A radio system interference handling method and apparatus

Abstract

The application discloses a kind of radio system interference processing method and equipment, it is related to radio anti-interference technical field, including real-time environment monitoring and data acquisition;Interference source detection and accurate identification;Interference source dynamic prediction and trend analysis;Multi-dimensional interference avoidance strategy generation;Real-time feedback and intelligent decision optimization;Collaborative work and system integration;Final performance evaluation and adjustment.The priority of the application is, through real-time environment monitoring, accurate identification and dynamic prediction of interference source, realize the rapid positioning and accurate avoidance of radio interference source, compared with traditional fixed threshold or energy detection method, can respond to the dynamic change of frequency, bandwidth, power etc.of interference source in time, reduce misjudgment and omission, avoid blind avoidance or over-avoidance, improve the accuracy of interference avoidance.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of radio anti-jamming technology, in particular to a radio system interference processing method and device. BACKGROUND

[0002] In the field of radio communication, signal interference is a key problem that restricts communication quality and spectrum utilization efficiency. Traditional interference processing methods rely on single-dimensional passive response mechanisms, such as fixed frequency switching or static power adjustment, which are difficult to cope with complex dynamic electromagnetic environments. Existing technologies are often limited to single feature analysis in precise identification of interference sources, cannot effectively distinguish mixed interference types, and have weak interference prediction ability, lacking active prediction ability for sudden interference and trend changes. In addition, avoidance strategies usually operate in isolation, making it difficult to dynamically coordinate frequency, space, power, and time domain response, resulting in resource conflicts, switching delays, and other problems when the system faces complex interference, making it difficult to guarantee communication error rate and spectrum utilization.

[0003] The current technical bottleneck mainly manifests in three aspects: first, the identification ability is limited, most systems rely on preset interference thresholds or artificial experience, and cannot realize intelligent classification and positioning of interference sources through multi-dimensional signal features; second, the dynamic response is insufficient, lacking modeling ability for interference intensity time characteristics, making it difficult to predict interference evolution trend using historical data and environmental factors; third, the strategy coordination is missing, existing avoidance schemes mostly use static combinations, and cannot dynamically adjust strategy weights based on real-time performance feedback, resulting in unstable long-term anti-interference performance. Therefore, there is an urgent need for a systematic solution that integrates precise identification, multi-source prediction, multi-dimensional collaborative avoidance, and intelligent feedback optimization. SUMMARY

[0004] To solve the above technical problems, a radio system interference processing method and device are provided, which solves the problem that existing technologies are often limited to single feature analysis in precise identification of interference sources, cannot effectively distinguish mixed interference types, and have weak interference prediction ability, lacking active prediction ability for sudden interference and trend changes.

[0005] To achieve the above purposes, the technical solution adopted by the present application is:

[0006] A radio system interference processing method, comprising:

[0007] Real-time environment monitoring and data collection: using multiple sensors to perform real-time monitoring of the radio environment, collecting spectrum, interference signal strength, signal quality key information, performing spectrum analysis on all radio signals existing in the environment, collecting and analyzing the use of different frequency bands, and judging the signal quality and interference strength of each frequency band;

[0008] Interference source detection and accurate identification: process the collected signals through signal analysis algorithms to determine possible interference sources, use machine learning algorithms combined with historical interference data to classify signals and accurately identify the type of interference source, its interference frequency, intensity, and periodicity, and on this basis, conduct positioning analysis of the interference source;

[0009] Interference source dynamic prediction and trend analysis: based on machine learning or deep learning algorithms, combined with real-time monitoring data, historical interference data and environmental factors, predict the future behavior of the interference source through the following steps to analyze the trend of the dynamic change of the interference source and predict the change of interference intensity in advance;

[0010] Multi-dimensional interference avoidance strategy generation: based on the positioning analysis results of the interference source and the predicted trend of the interference intensity, dynamically generate combined strategies from four dimensions of frequency, space, power and time domain, including: selecting idle frequency bands according to the frequency distribution of the interference source, avoiding the interference direction through beamforming technology to adjust the antenna pattern, reducing the transmission power of the interfered frequency band, and allocating standby time slots for communication during the high interference period;

[0011] Real-time feedback and intelligent decision optimization: during system operation, real-time performance data of the system is collected through sensors and feedback modules, including communication error rate, signal-to-noise ratio and spectrum utilization rate:

[0012] Collaboration and system integration: modules interact with each other through standardized interfaces to form a highly integrated system, and different avoidance strategies are intelligently switched or combined according to environmental and interference changes;

[0013] Final performance evaluation and adjustment: after the system has been running for a period of time, the interference avoidance effect is comprehensively evaluated based on the expected return, and according to the evaluation results, the interference identification algorithm, avoidance strategy and feedback mechanism of the system are further optimized.

