Self-evolution intelligent communication anti-interference method and system
By introducing biological evolution mechanisms into the wireless communication system, interference cognition, decision-making and waveform optimization are achieved, the problem of improving anti-interference ability under intelligent interference is solved, and the system's adaptability and communication reliability are improved.
Patent Information
- Application Number
- CN202410992654.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-23
- Publication Date
- 2024-11-12
AI Technical Summary
When facing intelligent interference, the existing intelligent communication anti-interference methods have fast forgotten tasks, slow convergence speed, low optimization efficiency, poor adaptability to interference environments, making it difficult to achieve continuous improvement of anti-interference capability.
Drawing on biological evolution mechanisms, through interference cognition, communication anti-interference decision-making, waveform evolution and effect evaluation, an autonomously evolved wireless communication system is formed to dynamically adapt to the electromagnetic environment, and optimize communication waveforms to improve anti-interference ability.
The independent evolution of wireless communication systems in complex electromagnetic environments has been realized, the dynamic adaptability and communication reliability of anti-interference capabilities have been improved, and effective communication under intelligent interference has been ensured.
Smart Images

Figure CN118944769A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of wireless communication anti-interference, and in particular to a self-evolving intelligent communication anti-interference method and system. Background Art
[0002] In recent years, with the rapid development of wireless systems such as 5G mobile communications, the Internet of Things (IoT), wireless ad hoc networks (MANETs), and unmanned trunked systems, a vast number of wireless communication devices have become indispensable in people's lives. However, the increasing presence of unauthorized and malicious users has further complicated the electromagnetic environment, severely impacting the reliability and effectiveness of wireless communication transmissions. With the advancement of technologies such as cognitive radio and artificial intelligence, the intelligence level of various unauthorized and malicious users has continued to increase, gradually acquiring intelligent features such as environmental awareness, online learning, waveform variability, collaborative operation, and cross-layer attacks. It is foreseeable that wireless communication systems will face increasingly intelligent interference threats. Therefore, wireless communication systems urgently need to improve the intelligence level of interference mitigation to effectively combat rapidly evolving intelligent interference. Currently, existing intelligent communication interference mitigation methods mostly use machine learning algorithms such as reinforcement learning and deep learning to learn the dynamic behavior of interference and develop optimized interference mitigation strategies. This has significantly improved their ability to cope with intelligent interference. However, these methods still suffer from limitations such as rapid task forgetting, slow convergence, low optimization efficiency, and poor adaptability to interference environments. Therefore, it is urgent to solve the problem of how to enable wireless communication systems to transform from stress-reflex anti-interference to learning interference behavior, so that the anti-interference capability can be continuously improved or evolved. Summary of the Invention
[0003] The present invention provides a self-evolving intelligent communication anti-interference method, which draws on the biological evolution mechanism to enable the wireless communication system to achieve "autonomous evolution" in an interference environment, with the aim of continuously improving its anti-interference capability according to the changing interference environment.
[0004] The technical solution to implement the present invention is: a self-evolving intelligent communication anti-interference method, the steps of which are:
[0005] Step 1 (S1), interference recognition, obtains interference knowledge such as interference type and interference characteristic parameters from the electromagnetic environment; recognition methods include but are not limited to signal processing-based methods and neural network-based methods; representation methods of interference knowledge include vectors, tensors and knowledge graphs;
[0006] Step 2 (S2), communication anti-interference decision-making, its main task is to provide an anti-interference strategy based on the results of interference cognition; the methods for providing an anti-interference strategy include but are not limited to decision-making methods based on anti-interference knowledge graphs and decision-making methods based on reinforcement learning;
[0007] Step 3 (S3), communication waveform evolution, according to the anti-interference strategy output by the communication anti-interference decision module, to achieve active adjustment of the communication waveform;
[0008] Step 4 (S4), communication effect evaluation, according to the anti-interference strategy output by the communication anti-interference decision module, to achieve active adjustment of the communication waveform;
[0009] Step 5 (S5), anti-interference knowledge is formed, and anti-interference knowledge is generated based on the interference knowledge sent by the interference recognition module, the anti-interference strategy sent by the anti-interference decision module, and the communication evaluation effect sent by the communication effect evaluation module; the anti-interference knowledge includes interference knowledge, corresponding anti-interference strategy and the communication evaluation effect of the strategy; after the anti-interference knowledge is generated, it is sent to the anti-interference knowledge base for storage.
