Unmanned aerial vehicle frequency system identification system based on AI intelligent analysis

Abstract

The invention relates to the technical field of radio signal processing, in particular to an unmanned aerial vehicle frequency system identification system based on AI intelligent analysis. The method comprises the steps that a signal acquisition unit captures potential radio signals in an environment based on a radio frequency receiver module; the information processing unit performs data processing by using a digital signal processing technology based on the potential radio signals acquired by the signal acquisition unit; the feature extraction unit performs feature matching and classification through a Bayesian algorithm based on the potential radio signals processed by the information processing unit; and the frequency modulation unit selects a frequency and a modulation mode to generate a new signal based on a classification result of the feature extraction unit. By using a high-performance radio frequency receiver module, weak signals can be captured in a low signal-to-noise ratio environment. Different frequencies can be quickly adjusted, and the signal capturing efficiency and response speed are improved. Through an advanced digital signal processing technology, background noise is effectively removed, and the signal-to-noise ratio is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of radio signal processing, and in particular to a UAV frequency system identification system based on AI intelligent analysis. Background Art

[0002] The drone frequency system identification system based on AI intelligent analysis is a comprehensive solution that integrates advanced artificial intelligence, signal processing, radio spectrum management, sensor technology, data fusion, network security, cloud computing, geographic information system (GIS), and automatic control. The system is designed to monitor and capture the radio signals emitted by drones in real time by deploying highly sensitive radio frequency receiving equipment, and use machine learning algorithms to conduct in-depth analysis of these signals to automatically identify the type, model, flight status and operator information of the drone.

[0003] In the existing drone frequency system recognition system based on AI intelligent analysis, the training data set is not comprehensive or representative enough, which may lead to recognition errors, especially for new models or uncommon types of drones, whose signal characteristics may not be fully learned, thus affecting the recognition rate. The radio signal environment is complex and changeable, and there are many interference sources, such as other radio equipment, natural noise, etc. These factors may reduce the sensitivity and clarity of signal capture, thereby weakening the detection ability of the system. Due to the diversity of drone communication protocols and the existence of encryption measures, it is difficult to effectively decode or parse the signal content in some cases, which limits the in-depth understanding of the drone's operating intentions. Therefore, a drone frequency system recognition system based on AI intelligent analysis is designed. Summary of the invention

[0004] The purpose of the present invention is to provide a UAV frequency system identification system based on AI intelligent analysis, so as to solve the problem raised in the above background technology that in the existing UAV frequency system identification system based on AI intelligent analysis, the training data set is not comprehensive or representative enough, which may lead to recognition errors, especially for new models or uncommon types of UAVs, whose signal characteristics may not be fully learned, thus affecting the recognition rate. The radio signal environment is complex and changeable, and there are many interference sources, such as other radio equipment, natural noise, etc. These factors may reduce the sensitivity and clarity of signal capture, thereby weakening the detection ability of the system. Due to the diversity of UAV communication protocols and the existence of encryption measures, it is difficult to effectively decode or parse the signal content in some cases, which limits the in-depth understanding of the UAV operation intention.

[0005] To achieve the above-mentioned object, the present invention aims to provide a UAV frequency system identification system based on AI intelligent analysis, including a signal acquisition unit, which captures potential radio signals in the environment based on a radio frequency receiver module; An information processing unit, wherein the information processing unit performs data processing using a digital signal processing technology based on the potential radio signal collected by the signal collection unit; A feature extraction unit, wherein the feature extraction unit performs feature matching and classification through a Bayesian algorithm based on the potential radio signal processed by the information processing unit; The frequency modulation unit selects a frequency and a modulation method to generate a new signal based on the classification result of the feature extraction unit.

[0006] As a further improvement of the technical solution, the signal acquisition unit includes a radio frequency receiver module, and the radio frequency receiver module is used to capture potential radio signals in the environment through multi-band scanning and fast tuning; Among them, potential radio signals include frequency, frequency band and power.

