit fly optimization algorithm, and update the interference heat map generation mechanism and the fruit fly jump strategy.

[0016] The interference path searching module is configured to initialize a plurality of fruit fly individuals in the high-heat area of the heat map, perform a jump search by using the improved fruit fly optimization algorithm, and generate an interference path candidate set;

[0017] S1, constructing a communication topology graph by using a channel detection device, setting dynamic attributes and disturbance direction labels, and forming a communication graph;

[0018] S2, constructing a multi-layer rotating perception graph neural network, inputting the communication graph into the multi-layer rotating perception graph neural network, perceiving the connection relationship and direction change in the communication graph, extracting multi-scale disturbance propagation features, and outputting an interference heat map;

[0019] S3, initializing a plurality of fruit fly individuals in the high-heat area of the interference heat map by using the improved fruit fly optimization algorithm, performing a jump search according to the heat value, the path direction and the disturbance feature, and generating an interference path candidate set;

[0020] S4, performing path evaluation on the interference path candidate set according to the interference path heat coverage and the propagation rationality index, and determining the optimal interference path and the interference source position;

[0021] S5, controlling the communication hardware in linkage according to the optimal interference path and the interference source position, dynamically adjusting the compensation loop and the buffer amplifier, and realizing interference suppression and communication link stable maintenance;

[0022] S6, collecting real-time communication link signal quality data, and feeding back to the multi-layer rotating perception graph neural network and the improved fruit fly optimization algorithm, and updating the interference heat map generation mechanism and the fruit fly jump strategy.

[0023] Optionally, the S1 specifically includes:

[0024] S11, initializing scanning of the dual-mode communication system by using the channel detection device, identifying all terminal nodes, relay nodes and communication coupling devices, and constructing a node set;

[0025] S12, identifying an effective communication link between each node according to a physical connection topology relationship, representing the effective communication link between each pair of nodes as an edge to form a communication topology graph;

[0026] S13, collecting and labeling channel parameters on each edge of the communication topology graph, including real-time instantaneous attenuation, interference intensity, frequency stability, and path propagation delay;

[0027] S14, configuring a dynamic working feature of each node in the communication process, the dynamic working feature including a node real-time voltage, a current, a harmonic amplitude index, and a signal-to-noise ratio of a connected path, for reflecting an operation state of the node and a sensitivity degree to interference;

[0028] S15, further defining a perturbation direction label on each edge, for quantifying a direction change of an interference signal on a propagation path, recording a perturbation direction angle as the perturbation direction label:

[0029]

[0030] wherein θ ij represents a perturbation direction angle of the edge e ij , represents a direction estimation of an interference electric field vector on the edge e ij at a time t+Δt, represents a direction estimation of an interference electric field vector on the edge e ij at a time t, ||·|| represents a norm, represents a spatial position vector of the jth node, represents a spatial position vector of the ith node, ∈ represents a smoothing factor, λ represents a weight coefficient, S ij represents a channel stability index on the edge e ij , |·| represents an absolute value operation;

[0031] S16, obtaining a communication graph containing a node physical state, an edge channel attribute, and a perturbation direction label.

[0032] Optionally, the S2 specifically includes:

[0033] S21, constructing a multi-layer rotation perception graph neural network, the multi-layer rotation perception graph neural network being composed of L layers of graph convolution units, each graph convolution unit containing an adjacency perception module and a direction perception module, the adjacency perception module being used to extract a structural connection relationship in a graph, and the direction perception module being used to capture an edge direction change and a perturbation propagation angle feature;

[0034] S22, initializing a node attribute set of the communication graph as an input feature matrix X (0), wherein each row represents the original dynamic characteristics of a node, input into the first layer graph convolution unit;

[0035] S23, in the first layer graph convolution unit, set the neighbor set of node v i as N(v i ), and update the node characteristics using a rotation-aware convolution operator combined with the disturbance direction:

[0036]

[0037] , wherein, represents the feature vector of the i-th node in the l-th layer, σ represents the activation function, W (l) represents the adjacent perception module weight matrix, represents the feature vector of the j-th node in the l-1-th layer, γ (l) represents the direction adjustment coefficient, θ ij represents the disturbance direction angle of edge e ij , D ij represents the direction embedding vector of the edge, α ij represents the disturbance intensity normalization coefficient;

[0038] S24, in each layer, the direction perception module is used to capture the edge direction change and the disturbance propagation angle characteristics, and is updated, and finally the node feature matrix X (L) of the L-th layer is output;

[0039] S25, the node features output by multiple network layers are spliced and weighted aggregated to form a final fusion feature matrix, capturing the interference propagation patterns at different scales;

[0040] S26, based on the final fusion feature matrix, the interference heat value of each node is obtained by using full connection mapping and normalization operation, and an interference heat map is generated, which is used to represent the possibility of each node in the communication graph being disturbed.

[0041] Optionally, the S3 specifically comprises:

[0042] S31, based on the interference heat map, a node set with a heat value higher than a threshold τ h is extracted as a candidate interference source area, and a plurality of fruit fly individuals are initialized in the candidate interference source area, each fruit fly individual corresponding to a search starting point of a potential interference path;

[0043] S32, an improved fruit fly optimization algorithm is used, and the improvement is that a multi-factor perception function is introduced to replace the search mechanism based on single value judgment of the objective function:

[0044]

[0045] wherein, F krepresents the kth fruit fly individual, J(F k ) represents the comprehensive perception function value calculated by the kth fruit fly individual in the current jump decision, η represents the jump amplitude adjustment factor, h k represents the interference heat of the node where the kth fruit fly individual is currently located, ρ kj represents the consistency coefficient of the current path direction and the historical disturbance direction, I kj represents the interference intensity, SNR kj represents the signal-to-noise ratio, θ kj represents the disturbance direction angle of the edge e kj .