[0010] Preferably, the method for obtaining interference knowledge includes: when the wireless communication system encounters interference, the interference recognition module extracts the characteristics of the interference signal and estimates the interference parameters; then the obtained interference knowledge is sent to the communication anti-interference decision module and the anti-interference knowledge formation module; the interference knowledge includes interference characteristics and / or interference parameters.
[0011] Preferably, after receiving the interference knowledge, the communication anti-interference decision module searches the anti-interference knowledge base according to the interference knowledge to obtain the anti-interference knowledge corresponding to the current interference knowledge; then the communication anti-interference decision module obtains the anti-interference strategy according to the interference knowledge and its corresponding anti-interference knowledge; the anti-interference strategy is sent to the communication waveform evolution module on the one hand, and to the anti-interference knowledge formation module on the other hand.
[0012] Preferably, the communication waveform evolution module adjusts the parameters or type of the communication waveform according to the anti-interference strategy, and establishes communication with the communication waveform.
[0013] Preferably, the communication effect evaluation module evaluates the reliability and effectiveness of communication transmission in an interference environment after adopting an anti-interference strategy and performing communication waveform evolution, providing a basis for the effectiveness of the anti-interference strategy.
[0014] Preferably, when the interference pattern is known, the anti-interference decision module directly searches for the anti-interference strategy corresponding to the known pattern and pattern parameters in the spectrum; when the interference pattern is unknown, the anti-interference decision module needs to first extract the time-frequency domain features of the interference from the characteristic parameters of the unknown interference through feature extraction, and search for the anti-interference strategy corresponding to the time-frequency domain features in the spectrum.
[0015] Preferably, when the interference pattern is unknown, the interference parameters are first processed to extract the time-frequency domain features of the interference pattern; then the time-frequency domain features are used as head nodes to search in the anti-interference knowledge graph to obtain the corresponding anti-interference strategy.
[0016] Preferably, the strategy fusion method includes: expressing each set of "strategy-parameters" as (a i ,p i ); where a i represents the anti-interference strategy, p i ={p i1 ,p i2 ,…,p in} represents the decision parameter corresponding to the anti-interference strategy; multiple anti-interference strategies are integrated. First, the intersection of the anti-interference strategies in the same domain is taken, and the union of the anti-interference strategies in different domains is taken to obtain the preliminary fused anti-interference strategy (A, P), where A = {a1, a2, ... a i} represents the policy set, P={p1,p2,…p n} represents the set of parameter sets corresponding to each strategy; the fused strategy is further optimized according to the communication requirements, redundant strategies are discarded while meeting the requirements for successful communication transmission, and the strategy with the smallest communication overhead is selected to finally form an anti-interference decision.
[0017] Preferably, the communication waveform evolution method includes: assuming that the communication parameter set of the t0 time slot of the frequency hopping communication system is
[0018] θ0={F0,M0,C0,P0,T0}
[0019] The frequency set used is F0, the modulation mode is M0, the coding mode is C0, the transmission power is P0, and the hopping period is T0. At the communication receiver, the anti-interference decision module runs and makes a decision at the t0 time slot, and the parameter set is changed to θ1 at the t1 time slot. The changed communication parameters are fed back from the communication receiver to the communication transmitter through the feedback link. When the transmitter and receiver enter the t1 time slot, the new parameter set θ1 is enabled at the same time, thereby achieving synchronous switching between transmission and reception and completing waveform evolution.
[0020] The present application also relates to a self-evolving intelligent communication anti-interference system, comprising an interference recognition module, a communication anti-interference decision module, a communication waveform evolution module, a communication effect evaluation module, an anti-interference knowledge formation module, and an anti-interference knowledge base;
[0021] Interference recognition module, used to realize the recognition of interference and send the interference knowledge to the communication anti-interference decision module;
[0022] The communication anti-interference decision module uses an anti-interference decision algorithm based on a knowledge graph to output an anti-interference strategy; the anti-interference decision algorithm includes strategy extraction and strategy fusion;
[0023] The communication waveform evolution module actively adjusts the communication waveform according to the anti-interference strategy output by the communication anti-interference decision module;
[0024] The communication effect evaluation module evaluates the performance of communication transmission in an interference environment after adopting anti-interference strategies and performing communication waveform evolution;
[0025] The anti-interference knowledge formation module generates anti-interference knowledge based on the interference knowledge sent by the interference recognition module, the anti-interference strategy sent by the anti-interference decision module, and the communication evaluation effect sent by the communication effect evaluation module; the anti-interference knowledge includes the interference knowledge, the corresponding anti-interference strategy, and the communication evaluation effect of the strategy;
[0026] The anti-interference knowledge base is used to store anti-interference knowledge.