[0014] Preferably, in the real-time environment monitoring and data collection, further specifically including classifying different types of radio signals and accurately analyzing the spectrum according to the signal type.

[0015] Preferably, the interference source detection and accurate identification specifically includes:

[0016] Extracting multi-dimensional feature vectors from signals: extracting signal amplitude fluctuation features from the time domain and arrival angle features from the spatial domain to form multi-dimensional feature vectors;

[0017] Training statistical characteristic parameters of various interference sources through historical interference sample data, including mean vector and covariance matrix;

[0018] Calculate the posterior probability that the signal belongs to a specific interference source class under the condition of the extracted current signal feature vector;

[0019] Identify the interference source and locate its position based on the calculated posterior probability.

[0020] Preferably, the interference source dynamic prediction and trend analysis specifically includes:

[0021] Collect interference intensity data continuously in minutes and organize it into time series data;

[0022] Build a long short-term memory prediction model, which contains two hidden layers designed to capture long-term temporal dependencies of interference signals;

[0023] Input the historical interference intensity time series data containing 10 consecutive time points into the trained LSTM model for inference;

[0024] The LSTM model outputs the predicted value of the interference intensity after k time steps;

[0025] The model output weights and environmental factor bias terms are optimized and adjusted through the backpropagation algorithm, and the environmental factors include weather, geographical location and device density, which are integrated into the prediction model through feature encoding.

[0026] Preferably, in the multi-dimensional interference avoidance strategy generation, the strategy generation logic is: when a sudden interference is predicted, the time slot allocation is enabled first; for narrowband interference, frequency avoidance is adopted; for directional interference, beamforming is started; power control is triggered only when the interference intensity exceeds the threshold.

[0027] Preferably, in the step of system integration, each module interacts with data through standardized interfaces.

[0028] Preferably, the real-time feedback and intelligent decision optimization specifically includes:

[0029] When the performance index deviates from the preset threshold, trigger the strategy adjustment mechanism;

[0030] Define the immediate reward, which is the weighted sum of the spectrum utilization rate improvement ratio and the bit error rate reduction ratio, reflecting the immediate effect of the current interference avoidance strategy;

[0031] Define the expected return of the strategy, which is the sum of the immediate rewards of all future time points after attenuation by a discount factor, reflecting the overall long-term performance of the strategy;

[0032] Build a reinforcement learning model, which uses the policy gradient method to dynamically update the strategy parameters according to the real-time feedback performance data and reward signals, and continuously optimizes and adjusts the avoidance strategy.

[0033] Further, a radio system interference processing device is proposed for implementing the radio system interference processing method as described above, comprising:

[0034] A signal detection module: responsible for real-time monitoring and collection of signals in the radio spectrum, distinguishing between interference signals and target signals;

[0035] An interference identification and analysis module: analyzes the collected signal data, identifies the interference source and classifies it through feature extraction, pattern recognition and other techniques;

[0036] A dynamic spectrum management module: adjusts the spectrum usage strategy according to the real-time situation of the interference source and the target signal, avoiding conflicts of spectrum resources;

[0037] An interference source prediction module: combines historical data and real-time signal changes to predict the possible trend of the interference source;

[0038] An anti-interference strategy generation module: based on interference identification and spectrum management information, automatically generates reasonable interference avoidance strategies;

[0039] An adaptive adjustment module: dynamically adjusts the avoidance strategy according to system performance feedback and real-time environmental changes;

[0040] An intelligent decision support module: combines machine learning and data analysis algorithms to provide the best interference avoidance decisions according to different environments and task requirements;

[0041] A feedback monitoring module: real-time monitoring of the interference avoidance effect of the system, and feeding back the data to the decision support module for adjustment;

[0042] A user interface and reporting module: provides system operation interface and monitoring panel, showing interference avoidance status, spectrum usage and other information;

[0043] A communication coordination module: ensures smooth information transmission and collaborative work between devices and modules within and outside the system.