[0046] S33, the fruit fly individual performs local neighborhood selection based on the multi-factor perception function, performs a jump search operation, and jumps to a neighboring node that maximizes J(F k ), while updating the trajectory sequence and direction history information of the fruit fly individual;

[0047] S34, to overcome the problem of slow path convergence or falling into the hotspot center of the traditional fruit fly algorithm in high-dimensional graph structure, an adaptive disturbance adjustment strategy is set, when the fruit fly individual appears heat value shock or disturbance direction sharp reversal in continuous jump, the direction disturbance update mechanism is triggered:

[0048]

[0049] wherein, represents the current direction perception variable of the individual, α represents the disturbance amplitude adjustment factor, and rand(·) represents a random number generation function;

[0050] S35, the jump trajectory of each fruit fly individual is recorded in sequence as a path set, and the path sequences of all individuals constitute an interference path candidate set.

[0051] Optionally, the S4 specifically comprises:

[0052] S41, each interference path in the interference path candidate set P = {P1, P2, …, P N} is evaluated, wherein P i represents the i th interference path, N represents the total number of interference paths, and the interference heat coverage and disturbance propagation rationality scores are calculated for each interference path respectively;

[0053] S42, the interference heat coverage is defined as the normalized mean value of the interference heat values of each node on the path:

[0054]

[0055] wherein, Ω i represents the interference heat coverage of the i th interference path, |P i| represents the length of the i-th interference path, h j represents the interference hotness value of node v j on the interference path, max(h j ) represents the maximum interference hotness value of all nodes in the i-th interference path;

[0056] S43, comprehensively evaluate the path direction continuity, consistency and disturbance intensity change trend, calculate the disturbance propagation rationality score:

[0057]

[0058] wherein, Ψ i represents the disturbance propagation rationality score of the i-th interference path, θ mn represents the disturbance direction angle of edge e mn , I mn represents the interference intensity of edge e mn , δf mn represents the frequency drift;

[0059] S44, normalize the interference hotness coverage and the disturbance propagation rationality score to combine them into a path comprehensive score function:

[0060] Φ i = β1·Ω i + β2·Ψ i ;

[0061] wherein, Φ i represents the comprehensive score of the i-th interference path, β1 and β2 represent weight coefficients;

[0062] S45, sort all interference paths P i according to the comprehensive score, select the interference path with the highest comprehensive score as the optimal interference path, and mark the starting node of the optimal interference path as the current interference source position;

[0063] S46, output the optimal interference path and the interference source position.

[0064] Optionally, the S5 specifically comprises:

[0065] S51, start the hardware linkage control interface according to the optimal interference path and the interference source position, send control instructions to the hardware modules of the communication circuit through the hardware control bus, and activate the filters, compensation loops and buffer amplifiers along the interference path;

[0066] S52, based on the interference signal spectrum characteristics on the optimal interference path, select the filter parameters corresponding to the interference frequency band, and suppress the interference signals in the interference frequency band from entering the communication receiving unit;

[0067] S53, real-time monitoring of the optimal interference path on the node voltage, current and harmonic component data, according to the node impedance characteristics, automatically adjust the compensation circuit capacitance and inductance unit, meet the impedance balance requirements in the process of signal transmission, inhibit signal reflection and channel distortion;

[0068] S54, start feedback compensation circuit, measure the signal phase and amplitude characteristics in the communication link, dynamically adjust the transistor, switch diode and phase shift circuit in the compensation circuit, real-time generate compensation signal with similar amplitude and opposite phase of the interference signal;

[0069] S55, real-time acquisition of signal quality data on the optimal interference path, when the communication signal level is found to be reduced, waveform distortion or bit error rate is increased, automatically trigger the buffer amplifier adjustment program, restore the disturbed low-level communication signal by adjusting the amplifier gain parameter, and perform waveform repair and signal standardization output;

[0070] S56, if the communication link quality does not reach the preset index after the interference suppression control of steps S51-S55, automatically execute the communication path switching mechanism according to the pre-defined alternative communication path data, and switch to the sub-optimal path for signal transmission.

[0071] Optionally, the S6 specifically comprises:

[0072] S61, real-time acquisition of communication link signal quality data after hardware linkage control, and establishment of real-time communication quality database;

[0073] S62, using the latest data in the real-time communication quality database, calculating the actual interference response deviation value of the communication node, and comparing it with the heat value of the corresponding node in the interference heat map generated in step S2, to obtain a node heat error set;

[0074] S63, according to the node heat error set, dynamically adjusting the adjacent perception module weight matrix and direction adjustment coefficient of the multi-layer rotating perception graph neural network through the back propagation mechanism;

[0075] S64, analyzing the path disturbance direction change characteristics in the real-time communication quality data, calculating the deviation between the actual direction angle change of path propagation and the disturbance direction angle in the fruit fly jumping search process, adjusting the direction perception variable and jumping amplitude adjustment factor in the fruit fly individual jumping strategy based on the direction deviation, and dynamically updating and improving the multi-factor perception function J(F k ) of the fruit fly optimization algorithm;

[0076] S65, according to the updated multi-layer rotating perception graph neural network parameters and improved fruit fly optimization algorithm strategy, re-executing steps S2 and S3 to generate a new interference heat map and interference path candidate set, realizing self-learning and strategy iteration ability.

[0077] Optionally, the communication link signal quality data includes node signal-to-noise ratio, bit error rate, interference strength, voltage and current variation.

[0078] The dual-mode carrier communication aggregated interference identification device based on the graph neural network according to the embodiment of the application is used for executing each module in the dual-mode carrier communication aggregated interference identification system based on the graph neural network, and realizing communication graph construction, graph neural network modeling, interference path search, path evaluation and source point positioning, linkage control and self-learning update operation in a hardware or software manner, so that the identification and suppression of the interference path in the dual-mode carrier communication environment are completed.

[0079] The beneficial effects of the application are:

[0080] Firstly, the application realizes high-fidelity modeling of the communication structure in a complex electromagnetic interference environment by constructing a communication topology graph containing dynamic node attributes, channel states and disturbance direction labels. Compared with the traditional modeling method relying on static topology or channel threshold, the application introduces key features such as disturbance propagation direction angle at the node and edge level, which can more comprehensively depict the propagation trajectory of the interference signal in the physical path, providing an accurate structural basis for subsequent identification.