[0027] Furthermore, the self-evolving intelligent communication anti-interference system achieves the communication anti-interference purpose according to any one of the aforementioned self-evolving intelligent communication anti-interference methods.
[0028] Compared with the existing anti-interference methods or architectures of wireless communication systems, the present invention has the following significant advantages: this application draws on the biological evolution mechanism, can realize the autonomous evolution of communication waveforms according to complex electromagnetic environments, dynamically adapt to the electromagnetic environment, and ensure the reliability and effectiveness of communication. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 It is a schematic diagram of the main functional components of the present invention;
[0030] Figure 2 This is a flowchart of the self-evolving intelligent communication anti-interference of the present invention;
[0031] Figure 3 This is a block diagram of the self-evolving intelligent communication anti-interference device of the present invention. DETAILED DESCRIPTION
[0032] The implementation steps of the present invention will be further described below with reference to the accompanying drawings.
[0033] like Figure 1-2 As shown, this application relates to a self-evolving intelligent communication anti-interference method:
[0034] Step 1 (S1), interference recognition, obtains interference knowledge such as interference type and interference characteristic parameters from the electromagnetic environment; recognition methods include but are not limited to signal processing-based methods and neural network-based methods; representation methods of interference knowledge include vectors, tensors and knowledge graphs;
[0035] Step 2 (S2), communication anti-interference decision-making, its main task is to provide an anti-interference strategy based on the results of interference cognition; the methods for providing an anti-interference strategy include but are not limited to decision-making methods based on anti-interference knowledge graphs and decision-making methods based on reinforcement learning;
[0036] Step 3 (S3), communication waveform evolution, according to the anti-interference strategy output by the communication anti-interference decision module, to achieve active adjustment of the communication waveform;
[0037] Step 4 (S4), communication effect evaluation, according to the anti-interference strategy output by the communication anti-interference decision module, to achieve active adjustment of the communication waveform;
[0038] Step 5 (S5), anti-interference knowledge is formed, and anti-interference knowledge is generated based on the interference knowledge sent by the interference recognition module, the anti-interference strategy sent by the anti-interference decision module, and the communication evaluation effect sent by the communication effect evaluation module; the anti-interference knowledge includes interference knowledge, corresponding anti-interference strategy and the communication evaluation effect of the strategy; after the anti-interference knowledge is generated, it is sent to the anti-interference knowledge base for storage.
[0039] Step 1 (S1), interference recognition, primarily involves acquiring interference knowledge from the electromagnetic environment. When a wireless communication system (hereinafter referred to as the "system") encounters interference, the interference recognition module extracts the characteristics of the interference signal and estimates the interference parameters. Cognition methods include, but are not limited to, signal processing-based and neural network-based methods. Interference knowledge is represented using vectors, tensors, and knowledge graphs. The obtained interference knowledge, including interference characteristics and parameters, is then fed into the communication anti-interference decision-making process (S2) and anti-interference knowledge generation process (S5).
[0040] The following uses a neural network-based method as an example to illustrate the interference recognition process. The anti-interference technology architecture proposed in the present invention is also applicable to other interference recognition methods.
[0041] The neural network-based interference recognition model was designed and trained using the PyTorch deep learning framework, the Python language, and related toolkits, combined with the MATLAB simulation platform. The trained model can be deployed on embedded platforms.
[0042] The process is as follows:
[0043] (1) MATLAB generates the combined interference signal samples and saves them as .mat files to form a data set;
[0044] (2) Preprocess the dataset;
[0045] (3) Import the preprocessed data into the convolutional neural network to start training the model and calculate the loss and accuracy;
[0047] (4) When the accuracy reaches the threshold, stop training and save the model;
[0048] (5) Use the trained model to predict the combined interference signal.
[0049] The convolutional neural network of this model consists of five convolutional layers, five pooling layers, and two fully connected layers. Data enters the neural network and first passes through the first convolutional layer, which increases the number of feature channels (1->8) and extracts features without reducing the data size of each feature channel (kernel_size=3, stride=1, padding=1). An activation function introduces nonlinearity into the neural network, allowing it to arbitrarily approximate any nonlinear function. A pooling layer (max pooling) further extracts features and reduces the data size to one-quarter (window size 2, stride defaults to 2), reducing computational complexity. After five hidden layers (one convolution layer followed by one pooling layer), the data is output through two linear layers and one dropout layer.