[0044] Optionally, the signal detection module includes a plurality of sensor arrays.

[0045] Optionally, the communication coordination module supports real-time communication and data synchronization between different modules within the system.

[0046] Compared with the prior art, the present application has the following advantages:

[0047] The application significantly improves the interference source identification accuracy through multi-dimensional feature fusion technology, realizes the active prediction of interference strength change trend based on time sequence modeling and long short-term memory network, and innovatively fuses four-dimensional dynamic avoidance strategies of frequency switching, beamforming, power adjustment and time slot allocation, and dynamically optimizes the strategy weight according to the real-time communication quality feedback combining with the reinforcement learning mechanism, solves the problems of limited identification ability, response lag and strategy isolation of traditional methods from the root, realizes the accurate suppression of interference signals, the cooperative improvement of real-time and adaptability of interference avoidance, and maximizes the spectrum utilization efficiency and reduces the bit error rate. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The radio system interference processing method flowchart is provided for the present scheme;

[0049] Figure 2 The interference source detection and accurate identification method flowchart is provided for the present scheme;

[0050] Figure 3 The interference source dynamic prediction and trend analysis method flowchart is provided for the present scheme;

[0051] Figure 4 The real-time feedback and intelligent decision optimization method flowchart is provided for the present scheme. DETAILED DESCRIPTION

[0052] The following description is used to disclose the present application so that those skilled in the art can implement the present application. The preferred embodiments in the following description are only as examples, and other obvious modifications can be thought of by those skilled in the art.

[0053] Referring to Figure 1 The radio system interference processing method shown, comprising:

[0054] Real-time environment monitoring and data collection: a variety of sensors are used for real-time monitoring of radio environment, collecting spectrum, interference signal strength, signal quality key information, performing spectrum analysis on all radio signals existing in the environment, collecting and analyzing the use of different frequency bands, judging the signal quality and interference strength of each frequency band, realizing panoramic perception of radio environment through multi-sensor cooperative monitoring and full-band spectrum analysis, providing high-precision data basis for interference identification, and significantly improving the real-time and comprehensiveness of signal quality evaluation;

[0055] In some preferred embodiments, real-time environmental monitoring and data collection also includes classifying different types of radio signals and performing precise spectral analysis based on signal type. By implementing classification filtering on multi-source radio signals and performing differentiated narrowband spectral analysis based on signal type, background noise interference from non-target frequency bands is eliminated at the source. This design significantly improves the detection accuracy of useful signal frequency bands, amplifies the abnormal fluctuations and spectral distortion characteristics of weak interference signals, and provides high-purity data basis for subsequent interference source identification. It also drives the synergistic leap of spectral resource analysis efficiency and dynamic calibration ability of anti-interference threshold, effectively solving the problem of feature flooding in sensitive frequency bands caused by traditional wide frequency scanning.

[0056] Interference source detection and accurate identification: By processing the collected signals through signal analysis algorithms, possible interference sources are determined. Machine learning algorithms are used in combination with historical interference data to classify signals and accurately identify the type of interference source, its interference frequency, intensity, and periodicity. Based on this, interference source positioning analysis is performed, combining multi-dimensional feature extraction and machine learning classification techniques to break through the limitations of traditional single feature recognition and accurately locate the type and physical location of the interference source, providing reliable target information for dynamic avoidance strategies.

[0057] Interference source dynamic prediction and trend analysis: Based on machine learning or deep learning algorithms, real-time monitoring data, historical interference data, and environmental factors are combined to predict future behavior of interference sources and analyze trends in dynamic changes. By predicting changes in interference intensity in advance, long short-term memory network modeling is used to model the timing changes of interference intensity, combined with environmental factor feature encoding to achieve proactive prediction of interference evolution and effectively solve the problem of response lag to sudden interference.

[0058] Multi-dimensional interference avoidance strategy generation: Based on the results of interference source positioning analysis and the predicted trends of interference intensity, combined strategies are dynamically generated in four dimensions: frequency, space, power, and time domain. Specifically, it includes selecting idle frequency bands based on interference source frequency distribution, avoiding interference direction through beamforming technology, reducing transmission power in interfered frequency bands, and allocating standby time slots for communication during high interference periods.