[0081] Secondly, relying on the multi-layer rotating perception graph neural network model, the application realizes deep perception of the connection relationship and direction change in the communication graph. By setting the adjacent perception module and the direction perception module, the model can fuse the topological information between nodes and the disturbance angle features of edges at multiple scales, thereby more effectively extracting the pattern features of interference propagation. The output interference heat map not only reflects the possibility of each node being disturbed, but also provides accurate heat zone positioning basis for the path identification process, significantly improving the focusing ability of the area around the interference source.

[0082] In addition, in terms of path search mechanism, the application introduces an improved fruit fly optimization algorithm, which comprehensively considers node heat value, direction consistency, channel interference strength and signal-to-noise ratio and other information through a multi-factor perception function, and constructs a jump search strategy with direction guidance and environmental adaptability. This method overcomes the problems of slow convergence and easy falling into local optimum of traditional heuristic algorithms, so that the search individual can quickly converge to the real interference path in the complex graph structure. The heat coverage and propagation rationality scores are fused in the path evaluation process, and the global optimality and physical rationality of the identification result are ensured through the comprehensive evaluation mechanism, effectively improving the accuracy of path identification.

[0083] More importantly, the system of the application closely couples the recognition result with the hardware response process through a linkage control module, realizing dynamic regulation and control of devices such as filters, compensation circuits and buffer amplifiers. This linkage control capability not only enables targeted suppression measures to be performed on different types of interference signals, but also supports automatic switching to a suboptimal path when communication quality decreases, ensuring continuous and stable operation of the communication link. Finally, with the help of a self-learning update module, the system can continuously collect communication quality data, dynamically update the graph neural network model parameters and fruit fly jumping strategy, gradually evolve the recognition model and control strategy, and realize rapid adaptation and response to new interference modes, thereby building an intelligent communication interference recognition platform with learning ability. BRIEF DESCRIPTION OF DRAWINGS

[0084] The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings:

[0085] Figure 1 Structure diagram of the dual-mode carrier communication aggregated interference recognition system based on a graph neural network proposed by the application;

[0086] Figure 2 Overall flowchart of the dual-mode carrier communication aggregated interference recognition method based on a graph neural network proposed by the application. DETAILED DESCRIPTION

[0087] The application will now be described in further detail, by way of example only, with reference to the accompanying drawings. These drawings are not drawn to scale and are included herewith to further explain various aspects of the application, like the basic structure of the application. In the drawings:

[0088] Reference Figure 1 The dual-mode carrier communication aggregated interference recognition system based on a graph neural network comprises the following modules:

[0089] A communication graph construction module is configured to identify various nodes and communication links in a dual-mode communication system through a channel detection device, construct a communication topology graph, and configure node dynamic attributes and disturbance direction labels.

[0090] A graph neural network modeling module is configured to construct a multi-layer rotation perception graph neural network, perceive structure connection relationships and direction changes, output node interference heat values, and generate an interference heat map.

[0091] An interference path search module is configured to initialize multiple fruit fly individuals in a high heat area of the heat map, perform jumping search using an improved fruit fly optimization algorithm, and generate an interference path candidate set.

[0092] A path evaluation and source location module is configured to evaluate the heat coverage and propagation rationality of the interference path candidate set, and determine the optimal interference path and the location of the interference source.

[0093] A linkage control module is configured to control the filter, compensation loop and buffer amplifier hardware devices, and dynamically implement interference suppression and path switching operations.

[0094] A self-learning update module is configured to collect communication link signal quality data, and feed back to the multi-layer rotating perception graph neural network and the improved fruit fly optimization algorithm, to update the interference heat map generation mechanism and the fruit fly jumping strategy.

[0095] The system integrates the six modules of communication graph construction, graph neural network modeling, path search, path evaluation, linkage control and self-learning update in the same system architecture, and constructs a dual-mode carrier communication interference identification system with complete functions, clear structure and closed-loop iterative optimization. In the graph modeling stage, the dynamic characteristics and disturbance direction of the nodes are fully considered, and the multi-layer graph neural network and the improved fruit fly optimization algorithm are combined to realize the full-process automation from interference heat map generation to optimal path identification and accurate positioning of the interference source, significantly improving the identification accuracy and system adaptability. Especially in the face of high noise, frequency drift and complex path interference, the system can autonomously optimize the identification strategy and control mechanism based on communication data, and has high deployability and engineering application value.

[0096] Reference Figure 2 The dual-mode carrier communication aggregated interference identification method based on the graph neural network comprises the following steps:

[0097] S1. Construct a communication topology graph through a channel detection device, and set dynamic attributes and disturbance direction labels to form a communication graph.

[0098] S2. Construct a multi-layer rotating perception graph neural network, and input the communication graph into the multi-layer rotating perception graph neural network, perceive the connection relationship and direction change in the communication graph, extract multi-scale disturbance propagation features, and output an interference heat map.

[0099] S3. Initialize multiple fruit fly individuals in the high-heat area of the interference heat map using the improved fruit fly optimization algorithm, perform jumping search according to the heat value, path direction and disturbance characteristics, and generate an interference path candidate set.

[0100] S4. Perform path evaluation on the interference path candidate set according to the interference path heat coverage and propagation rationality indexes, and determine the optimal interference path and the location of the interference source.

[0101] S5. Linkage control of the communication hardware according to the optimal interference path and the location of the interference source, dynamic adjustment of the compensation loop and the buffer amplifier, realization of interference suppression and stable maintenance of the communication link.

[0102] S6, real-time collection of communication link signal quality data and feedback to the multi-layer rotating perception graph neural network and improved fruit fly optimization algorithm for updating the interference heat map generation mechanism and fruit fly jumping strategy.