[0050] The Dropout layer is applied during the training of deep neural networks. During each training iteration, a certain number of neurons are randomly dropped from the neural network, forcing the network to rely on multiple subsets of neurons, thereby reducing overfitting. This project adds a Dropout layer between two linear layers to avoid overfitting.
[0051] The model training process is as follows:
[0052] (1) Initialization
[0053] First, convert the training device to a GPU to reduce CPU pressure, then load the convolutional neural network into the device, set relevant parameters (training set and validation set batch size, number of training rounds, learning rate, etc.), use the cross-entropy loss function as the loss function, initialize the optimizer, set the transformation method, and load the dataset.
[0054] (2) Training process
[0055] Start training the model, set the optimal accuracy, set the network model mode to training mode, and traverse the dataset with the set batch size for training in each training round: input data (make sure the data and the neural network model are on the same device), calculate the loss, clear the gradient information first, then calculate the parameter gradient, clip the gradient value to prevent gradient explosion, use the gradient to update the parameters, calculate the accuracy of the current batch, record the training accuracy and training loss of each batch, and take the average to get the training accuracy and training loss of each training round.
[0056] (3) Verification process
[0057] Verification process: Set the model mode to eval mode. Except for the non-gradient related calculations, the rest of the process is consistent with the training process. Finally, the verification accuracy and verification loss are obtained.
[0058] (4) Save the model
[0059] When the verification accuracy is greater than or equal to the optimal accuracy, the optimal accuracy at this time is updated to the verification accuracy, the log is updated and the training model is saved. If the tolerance reaches the upper limit but the verification accuracy does not reach the optimal accuracy, the training is completed in advance.
[0060] After the model training is completed, the interference type can be identified to form interference knowledge.
[0061] Step 2 (S2) involves the communication anti-interference decision-making process, which primarily develops an anti-interference strategy. After receiving the interference knowledge, the communication anti-interference decision-making process searches the anti-interference knowledge base based on the interference knowledge to obtain the anti-interference knowledge corresponding to the current interference knowledge. The communication anti-interference decision-making module then uses machine learning or other decision-making methods based on the interference knowledge and its corresponding anti-interference knowledge to develop an anti-interference strategy. This anti-interference strategy is fed into the communication waveform evolution process (S3) and the anti-interference knowledge formation process (S5).
[0062] The following uses the decision-making method based on the anti-interference knowledge graph as an example to illustrate the process of anti-interference decision-making. The anti-interference technical architecture proposed in this invention is also applicable to other anti-interference decision-making methods.
[0063] The knowledge graph-based anti-interference decision module consists of two main parts: ① If the interference pattern is known, the graph is directly searched for the anti-interference strategy corresponding to the known pattern and pattern parameters. ② If the interference pattern is unknown, feature extraction is first performed to extract the time-frequency domain features of the unknown interference from its characteristic parameters, and then the graph is searched for the anti-interference strategy corresponding to these time-frequency domain features.
[0064] To make decisions using the anti-interference knowledge graph, the first step is to search the anti-interference knowledge graph for corresponding anti-interference strategies based on the interference pattern. Known interference patterns have prior information in the anti-interference knowledge graph, so it is sufficient to understand how to use the anti-interference knowledge graph to search for anti-interference strategies for known patterns. Unknown interference patterns do not have directly corresponding anti-interference strategies in the knowledge graph. Furthermore, different patterns have different determinants of corresponding anti-interference strategies. Therefore, it is not possible to directly use the interference pattern as the head node for strategy search. Instead, it is necessary to extract searchable head nodes based on the parameter information of the unknown pattern. This section addresses unknown patterns by first processing the interference parameters to extract the time-frequency domain features of the interference pattern. These time-frequency domain features are then used as head nodes to search the anti-interference knowledge graph for corresponding anti-interference strategies. The relationship from parameters to features can also be viewed as a triple relationship, with nodes representing parameters and features, and edges representing computation rules. Because the Neo4j graph database cannot store computation rules, this step from parameters to features requires additional algorithms.