[0059] In some preferred embodiments, the multi-dimensional interference avoidance strategy generation logic is: when a burst interference is predicted, time slot allocation is preferentially enabled; frequency avoidance is adopted for narrowband interference; beamforming is started for directional interference; power control is triggered only when the interference intensity exceeds a threshold, by establishing an intelligent mapping mechanism of interference type and avoidance strategy, time slot allocation is automatically preferentially enabled when a burst interference is predicted to achieve communication timing space isolation, frequency avoidance is accurately triggered for narrowband interference to jump to an idle frequency band, beamforming technology is started in real time for directional interference to build a spatial barrier to block, and power control is activated only when the interference intensity breaks through a preset threshold to avoid invalid energy loss. This design breaks through the resource conflict problem of traditional multi-strategy parallel execution, realizes the optimal matching of interference characteristics and four-dimensional strategies with millisecond-level response speed, significantly improves the dynamic adaptation efficiency of spectrum resources, the communication link invulnerability, and the system energy efficiency self-optimization ability, forming a closed-loop decision-making system with precise targeting and dynamic lightness.

[0060] Real-time feedback and intelligent decision optimization: During system operation, the performance data of the system is collected in real time through sensors and feedback modules, including communication error rate, signal-to-interference-and-noise ratio, and spectrum utilization, and the performance indicators such as error rate and spectrum utilization are converted into strategy optimization signals through reinforcement learning mechanism to realize closed-loop dynamic optimization of avoidance strategies and ensure continuous and stable improvement of communication quality;

[0061] Collaborative work and system integration: Each module interacts with each other through standardized interfaces to form a highly integrated system, and different avoidance strategies are intelligently switched or combined according to the changes of the environment and interference, relying on standardized interfaces to realize intelligent collaboration of multiple modules, supporting seamless switching and combined execution of strategies, and significantly enhancing the overall response efficiency and robustness of the system in response to time-varying interference;

[0062] Final performance evaluation and adjustment: After the system has been running for a period of time, the interference avoidance effect is comprehensively evaluated based on expected return, and according to the evaluation results, the interference identification algorithm, avoidance strategy, and feedback mechanism of the system are further optimized, the long-term anti-interference performance is quantitatively evaluated based on the expected return model, the iterative optimization of the identification algorithm, avoidance strategy, and feedback mechanism is driven, and the self-adaptation ability of the system is continuously evolved.

[0063] Reference Figure 2 As shown, the interference source detection and accurate identification specifically includes:

[0064] Extracting a multi-dimensional feature vector from the signal: extracting signal amplitude fluctuation features from the time domain, extracting arrival angle features from the spatial domain, and combining them into a multi-dimensional feature vector;

[0065] The statistical characteristic parameters of each type of interference source are obtained by training historical interference sample data, and the statistical characteristic parameters include a mean vector and a covariance matrix;

[0066] The posterior probability of the signal belonging to a specific interference source category is calculated under the condition of the extracted current signal feature vector;

[0067] The interference source is identified and its position is located based on the calculated posterior probability;

[0068] Specifically, the calculation of the posterior probability is as follows:

[0069] ;

[0070] In the formula, is the posterior probability of the signal feature vector under the condition of the interference source , and are the mean vector and the covariance matrix of the interference source , is the feature vector extracted from multi-dimensional time domain and spatial domain data.

[0071] By fusing the time domain amplitude fluctuation and the spatial domain angle of arrival characteristics to construct a multi-dimensional feature vector, combining the mean vector and the covariance matrix trained by historical samples, and using the posterior probability model to quantify the interference source attribution confidence, the limitations of traditional threshold discrimination are broken through. This design significantly improves the classification ability and positioning accuracy of the interference source category in complex electromagnetic environments, provides a high-reliability target identification basis for subsequent dynamic avoidance strategies, and enhances the anti-noise analysis capability of the system for mixed interference signals.

[0072] As shown in Figure 3 , the interference source dynamic prediction and trend analysis specifically includes:

[0073] The interference intensity data is continuously collected in minutes and organized into time series data;

[0074] A long short-term memory prediction model is constructed, which contains two hidden layers and is designed to capture the long-term temporal dependence of interference signals;

[0075] The historical interference intensity time series data containing 10 consecutive time points are input into the trained LSTM model for inference;

[0076] The LSTM model outputs the predicted value of the interference intensity after k time steps in the future;

[0077] The model output weight and the environmental factor bias term are optimized and adjusted through the back propagation algorithm, and the environmental factors include weather, geographical location and device density, which are integrated into the prediction model through feature encoding;

[0078] Specifically, the expression of the prediction model is:

[0079] ;

[0080] wherein, is the predicted value of the interference intensity after k time steps, is the interference intensity data at time t, and b are the output weight and bias term of the LSTM model, denotes a long short-term memory network for capturing long-term dependencies of the interference signal.