[0103] The method clearly defines the cooperative operation logic between each functional module, and builds an end-to-end interference identification and response method process. From the construction of the communication graph, through the graph neural network perception and heat map output, fruit fly optimization search path, path score evaluation, and linkage hardware control, to the final self-learning feedback update, a complete closed-loop execution chain is formed. This method not only improves the real-time and traceability of interference identification, but also highly integrates identification and control, enabling the system to have continuous operation capabilities from detection to decision-making and response, greatly enhancing the system's autonomous adjustment and stable maintenance performance in complex and dynamic interference environments.

[0104] In this embodiment, S1 specifically includes:

[0105] S11, initializing scanning of the dual-mode communication system by the channel detection device, identifying all terminal nodes, relay nodes and communication coupling devices, and constructing a node set;

[0106] S12, identifying effective communication links between each node according to the physical connection topology relationship, representing each pair of nodes' effective communication link as an edge, and forming a communication topology graph;

[0107] S13, collecting and labeling channel parameters on each edge of the communication topology graph, including real-time instantaneous attenuation, interference strength, frequency stability and path propagation delay;

[0108] S14, configuring dynamic working characteristics of each node in the communication process, including node real-time voltage, current, harmonic amplitude index and signal-to-noise ratio of the connected path, for reflecting the running state of the node and the sensitivity to interference;

[0109] S15, further defining a disturbance direction label on each edge for quantifying the direction change of the interference signal on the propagation path, and recording the disturbance direction angle as the disturbance direction label:

[0110]

[0111] where θ ij represents the disturbance direction angle of edge e ij , represents the direction estimate of the interference electric field vector on edge e ij at time t+Δt, represents the direction estimate of the interference electric field vector on edge e ijdirection estimation on the upper, ||·|| represents the norm, represents the spatial position vector of the jth node, represents the spatial position vector of the ith node, ∈ represents the smoothing factor, λ represents the weight coefficient, S ij represents the channel stability index on the edge e ij , |·| represents the absolute value operation;

[0112] S16, obtaining a communication graph containing node physical state, edge channel attribute and disturbance direction label.

[0113] Steps S11-S16 systematically refine the communication graph construction step, define node set, edge set, channel parameter and disturbance direction angle label, and construct a structured communication graph with dynamic properties and direction propagation characteristics, providing high-quality data support for subsequent graph neural network model training and interference path modeling. Unlike traditional methods that only use static topology graph or channel strength graph, this communication graph introduces multi-dimensional attributes such as disturbance direction angle, frequency stability and path propagation delay in structure, which can accurately express the space-time trajectory of interference propagation in the system, effectively improving the model's ability to identify direction mutation interference paths.

[0114] In this embodiment, S2 specifically includes:

[0115] S21, constructing a multi-layer rotation perception graph neural network, the multi-layer rotation perception graph neural network is composed of L layers of graph convolution units, each graph convolution unit includes an adjacency perception module and a direction perception module, the adjacency perception module is used to extract the structural connection relationship in the graph, and the direction perception module is used to capture the edge direction change and disturbance propagation angle characteristics;

[0116] S22, initializing the node attribute set of the communication graph as an input feature matrix X (0) , wherein each row represents the original dynamic characteristics of a node, and is input into the first layer graph convolution unit;

[0117] S23, in the lth layer graph convolution unit, let the neighbor set of node v i be N(v i ), and update the node characteristics by using a rotation perception convolution operator combined with the disturbance direction:

[0118]

[0119] wherein, represents the feature vector of the ith node in the lth layer, σ represents the activation function, W (l) represents the adjacency perception module weight matrix, represents the feature vector of the jth node in the l-1th layer, γ (l)denotes a direction adjustment coefficient, θ ij denotes a disturbance direction angle, D ij ij denotes a direction embedding vector of an edge, a ij denotes a disturbance intensity normalization coefficient;

[0120] S24, in each layer, a direction perception module is used to capture edge direction change and disturbance propagation angle characteristics, and is updated, and finally a node feature matrix X of the Lth layer is output (L) ;

[0121] S25, the node features output by multiple network layers are spliced and weighted aggregated to form a final fusion feature matrix, capturing the interference propagation pattern at different scales;

[0122] S26, based on the final fusion feature matrix, a full connection mapping and a normalization operation are used to obtain the interference heat value of each node, and an interference heat map is generated, which is used to represent the possibility of each node being disturbed in the communication graph.

[0123] The above steps introduce a multi-layer rotation perception graph neural network structure at the graph neural network modeling level, innovatively combine the adjacency perception and direction perception modules, and simultaneously consider the topological connectivity and edge disturbance direction change characteristics in the node feature extraction process. The model can fully fuse multi-scale neighborhood information, improve the modeling ability of direction deflection, signal attenuation and propagation pattern evolution in the multi-hop interference path. The interference heat map output by the model has higher accuracy and more reasonable hot area determination, providing more reliable initial guide data for the subsequent fruit fly search algorithm, and significantly improving the upstream and downstream collaborative performance of the overall recognition system.

[0124] In the embodiment, the S3 specifically includes:

[0125] S31, based on the interference heat map, a node set with a heat value higher than a threshold τ h is extracted as a candidate interference source area, and a plurality of fruit fly individuals are initialized in the candidate interference source area, each fruit fly individual corresponding to a search starting point of a potential interference path;

[0126] S32, an improved fruit fly optimization algorithm is used, and the improvement lies in introducing a multi-factor perception function to replace the search mechanism based on the single value judgment of the objective function:

[0127]

[0128] wherein, F k represents the kth fruit fly individual, J(F k ) represents the comprehensive perception function value calculated by the kth fruit fly individual in the current jump decision, η represents a jump amplitude adjustment factor, h k ​represents the interference heat of the kth fruit fly individual at the current node, p kj represents the consistency coefficient of the current path direction and the historical disturbance direction, I kj represents the interference intensity, SNR kj represents the signal-to-noise ratio, θ kj represents the disturbance direction angle of the edge e kj .