[0065] When using the constructed anti-interference knowledge graph to make search decisions, it is necessary to consider how to generate search results for different search needs; secondly, since the knowledge graph stores knowledge relationships rather than cases, for different parameter conditions of the same interference style, the graph search results are the same but the actual decision results may be different. This is because the graph stores all possible decision results of the style. Therefore, after using the anti-interference knowledge graph to search for anti-interference decisions, it is also necessary to fuse the strategies and select the best strategy. The main solution of this section is to design a search decision algorithm that can select the best decision results under different parameter conditions and output the anti-interference strategy and corresponding parameters. The anti-interference decision algorithm based on knowledge graph studied in this section is carried out in two steps: strategy extraction and strategy fusion. "Strategy extraction" is the process of obtaining anti-interference strategies through anti-interference knowledge graph search and matching decision parameters. Assume that each set of "strategy-parameter" sets can be expressed as (a i ,p i ). Among them, a i represents an anti-interference strategy, p i ={p i1 ,p i2 ,…,p in} represents the decision parameter corresponding to the anti-interference strategy. During the search process, multiple anti-interference strategies may be obtained. "Strategy fusion" is to integrate the multiple anti-interference strategies that may exist after "extraction" to obtain the optimal strategy. First, the intersection of the anti-interference strategies in the same domain is taken, and the union of the anti-interference strategies in different domains is taken to obtain the preliminary fused anti-interference strategy (A, P), where A = {a1, a2, ... a i} represents the policy set, P={p1,p2,…p n} represents the set of parameters corresponding to each strategy. The fused strategy is then further optimized based on communication requirements. While ensuring successful communication transmission, redundant strategies are discarded, and the strategy with the lowest communication overhead is selected, ultimately forming an anti-interference decision. For example, if the initially "fused" anti-interference strategy includes both frequency and power domains, and if only the frequency domain strategy is sufficient for anti-interference, and the execution reward function shows that the frequency domain strategy has lower overhead than the power domain strategy, the power domain strategy can be discarded, and the frequency domain strategy can be used as the final decision.
[0066] Step 3 (S3), communication waveform evolution, has the main task of adjusting communication waveform parameters, type, etc. according to the anti-interference strategy, and establishing communication with the communication waveform.
[0067] The following uses a typical frequency hopping communication system as an example to illustrate the evolution of the communication waveform. The anti-interference technology architecture proposed in the present invention is also applicable to other communication systems.
[0068] Assume that the communication parameter set of the frequency hopping communication system t0 time slot is
[0069] θ0={F0,M0,C0,P0,T0}
[0070] The frequency set used is F0, the modulation scheme is M0, the coding scheme is C0, the transmit power is P0, and the hopping period is T0. At the receiver, the anti-interference decision module runs and makes a decision at time slot t0. At time slot t1, the parameter set is changed to θ1. The changed communication parameters are fed back from the receiver to the transmitter via a feedback link. When the transmitter and receiver enter time slot t1, the new parameter set θ1 is simultaneously activated, achieving synchronized transmission and reception switching and completing waveform evolution.
[0071] Step 4 (S4), communication effect evaluation, its main task is to evaluate the reliability and effectiveness of communication transmission in an interference environment after adopting an anti-interference strategy and performing communication waveform evolution, and provide a basis for the effectiveness of the anti-interference strategy.
[0072] The following uses a typical frequency hopping communication system and a score-based evaluation method as an example to illustrate the communication effect evaluation process. The anti-interference technology architecture proposed in the present invention is also applicable to other communication systems and other communication effect evaluation methods.
[0073] The evaluation considers several key communication indicators, including the frame error rate P e , information rate R, transmission power P and hop period T.
[0074] When evaluating, the weighted sum of each communication indicator is used to give a score that represents the communication capability of the system, thus giving an intuitive evaluation. The calculation method of the system performance score is:
[0075] U=w1R+w2T-w3P-w4logP e
[0076] Among them, w i ≥0, i=1, 2, 3, 4 are weighting coefficients, adjusted based on the importance of each parameter to the system. For communication systems, a higher information rate is more advantageous, so system performance is positively correlated with the information rate. A too short hop period requires more guard slots, reducing communication efficiency. Therefore, system performance is positively correlated with the hop period. For interference mitigation, communication systems aim to minimize transmission power, as higher transmission power incurs greater costs. Therefore, system performance is negatively correlated with transmission power. A lower frame error rate (FR) indicates better communication system performance. FR is typically measured in terms of magnitude, with a minimum of 0. Therefore, system performance is negatively correlated with the logarithm of the FR.