[0081] By constructing an interference intensity time series through minute-level high-density sampling, using a double-layer long short-term memory network to deeply mine the long-period dependence characteristics of the signal, and combining feature encoding and bias term optimization of environmental factors, the modeling capability of dynamic changes in interference intensity is significantly improved. This design breaks through the response delay limitation of traditional static prediction methods for sudden interference, realizes accurate prediction of future multi-step interference intensity, provides forward-looking decision basis for avoidance strategies, and continuously adapts to environmental changes through the back propagation mechanism, effectively suppressing external factor interference and predicting stability. Specifically, the LSTM structure enhances the ability to capture the long-term and short-term fluctuation patterns of the interference signal, overcoming the defects of simple regression models in dealing with periodic and sudden interference. Dynamic factors such as weather and device density are integrated into the model output bias term through feature encoding, significantly improving the prediction robustness in complex scenarios. Based on the back propagation algorithm, the output weight and environmental bias term are continuously optimized, enabling the prediction model to have online evolution capability.

[0082] Referring to Figure 4 , real-time feedback and intelligent decision optimization specifically includes:

[0083] When the performance index deviates from the preset threshold, trigger the strategy adjustment mechanism;

[0084] Define the immediate reward, which is the weighted sum of the spectrum utilization rate improvement ratio and the bit error rate reduction ratio, reflecting the immediate effect of the current interference avoidance strategy;

[0085] Define the expected return of the strategy, which is the sum of the immediate rewards of all future time points after attenuation by a discount factor, reflecting the overall long-term performance of the strategy;

[0086] Build a reinforcement learning model, which uses the policy gradient method to dynamically update the strategy parameters based on the real-time feedback performance data and reward signals, and continuously optimizes and adjusts the avoidance strategy.

[0087] Specifically, the calculation formula of the expected return is:

[0088] ;

[0089] In the formula, is the expected return of the strategy, representing the overall performance of the system under the current strategy, is the immediate reward at time t, reflecting the effect of the interference avoidance strategy, is the discount factor, representing the degree of influence of future rewards, is the parameter of the interference avoidance strategy, representing the decision-making strategy of the system to deal with interference.

[0090] The scheme encodes the core indicators such as communication error rate and spectrum utilization into immediate reward signals of reinforcement learning, establishes a direct association between environmental feedback and strategy optimization, and uses the expected return model to plan the current reward and future decay reward through the discount factor, to avoid the risk of short-sighted decision-making in strategy optimization. Combined with the policy gradient method, the model dynamically updates the parameters according to the real-time reward, so that the interference avoidance strategy has the ability of continuous self-optimization without the need for manual reconfiguration. By converting spectrum utilization and error rate into weighted immediate reward signals and combining the discount factor to construct an expected return model to quantify the long-term performance of the strategy, and based on the policy gradient method to realize the closed-loop dynamic optimization of the interference avoidance strategy. This design breaks through the adaptability limitations of traditional static strategies, enabling the system to autonomously evolve decision logic based on real-time communication quality, significantly improving the dynamic adaptability and long-term stability of interference avoidance. At the same time, the continuous optimization mechanism effectively balances immediate performance and system persistent anti-interference capability.

[0091] Further, based on the same inventive concept as the above radio system interference handling method, the scheme proposes a radio system interference handling device, comprising:

[0092] Signal detection module: responsible for real-time monitoring and collection of signals in the radio spectrum, and identifying the difference between interference signals and target signals;

[0093] Interference identification and analysis module: analyze the collected signal data, identify the interference source and classify it through feature extraction, pattern recognition and other technologies;

[0094] Dynamic spectrum management module: adjusts the spectrum usage strategy according to the real-time situation of the interference source and the target signal to avoid conflicts of spectrum resources;

[0095] Interference source prediction module: combines historical data and real-time signal changes to predict the possible trend of the interference source;

[0096] Anti-interference strategy generation module: automatically generates a reasonable interference avoidance strategy based on interference identification and spectrum management information;

[0097] Adaptive adjustment module: dynamically adjusts the avoidance strategy according to system performance feedback and real-time environmental changes;

[0098] Intelligent Decision Support Module: Combining machine learning and data analysis algorithms, it provides optimal interference avoidance decisions based on different environments and task requirements.