[0129] S33, the fruit fly individual performs local neighborhood selection based on the multi-factor perception function, performs a jump search operation, and jumps to the adjacent node that maximizes J(F k ) while updating the trajectory sequence and direction history information of the fruit fly individual;

[0130] S34, to overcome the problem of slow path convergence or falling into the hotspot center in the high-dimensional graph structure of the traditional fruit fly algorithm, an adaptive disturbance adjustment strategy is set, when the fruit fly individual appears heat value shock or disturbance direction sharp reversal in continuous jump, the direction disturbance update mechanism is triggered:

[0131]

[0132] wherein, represents the current direction perception variable of the individual, a represents the disturbance amplitude adjustment factor, and rand(·) represents the random number generation function;

[0133] S35, the jump trajectory of each fruit fly individual is recorded in sequence as a path set, and the path sequences of all individuals constitute an interference path candidate set.

[0134] By improving the fruit fly optimization algorithm, a multi-factor perception function is introduced to integrate key features such as heat value, path direction consistency, interference intensity, and signal-to-noise ratio, and a more efficient and direction-guided jump path search mechanism is realized. Compared with the traditional search algorithm, this method not only avoids falling into local optimum, but also improves the path diversity and convergence speed by introducing the disturbance adjustment mechanism. The fruit fly individual search process is deeply combined with the communication heat map, effectively realizing the rapid locking of the real interference propagation path in the graph structure, and providing a more feasible candidate set for subsequent path evaluation.

[0135] In the embodiment, the S4 specifically comprises:

[0136] S41, each interference path in the interference path candidate set P = {P1, P2, …, P N} is evaluated, wherein P i represents the ith interference path, and N represents the total number of interference paths. The interference heat coverage and disturbance propagation rationality scores are calculated for each interference path respectively;

[0137] S42, the interference heat coverage is defined as the normalized mean of the interference heat values of each node on the path:

[0138]

[0139] wherein Ω i represents the interference heat coverage of the i-th interference path, |P i | represents the length of the i-th interference path, h j represents the interference heat value of node v j on the interference path, and max(h j ) represents the maximum interference heat value of all nodes in the i-th interference path.

[0140] S43, the path direction continuity, consistency and disturbance intensity change trend are comprehensively evaluated, and the disturbance propagation rationality score is calculated:

[0141]

[0142] wherein Ψ i represents the disturbance propagation rationality score of the i-th interference path, θ mn represents the disturbance direction angle of edge e mn , I mn represents the interference intensity of edge e mn , and δf mn represents the frequency drift.

[0143] S44, the interference heat coverage and the disturbance propagation rationality score are normalized and combined into a path comprehensive score function:

[0144] Φ i = β1·Ω i + β2·Ψ i ;

[0145] wherein Φ i represents the comprehensive score of the i-th interference path, and β1 and β2 represent weight coefficients.

[0146] S45, all interference paths P i are sorted according to the comprehensive score, the interference path with the highest comprehensive score is selected as the optimal interference path, and the starting node of the optimal interference path is recorded as the current interference source position.

[0147] S46, the optimal interference path and the interference source position are output.

[0148] The above steps propose a path evaluation mechanism based on interference heat coverage and propagation rationality score, which can scientifically quantify the advantages and disadvantages of different candidate paths in terms of interference intensity, direction consistency and structural coherence, form a comprehensive score function through weighted combination, and select the optimal interference path. Unlike traditional algorithms that target the shortest path or the path with the maximum local heat, this method considers the global quality and propagation rationality of the path, making the selected path not only match the actual logic of interference propagation, but also facilitate subsequent suppression control. The introduction of this step significantly improves the stability and reliability of the interference identification results.

[0149] In the present embodiment, the S5 specifically includes:

[0150] S51, according to the optimal interference path and the position of the interference source, starting the hardware linkage control interface, sending control instructions to the hardware modules of the communication circuit through the hardware control bus, and activating the filters, compensation circuits and buffer amplifiers along the interference path;

[0151] S52, based on the interference signal spectrum characteristics on the optimal interference path, selecting the filter parameters corresponding to the interference frequency band to suppress the interference signals in the interference frequency band from entering the communication receiving unit;

[0152] S53, real-time monitoring of node voltage, current and harmonic component data on the optimal interference path, automatically adjusting the capacitance and inductance units of the compensation circuit according to the node impedance characteristics, meeting the impedance balance requirements in the signal transmission process, and suppressing signal reflection and channel distortion;

[0153] S54, starting the feedback compensation circuit, measuring the signal phase and amplitude characteristics in the communication link, dynamically adjusting the transistors, switch diodes and phase shift circuits in the compensation circuit, and real-time generating compensation signals with similar amplitude and opposite phase to the interference signals;

[0154] S55, real-time acquisition of signal quality data on the optimal interference path, when the communication signal level drops, the waveform distorts or the bit error rate rises, automatically triggering the buffer amplifier adjustment program, adjusting the amplifier gain parameters to restore the disturbed low-level communication signal, and performing waveform repair and signal standardization output;

[0155] S56, if the communication link quality does not meet the preset indicators after the interference suppression control of steps S51-S55, automatically executing the communication path switching mechanism according to the pre-defined alternative communication path data, and switching to the sub-optimal path for signal transmission.

[0156] The above steps propose a linkage control mechanism based on optimal interference path and source point information. The system can dynamically adjust the filter frequency band parameters, compensate the capacitance and inductance values of the loop, and configure the gain of the buffer amplifier, forming a directional suppression control capability for the interference path. The system can real-time perceive the interference situation in the communication link, and automatically implement interference cancellation and waveform repair operations based on the current path state. If the path control fails, the communication path switching mechanism can also be triggered to ensure the overall communication continuity of the system.