[0077] Step 5 (S5) involves forming anti-interference knowledge. This involves generating a piece of anti-interference knowledge based on the interference knowledge provided by the interference recognition process, the anti-interference strategy provided by the anti-interference decision process, and the communication evaluation results of the strategy provided by the communication evaluation process, thereby evolving anti-interference capabilities. A piece of anti-interference knowledge consists of the interference knowledge, the corresponding anti-interference strategy, and the communication evaluation results of the strategy. This knowledge indicates the anti-interference strategy to be adopted in a given interference environment and the effect of the strategy. After the anti-interference knowledge piece is generated, it is stored in the anti-interference knowledge base.
[0078] In addition, if Figure 3 The present invention provides a self-evolving intelligent communication anti-interference system, which specifically includes:
[0079] B01: Interference recognition module, used to realize interference recognition; a dedicated interference recognition channel can be set in the communication receiver, or a conventional communication receiving channel can be used for implementation.
[0080] B02: Anti-interference decision module, used to provide anti-interference strategy;
[0081] B03: Communication waveform evolution module, used to achieve active adjustment of the communication waveform;
[0082] B04: Anti-interference knowledge formation module, used to generate anti-interference knowledge;
[0083] B05: Anti-interference knowledge base, used to store anti-interference knowledge.
[0084] B06: Communication effect evaluation module, used to evaluate the communication effect of the current communication waveform.
[0085] Furthermore, the self-evolving intelligent communication anti-interference system achieves the communication anti-interference purpose according to the aforementioned self-evolving intelligent communication anti-interference method. This part of the content is described here by reference and will not be repeated here.
[0086] Those skilled in the art will appreciate that some or all of the steps of the above embodiments may be accomplished by hardware, or may be accomplished by programs instructing related hardware. The programs may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk, or an optical disk, etc.
[0087] The technical features of the above embodiments can be combined in any way. To simplify the description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction between the combinations of these technical features, they should be considered within the scope of this specification.
[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
[0089] The contents not described in detail in the present application belong to the prior art known to professional and technical personnel in this field.
Claims
1. A self-evolving intelligent communication anti-interference method, characterized in that: The interference recognition module obtains interference knowledge from the electromagnetic environment to realize the recognition of interference, and sends the interference knowledge to the communication anti-interference decision module; After receiving the interference knowledge, the communication anti-interference decision module outputs the anti-interference strategy based on the anti-interference decision algorithm of the knowledge graph; the anti-interference decision algorithm includes strategy extraction and strategy fusion; The communication waveform evolution module realizes active adjustment of the communication waveform according to the anti-interference strategy output by the communication anti-interference decision module; The communication effect evaluation module evaluates the performance of communication transmission in an interference environment after adopting anti-interference strategies and performing communication waveform evolution. The anti-interference knowledge formation module generates anti-interference knowledge based on the interference knowledge sent by the interference recognition module, the anti-interference strategy sent by the anti-interference decision module and the communication evaluation effect sent by the communication effect evaluation module; the anti-interference knowledge includes interference knowledge, corresponding anti-interference strategy and the communication evaluation effect of the strategy; after the anti-interference knowledge is generated, it is sent to the anti-interference knowledge base for storage.
2. The self-evolving intelligent communication anti-interference method according to claim 1 is characterized in that: The method for acquiring interference knowledge includes: when the wireless communication system encounters interference, the interference recognition module extracts the characteristics of the interference signal and estimates the interference parameters; then the obtained interference knowledge is sent to the communication anti-interference decision module and the anti-interference knowledge formation module; the interference knowledge includes interference characteristics and / or interference parameters.
3. The self-evolving intelligent communication anti-interference method according to claim 1, characterized in that: After receiving the interference knowledge, the communication anti-interference decision module searches the anti-interference knowledge base according to the interference knowledge to obtain the anti-interference knowledge corresponding to the current interference knowledge; Then the communication anti-interference decision module obtains the anti-interference strategy based on the interference knowledge and its corresponding anti-interference knowledge; The anti-interference strategy is sent to the communication waveform evolution module on the one hand, and to the anti-interference knowledge formation module on the other hand.
4. The self-evolving intelligent communication anti-interference method according to claim 1, characterized in that: The communication waveform evolution module adjusts the parameters or type of the communication waveform according to the anti-interference strategy and establishes communication with the communication waveform.