[0099] Feedback Monitoring Module: Real-time monitoring of the system's interference avoidance effect and feeding back data to the decision support module for adjustment.

[0100] User Interface and Reporting Module: Provides system operation interface and monitoring panel, showing interference avoidance status, spectrum usage, etc.

[0101] Communication Coordination Module: Ensures smooth information transmission and collaborative work between devices and modules within and outside the system.

[0102] Specifically, the signal detection module includes multiple sensor arrays that can cover a wider area, providing more signal collection points and improving the accuracy and reliability of interference signal detection. Each sensor can collect data from different angles and positions, better identifying and locating the interference source. The diversity of data collected by multiple sensor arrays helps more accurate analysis of interference sources, and redundant data also improves the system's fault tolerance, ensuring overall system operation even if some sensors fail. Multi-point monitoring by multiple sensor arrays can effectively reduce performance degradation caused by single sensors limited by physical location or interference factors, enhancing the system's anti-interference ability.

[0103] The communication coordination module supports real-time communication and data synchronization between different modules in the system. Real-time communication and data synchronization ensure that each module can quickly respond after obtaining the latest information, improving the overall system's response speed and decision-making efficiency. Seamless exchange of information between modules allows closer collaboration, avoiding decision-making errors or resource conflicts caused by information lag or inconsistency. Through real-time communication and data synchronization, the system can dynamically adjust interference avoidance strategies based on the latest interference information, environmental changes, and task requirements, ensuring more accurate and effective avoidance measures.

[0104] In summary, the advantages of the present application are: through multi-dimensional feature fusion technology, the identification accuracy of interference sources is significantly improved; based on time series modeling and long short-term memory network, the trend of interference strength change is actively predicted; innovatively, four-dimensional dynamic avoidance strategies of frequency switching, beamforming, power adjustment, and time slot allocation are combined; and combined with reinforcement learning mechanism, the strategy weight is dynamically optimized according to real-time communication quality feedback, solving the problems of limited identification ability, response lag, and isolated strategies of traditional methods from the root, realizing precise suppression of interference signals, real-time and adaptability of interference avoidance, maximizing spectrum utilization efficiency and reducing bit error rate.

[0105] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only the principles of the present application. Various changes and improvements can be made without departing from the spirit and scope of the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method of radio system interference handling, characterized by, Comprise: Real-time environmental monitoring and data collection: using various sensors for real-time monitoring of the radio environment, collecting spectrum, interference signal strength, signal quality key information, performing spectrum analysis on all radio signals existing in the environment, collecting and analyzing the usage of different frequency bands, judging the signal quality and interference strength of each frequency band; Interference source detection and accurate identification: through signal analysis algorithm processing the collected signal, determining the possible interference source, using machine learning algorithm combined with historical interference data to classify the signal, accurately identifying the type of interference source and its interference frequency, strength, periodicity, on this basis, positioning analysis of the interference source; Dynamic prediction and trend analysis of interference sources: based on machine learning or deep learning algorithms, combined with real-time monitoring data, historical interference data and environmental factors, trend analysis of the dynamic changes of interference sources, predicting the change of interference strength in advance, specifically, the future behavior of the interference source is predicted by the following steps: Collecting interference strength data continuously in minutes, organizing into time series data; Building a long short-term memory prediction model, which contains two hidden layers, designed to capture the long-term temporal dependence of interference signals; Input the historical interference strength time series data containing 10 consecutive time points into the trained LSTM model for inference; The LSTM model outputs the predicted value of the interference strength after k time steps; The model output weight and environmental factor bias term are optimized and adjusted by the back propagation algorithm, the environmental factors include weather, geographical location and device density, which are integrated into the prediction model through feature encoding; Multi-dimensional interference avoidance strategy generation: based on the positioning analysis results of the interference source and the predicted trend of the interference strength, dynamically generate combined strategies from four dimensions of frequency, space, power and time domain, including: selecting idle frequency bands according to the frequency distribution of interference sources, avoiding interference direction by adjusting antenna pattern through beamforming technology, reducing transmission power in interference frequency bands, and allocating standby time slots for communication during high interference periods; Real-time feedback and intelligent decision optimization: during system operation, the performance data of the system is collected in real time through sensors and feedback modules, including communication error rate, signal-to-noise ratio and spectrum utilization; Collaboration and system integration: modules interact with each other through standardized interfaces to form a highly integrated system, and different avoidance strategies are intelligently switched or combined according to environmental and interference changes; Final performance evaluation and adjustment: after the system runs for a period of time, the interference avoidance effect is comprehensively evaluated based on the expected return, and the interference identification algorithm, avoidance strategy and feedback mechanism of the system are further optimized according to the evaluation results.