[0157] In the embodiment, S6 specifically includes:

[0158] S61, real-time acquisition of communication link signal quality data after hardware linkage control, and establishment of real-time communication quality database;

[0159] S62, using the latest data in the real-time communication quality database, calculating the actual interference response deviation value of the communication node, and comparing it with the heat value of the corresponding node in the interference heat map generated in step S2, to obtain a node heat error set;

[0160] S63, according to the node heat error set, dynamically adjusting the adjacent perception module weight matrix and direction adjustment coefficient of the multi-layer rotating perception graph neural network through the back propagation mechanism;

[0161] S64, analyzing the path disturbance direction change characteristics in the real-time communication quality data, calculating the deviation between the actual direction angle change of the path propagation and the disturbance direction angle in the fruit fly jump search process, adjusting the direction perception variable and jump amplitude adjustment factor in the fruit fly individual jump strategy based on the direction deviation, and dynamically updating and improving the multi-factor perception function J(F k ) of the fruit fly optimization algorithm;

[0162] S65, according to the updated multi-layer rotating perception graph neural network parameters and the improved fruit fly optimization algorithm strategy, re-executing steps S2 and S3 to generate a new interference heat map and interference path candidate set, realizing the self-learning and strategy iteration capability.

[0163] The above steps establish a system self-learning feedback mechanism, so that the communication quality data is not only used for performance monitoring, but also as a dynamic input source for model optimization. By feeding back the node heat prediction error to the graph neural network and adjusting the fruit fly jump strategy based on the direction deviation, the system realizes the functional leap from static reasoning to dynamic evolution. This mechanism ensures that the system can adapt to new interference modes, path changes and network topology adjustments in the long-term operation, truly realizes the evolution ability of the interference identification strategy, and significantly enhances the robustness and generalization ability of the system in complex scenarios.

[0164] In the embodiment, the communication link signal quality data includes node signal-to-noise ratio, bit error rate, interference strength, voltage and current change.

[0165] The dual-mode carrier communication aggregation interference identification device based on a graph neural network is used for executing each module in the dual-mode carrier communication aggregation interference identification system based on a graph neural network, and realizing communication graph construction, graph neural network modeling, interference path searching, path evaluation and source point positioning, linkage control and self-learning updating operation in a hardware or software manner, so that the identification and suppression of the interference path in the dual-mode carrier communication environment are completed.

[0166] Embodiment 1:

[0167] In order to verify the feasibility of the application in implementation, the application is applied to the communication network of a certain power substation. The site is located in the eastern region and is an important node of regional load scheduling, and adopts a dual-mode carrier communication mode for remote data acquisition and control instruction issuing. Due to the complexity of field devices, cable density and frequent operation of high-voltage devices, the communication link is frequently disturbed, including harmonic disturbance from inverters, electromagnetic transient pulses from switching devices, and occasional crosstalk signals, so that the communication reliability fluctuates obviously, and the traditional interference processing method based on fixed threshold and rule base cannot effectively identify the interference path, and it is more difficult to locate the interference source point in time, causing frequent abnormal alarms of the system, high data delay rate, and serious influence on the scheduling stability of the station control system.

[0168] In the field, we deployed the complete identification and control process of the system. First, the channel detection device is connected to the dual-mode communication network connected to the main control layer and primary equipment of the substation, real-time acquisition of the connection relationship and dynamic operation parameters between each terminal node, relay device and communication coupler, and construction of a communication topology graph with node voltage, current, harmonic amplitude and other information. At the same time, the disturbance direction label is introduced on each path edge in the graph, which is used to describe the propagation trajectory characteristics of signal direction jump under the condition of multiple source interference. The topology graph is then input into the multi-layer rotating perception graph neural network of the system, and the network model extracts the connection mode and disturbance propagation direction relationship in the graph structure, and outputs the interference heat map of the whole station. Through the visual interface, the area where the interference may occur or will occur can be directly identified.

[0169] On the interference heat map, the system automatically identifies multiple heat aggregation areas, initializes fruit fly individuals in these areas, and performs a jump path search based on heat values, direction continuity, signal-to-noise ratio changes, and other factors. In a packet loss event of a sudden communication, the system quickly locates the link between the No. 1 switch cabinet and the control room main communication machine as a high-heat path, and determines that the interference heat coverage of the path is 0.78 and the disturbance propagation rationality score is 0.85 through the path evaluation module. The comprehensive score is the highest among all paths, and the system determines it as the main interference path and marks the starting point as the interference source candidate area. Then, the system linkage control module is started, the filter parameters on the path are dynamically adjusted, the interference frequency band is set to 8-20 kHz, the inductance group and the capacitor group in the compensation loop are adjusted to meet the current channel impedance balance requirement, and the system also automatically activates the buffer amplifier to repair the signal distortion caused by interference. After one round of adjustment, the communication signal-to-noise ratio is increased from 19.2 dB to 31.5 dB, the bit error rate is reduced from 1.02x10 -3 to 4.37x10 -5 , and the communication link is restored to stable operation.

[0170] More importantly, the system feeds the communication quality change data collected in the above process to the graph neural network model and the fruit fly optimization algorithm module in real time, automatically corrects the heat prediction error of part of the path, and updates the jump strategy in the next round of interference identification process, so that the path identification time is shortened from the initial 2.8 seconds to 1.6 seconds, and the interference source positioning accuracy is improved to 94.3%. Through continuous 7-day on-site deployment data analysis, it is shown that after the system is enabled, the number of communication system abnormal alarm times is reduced by 71.5%, the data delay rate is reduced by an average of 48.9%, and in the extreme case, it is compressed from more than 3 seconds to less than 1 second, and the overall operation stability of the system is significantly improved.

[0171] It is verified through the embodiment that the system can not only accurately identify the interference path in the carrier communication link in a complex environment, but also link the physical control equipment to implement dynamic interference suppression, and has self-learning ability, which can continuously optimize the identification strategy during operation. It is especially suitable for deployment in power systems, rail transit, industrial sites and other electromagnetic interference complex but highly reliable communication scenarios.