5. The self-evolving intelligent communication anti-interference method according to claim 1, characterized in that: The communication effect evaluation module evaluates the reliability and effectiveness of communication transmission in an interference environment after adopting an anti-interference strategy and performing communication waveform evolution, providing a basis for the effectiveness of the anti-interference strategy.
6. The self-evolving intelligent communication anti-interference method according to claim 1, characterized in that: When the interference pattern is known, the anti-interference decision module directly searches for the anti-interference strategy corresponding to the known pattern and pattern parameters in the spectrum; when the interference pattern is unknown, the anti-interference decision module needs to first extract the time-frequency domain features of the interference from the feature parameters of the unknown interference through feature extraction, and search the anti-interference strategy corresponding to the time-frequency domain features in the spectrum.
7. The self-evolving intelligent communication anti-interference method according to claim 6, characterized in that: When the interference pattern is unknown, the interference parameters are first processed to extract the time-frequency domain features of the interference pattern. Then, the time-frequency domain features are used as the head nodes to search in the anti-interference knowledge graph to obtain the corresponding anti-interference strategy.
8. The self-evolving intelligent communication anti-interference method according to claim 1, characterized in that: Strategy fusion methods include: Each set of "strategy-parameters" is represented as (a i ,p i ), where a i represents the anti-interference strategy, p i ={p i1 , p 12 …,p in } represents the decision parameter corresponding to the anti-interference strategy; multiple anti-interference strategies are integrated, firstly the intersection of the anti-interference strategies in the same domain is taken, and the union of the anti-interference strategies in different domains is taken to obtain the initially integrated anti-interference strategy (A, P), where A = {a1, a2, ...a i } represents the strategy set, P = {p1, p2…p n } represents the set of parameter sets corresponding to each strategy; the fused strategy is further optimized according to the communication requirements, redundant strategies are discarded while meeting the requirements for successful communication transmission, and the strategy with the smallest communication overhead is selected to finally form an anti-interference decision.
9. The self-evolving intelligent communication anti-interference method according to claim 1, characterized in that: Communication waveform evolution methods include: Assume that the communication parameter set of the frequency hopping communication system t0 time slot is θ0={F0,M0,C0,P0,T0} The frequency set used is F0, the modulation mode is M0, the coding mode is C0, the transmission power is P0, and the hopping period is T0; at the communication receiver end, the anti-interference decision module is run, and a decision is given in the t0 time slot, and the parameter set is changed to θ1 in the t1 time slot; through the feedback link, the changed communication parameters are fed back from the communication receiver to the communication transmitter; when the transmitter and receiver enter the t1 time slot, the new parameter set θ1 is enabled at the same time, thereby realizing synchronous switching of transmission and reception and completing waveform evolution.
10. A self-evolving intelligent communication anti-interference system, characterized in that: It includes interference recognition module, communication anti-interference decision module, communication waveform evolution module, communication effect evaluation module, anti-interference knowledge formation module and anti-interference knowledge base; Interference recognition module, used to realize the recognition of interference and send the interference knowledge to the communication anti-interference decision module; The communication anti-interference decision module uses the anti-interference decision algorithm based on the knowledge graph to output the anti-interference strategy; the anti-interference decision algorithm includes strategy extraction and strategy fusion; The communication waveform evolution module realizes active adjustment of the communication waveform according to the anti-interference strategy output by the communication anti-interference decision module; The communication effect evaluation module evaluates the performance of communication transmission in an interference environment after adopting anti-interference strategies and performing communication waveform evolution; The anti-interference knowledge forming module generates anti-interference knowledge according to the interference knowledge sent by the interference recognition module, the anti-interference strategy sent by the anti-interference decision module and the communication evaluation effect sent by the communication effect evaluation module; the anti-interference knowledge includes the interference knowledge, the corresponding anti-interference strategy and the communication evaluation effect of the strategy; The anti-interference knowledge base is used to store anti-interference knowledge.
Citation Information
Cited By
Communication anti-interference method based on deep reinforcement learning and related device
CN120378042A
Method and device for processing low-altitude network interference and electronic equipment
CN120751423A
Method, device and electronic equipment for handling low altitude network interference
CN120751423B
Communication chip signal interference prediction and optimization system based on artificial intelligence
CN121173404A
An artificial intelligence-based communication chip signal interference prediction and optimization system
CN121173404B