2. A method of radio system interference handling according to claim 1, characterized in that, In the real-time environmental monitoring and data collection, further specifically includes classifying different types of radio signals and performing accurate spectrum analysis according to the signal type.

3. A method of radio system interference handling according to claim 2, characterized in that, The interference source detection and accurate identification specifically includes: Extracting multi-dimensional feature vectors from signals: extracting signal amplitude fluctuation features from time domain and arrival angle features from spatial domain to form multi-dimensional feature vectors; Statistical characteristic parameters of each type of interference source are obtained by training historical interference sample data, and the statistical characteristic parameters include a mean vector and a covariance matrix; A posterior probability that the signal belongs to a specific interference source category is calculated under the condition of the extracted current signal characteristic vector; The interference source is identified and its position is located based on the calculated posterior probability.

4. A method of radio system interference handling according to claim 3, characterized in that, In the multi-dimensional interference avoidance strategy generation, the strategy generation logic is: when a sudden interference is predicted, time slot allocation is preferentially enabled; frequency avoidance is adopted for narrowband interference; beamforming is started for directional interference; power control is triggered only when the interference intensity exceeds a threshold.

5. A method of radio system interference handling according to claim 4, characterized in that, In the step of system integration, the modules interact with each other through standardized interfaces.

6. A method of radio system interference handling according to claim 5, characterized in that, The real-time feedback and intelligent decision optimization specifically includes: When the performance index deviates from the preset threshold, a strategy adjustment mechanism is triggered; An immediate reward is defined, which is the weighted sum of the spectrum utilization rate improvement ratio and the bit error rate reduction ratio, reflecting the immediate effect of the current interference avoidance strategy; An expected return of the strategy is defined, which is the sum of the immediate rewards at all future time points after attenuation by a discount factor, reflecting the overall long-term performance of the strategy; A reinforcement learning model is constructed, which uses a policy gradient method to dynamically update the strategy parameters according to the real-time feedback performance data and reward signals, and continuously optimizes and adjusts the avoidance strategy.

7. A radio system interference handling device characterized by The method for implementing the interference processing method of the radio system as claimed in any one of claims 1-6, comprising: a signal detection module: responsible for real-time monitoring and collection of signals in the radio spectrum, and distinguishing interference signals from target signals; an interference identification and analysis module: analyzing the collected signal data, identifying and classifying interference sources through feature extraction and pattern recognition techniques; a dynamic spectrum management module: adjusting spectrum usage strategies based on real-time conditions of interference sources and target signals to avoid conflicts of spectrum resources; an interference source prediction module: combining historical data and real-time signal changes to predict possible trends of interference sources; an anti-interference strategy generation module: automatically generating reasonable interference avoidance strategies based on interference identification and spectrum management information; an adaptive adjustment module: dynamically adjusting avoidance strategies according to system performance feedback and real-time environmental changes; an intelligent decision support module: combining machine learning and data analysis algorithms to provide optimal interference avoidance decisions according to different environments and task requirements; a feedback monitoring module: real-time monitoring of the interference avoidance effect of the system and feeding back data to the decision support module for adjustment; a user interface and reporting module: providing system operation interface and monitoring panel to display interference avoidance status and spectrum usage information; a communication coordination module: ensuring smooth information transmission and collaborative work between devices and modules within and outside the system.

8. A radio system interference handling device according to claim 7, characterized in that, The signal detection module includes a plurality of sensor arrays.

9. A radio system interference handling device according to claim 7, characterized in that, The communication coordination module supports real-time communication and data synchronization between different modules within the system.

Images