[0172] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can make equivalent substitutions or changes to the technical solutions and inventive concepts of the present application within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A dual-mode carrier communication aggregation interference identification system based on graph neural network, characterized by: Includes the following modules: The communication graph construction module is used to identify various nodes and communication links in the dual-mode communication system through channel detection equipment, build a communication topology graph, and configure node dynamic attributes and disturbance direction labels; The graph neural network modeling module is used to build a multi-layer rotation-aware graph neural network, perceive the connection relationship and direction changes in the communication graph, output the node interference heat value, and generate an interference heat map; The interference path search module is used to initialize multiple fruit fly individuals in the high-temperature area of ​​the heat map, and use the improved fruit fly optimization algorithm to perform a jump search to generate a candidate set of interference paths; The path evaluation and source location module is used to evaluate the heat coverage and propagation rationality of the interference path candidate set and determine the optimal interference path and interference source location; A linkage control module is used to control the filter, compensation loop, and buffer amplifier hardware devices to dynamically implement interference suppression and path switching operations; The self-learning update module is used to collect communication link signal quality data and feed it back to the multi-layer rotation perception graph neural network and improved fruit fly optimization algorithm to update the interference heat map generation mechanism and fruit fly jumping strategy.

2. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 1 is characterized in that: The modules are implemented as follows: S1. Build a communication topology map through the channel detection device and set dynamic attributes and disturbance direction labels to form a communication map; S2. Construct a multi-layer rotation perception graph neural network, and input the communication graph into the multi-layer rotation perception graph neural network, perceive the connection relationship and direction change in the communication graph, extract multi-scale disturbance propagation characteristics, and output an interference heat map; S3. Initialize multiple fruit fly individuals in the high-heat area of ​​the interference heat map using the improved fruit fly optimization algorithm, perform a jump search based on the heat value, path direction, and disturbance characteristics, and generate a candidate set of interference paths; S4. Based on the interference path heat coverage and propagation rationality index, the interference path candidate set is evaluated to determine the optimal interference path and interference source location; S5. Based on the optimal interference path and the location of the interference source, the communication hardware is controlled and the compensation loop and the buffer amplifier are dynamically adjusted to achieve interference suppression and maintain a stable communication link. S6. Collect communication link signal quality data in real time and feed it back to the multi-layer rotation perception graph neural network and improved fruit fly optimization algorithm to update the interference heat map generation mechanism and fruit fly jumping strategy.

3. The dual-mode carrier communication aggregation interference identification system and device based on graph neural network according to claim 2 is characterized in that: Said S1 specifically includes: S11. Initialize scanning of the dual-mode communication system through a channel detection device, identify all terminal nodes, relay nodes, and communication coupling devices, and construct a node set; S12. Identify valid communication links between nodes based on the physical connection topology relationship, represent the valid communication links between each pair of nodes as edges, and form a communication topology graph; S13. Collect and mark channel parameters on each edge of the communication topology graph, including real-time instantaneous attenuation, interference intensity, frequency stability, and path propagation delay; S14. Configuring dynamic operating characteristics for each node during the communication process. The dynamic operating characteristics include the node's real-time voltage, current, harmonic amplitude index, and the signal-to-noise ratio of the connected path, which are used to reflect the node's operating status and sensitivity to interference. S15. Further define a disturbance direction label on each edge to quantify the direction change of the interference signal on the propagation path, and record the disturbance direction angle as the disturbance direction label: Among them, θ ij Represents edge e ij The disturbance direction angle, Indicates that the interference electric field vector at time t+Δt is at the edge e ij The direction of the estimate, Indicates the interference electric field vector at time t at edge e ij The direction estimate on , ||·|| represents the norm, represents the spatial position vector of the j-th node, represents the spatial position vector of the i-th node, ∈ represents the smoothing factor, λ represents the weight coefficient, S ij Represents edge e ij The channel stability index on ,|·| represents the absolute value operation; S16. Obtain a communication graph that includes node physical states, side channel attributes, and disturbance direction labels.

4. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 2 is characterized in that The S2 specifically includes: S21. Construct a multi-layer rotation-aware graph neural network, where the multi-layer rotation-aware graph neural network is composed of L layers of graph convolution units. Each graph convolution unit includes an adjacency-aware module and a direction-aware module. The adjacency-aware module is used to extract structural connectivity in the graph, and the direction-aware module is used to capture edge direction changes and disturbance propagation angle features. S22. Initialize the node attribute set of the communication graph to the input feature matrix X (0) , where each row represents the original dynamic features of a node, which is input into the first layer of graph convolution unit; S23. In the lth layer graph convolution unit, let node v i The neighbor set of i ), the node features are updated using a rotation-aware convolution operator combined with the perturbation direction: in, represents the feature vector of the i-th node in the l-th layer, σ represents the activation function, W (l) represents the adjacency perception module weight matrix, represents the feature vector of the jth node in the l-1th layer, γ (l) represents the direction adjustment coefficient, θ ij Represents edge e ij The disturbance direction angle, D ij represents the direction embedding vector of the edge, α ij represents the normalization coefficient of disturbance intensity; S24. In each layer, the direction perception module is used to capture the edge direction change and disturbance propagation angle characteristics, and update them, and finally output the node feature matrix X of the Lth layer. (L) ; S25, concatenate and weight the node features output by multiple network layers to form a final fusion feature matrix to capture interference propagation patterns at different scales; S26. Based on the final fusion feature matrix, the interference heat value of each node is obtained by using full connection mapping and normalization operations, and an interference heat map is generated to characterize the possibility of interference to each node in the communication graph.

5. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 2 is characterized in that: The S3 specifically includes: S31, based on the interference heat map, extract the heat value higher than the threshold τ h The node set is used as a candidate interference source area, and multiple fruit fly individuals are initialized in the candidate interference source area, each fruit fly individual corresponds to a search starting point of a potential interference path; S32. Using the improved fruit fly optimization algorithm, the improvement is to introduce a multi-factor perception function to replace the search mechanism based on the single value judgment of the objective function: Among them, F k represents the kth fruit fly individual, J(F k ) represents the comprehensive perception function value calculated by the kth fruit fly individual in the current jump decision, η represents the jump amplitude adjustment factor, h k Indicates the interference heat of the node where the kth fruit fly individual is currently located, ρ kj Indicates the consistency coefficient between the current path direction and the historical disturbance direction, I kj Indicates the interference strength, SNR kj represents the signal-to-noise ratio, θ kj Represents edge e kj The disturbance direction angle; S33, the fruit fly individual performs local neighborhood selection based on the multi-factor perception function, performs a jump search operation, and jumps to a region where J(F k ) maximizes the adjacent nodes, and updates the trajectory sequence and direction history information of the fruit fly individual; S34. Set up an adaptive disturbance adjustment strategy. When the heat value of a fruit fly individual fluctuates or the disturbance direction reverses sharply during continuous jumping, the direction disturbance update mechanism is triggered: in, represents the individual’s current direction perception variable, α represents the disturbance amplitude adjustment factor, and rand(·) represents the random number generation function; S35. The jumping trajectory of each fruit fly individual is recorded in sequence as a path set, and the path sets of all individuals constitute the interference path candidate set.

6. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 2 is characterized in that: The S4 specifically includes: S41, for the interference path candidate set P = {P1, P2, ..., P N Each interference path in} is evaluated, where P i represents the i-th interference path, N represents the total number of interference paths, and the interference heat coverage and disturbance propagation rationality score are calculated for each interference path; S42. Interference heat coverage is defined as the normalized mean of the interference heat values ​​of each node on the path: Among them, Ω i represents the interference heat coverage of the i-th interference path, |P i | represents the length of the i-th interference path, h j represents the node v on the interference path j Interference heat value, max(h j ) represents the maximum interference heat value of all nodes in the i-th interference path; S43. Comprehensively evaluate the path direction continuity, consistency, and disturbance intensity change trend, and calculate the disturbance propagation rationality score: Among them, i represents the rationality score of disturbance propagation of the i-th interference path, θ mn Represents edge e mn The disturbance direction angle, I mn Represents edge e mn The interference intensity, δf mn Indicates frequency drift; S44. Normalize the interference heat coverage and the disturbance propagation rationality score and combine them into a path comprehensive score function: F i =β1·Ω i +β2·Ψ i ; Among them, Φ i represents the comprehensive score of the i-th interference path, β1 and β2 represent weight coefficients; S45, all interference paths P are evaluated based on the comprehensive score. i Sort them, select the interference path with the highest comprehensive score as the optimal interference path, and record the starting node of the optimal interference path as the current interference source position; S46: Output the optimal interference path and interference source position.

7. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 2 is characterized in that: The S5 specifically includes: S51, starting the hardware linkage control interface according to the optimal interference path and the interference source location, sending a control instruction to the hardware module of the communication circuit through the hardware control bus, and activating the filter, compensation loop and buffer amplifier along the interference path; S52. Based on the spectrum characteristics of the interference signal on the optimal interference path, select a filtering parameter corresponding to the interference frequency band to suppress the interference signal in the interference frequency band from entering the communication receiving unit; S53. Real-time monitoring of node voltage, current, and harmonic component data on the optimal interference path, and automatically adjusting the capacitance and inductance units of the compensation loop according to the node impedance characteristics to meet the impedance balance requirement during signal transmission and suppress signal reflection and channel distortion; S54, starting a feedback compensation loop, measuring the phase and amplitude characteristics of the signal in the communication link, dynamically adjusting the transistors, switching diodes, and phase shift circuit in the compensation loop, and generating a compensation signal with a similar amplitude and opposite phase to the interference signal in real time; S55, collecting signal quality data on the optimal interference path in real time, and automatically triggering a buffer amplifier adjustment program when it is found that the communication signal level drops, the waveform is distorted, or the bit error rate increases, and adjusting the amplifier gain parameters to restore the interfered low-level communication signal, perform waveform repair, and output a standardized signal; S56: If the communication link quality does not reach the preset indicator after the interference suppression control in steps S51-S55, the communication path switching mechanism is automatically executed according to the predefined alternative communication path data to switch to the suboptimal path for signal transmission.

8. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 2 is characterized in that: The S6 specifically includes: S61, collecting communication link signal quality data after hardware linkage control in real time, and establishing a real-time communication quality database; S62: Using the latest data in the real-time communication quality database, calculate the actual interference response deviation value of the communication node, and compare it with the heat value of the corresponding node in the interference heat map generated in step S2 to obtain a node heat error set; S63. Dynamically adjust the adjacency perception module weight matrix and direction adjustment coefficient of the multi-layer rotation perception graph neural network through a back-propagation mechanism according to the node heat error set; S64. Analyze the characteristics of path disturbance direction changes in real-time communication quality data, calculate the deviation between the actual direction angle change of path propagation and the disturbance direction angle during the fruit fly jump search process, adjust the direction perception variable and jump amplitude adjustment factor in the fruit fly individual jump strategy based on the direction deviation, and dynamically update the multi-factor perception function J(F) of the improved fruit fly optimization algorithm. k ): Among them, F k represents the kth fruit fly individual, J(F k ) represents the comprehensive perception function value calculated by the kth fruit fly individual in the current jump decision, η represents the jump amplitude adjustment factor, h k Indicates the interference heat of the node where the kth fruit fly individual is currently located, ρ kj Indicates the consistency coefficient between the current path direction and the historical disturbance direction, I kj Indicates the interference strength, SNR kj represents the signal-to-noise ratio, θ kj Represents edge e kj The disturbance direction angle; S65. Based on the updated multi-layer rotation perception map neural network parameters and the improved fruit fly optimization algorithm strategy, re-execute steps S2 and S3 to generate a new interference heat map and interference path candidate set, thereby realizing self-learning and strategy iteration capabilities.

9. The dual-mode carrier communication aggregation interference identification system based on graph neural network according to claim 8, characterized in that: The communication link signal quality data includes node signal-to-noise ratio, bit error rate, interference intensity, voltage and current changes.

Images