[0007] As a further improvement of the technical solution, the specific steps involved in performing data processing using digital signal processing technology based on the potential radio signal collected by the signal collection unit (1) are as follows: S3.1. Apply a bandpass filter to remove unnecessary frequency components and retain the signal within the target frequency band to reduce noise interference; Among them, the expression involved in applying a bandpass filter to remove unnecessary frequency components and retaining the signal within the target frequency band is: ; in, represents the coefficients of the filter; Indicates the length of the filter; represents a discrete time index variable; represents the input signal; represents discrete time; Indicates input signal In time index The value at S3.2, dynamically adjust the gain according to the received signal strength to ensure that the signal is within the strength range to avoid overload and distortion; The expression involved in dynamically adjusting the gain is: ; in, Represents the power of the input signal; represents the target power level; Indicates gain; S3.3, converting the time domain signal into the frequency domain signal by means of a fast Fourier transform algorithm; Among them, the expression involved in converting the time domain signal into the frequency domain signal through the fast Fourier transform algorithm is: ; in, Indicates the number of signal samples; represents a frequency domain signal; represents an imaginary unit; S3.4, extracting amplitude spectrum and phase spectrum from frequency domain signal; S3.5. Use the Welch method to estimate the power spectral density of the signal to evaluate the energy distribution and noise level of the signal; Among them, the expression involved in estimating the power spectral density of the signal using the Welch method is: ; in, represents the estimated power spectral density; Indicates the number of segments; Indicates the paragraph number; Indicates Periodogram of the segment; S3.6, using an adaptive filter to suppress noise on the estimated signal; Among them, the expression involved in using the adaptive filter to suppress the noise of the estimated signal is: ; in, represents the step size parameter; represents the weight vector of the adaptive filter; represents the error signal; Represents the weight vector of the adaptive filter at time index The value at S3.7. Convert the noise-suppressed frequency domain signal back to a time domain signal using an inverse fast Fourier transform algorithm.

[0008] As a further improvement of the technical solution, in S3.4, the process of extracting the amplitude spectrum and the phase spectrum from the frequency domain signal is: Extract the magnitude spectrum: ; in, Represents frequency domain signal The amplitude spectrum of express The real part of express The imaginary part of Extract the phase spectrum: ; in, Represents frequency domain signal The phase spectrum.

[0009] As a further improvement of the technical solution, in S3.7, the expression involved in converting the noise-suppressed frequency domain signal back to the time domain signal by the inverse fast Fourier transform algorithm is: ; in, represents the time domain signal after inverse fast Fourier transform; Represents the frequency domain signal after noise suppression.

[0010] As a further improvement of the technical solution, the steps involved in performing feature matching and classification by using a Bayesian algorithm based on the potential radio signal processed by the information processing unit are: S6.1, selecting potential radio signal features from the processed signal; S6.2, extracting key parameters from selected potential radio signal features; Among them, key parameters include frequency and power; S6.3, normalize features of different scales to the same range; S6.4, match the extracted features with the predefined rule base through the SVM algorithm to find the most similar category; S6.5. Build a rule base to guide the feature matching and classification process through the Bayesian algorithm.

[0011] Specifically, the process of building a rule base includes collecting and annotating a large amount of training data, extracting key features from it, and using the Bayesian algorithm for training to generate a series of classification rules. These rules describe different categories of signals based on the conditions of the features and are stored in a data structure that is easy to query and update. The rule base not only contains category labels and feature descriptions, but also integrates probability models and support vector machine model parameters to ensure the efficiency and accuracy of classification, and can be dynamically updated and optimized based on new data.

[0012] As a further improvement of the technical solution, in S6.4, the extracted features are matched with the predefined rule base by the SVM algorithm, and the matching process involved in finding the most similar category is: Compute the Euclidean distance between two eigenvectors: ; in, and represents two eigenvectors; Represents the feature vector In the The value of the dimension; Represents the feature vector In the The values ​​in the dimensions; Indicates the number of dimensions of the feature vector; Represents the feature vector and The Euclidean distance between Represents dimension index variable; Measure the cosine of the angle between two eigenvectors: ; in, Represents the feature vector and The cosine value between ; Represents the feature vector The Euclidean norm of ; Represents the feature vector The Euclidean norm of ; Represents the feature vector and The dot product of Consider the covariance structure between features: ; in, Represents the feature vector and The Mahalanobis distance between represents the covariance matrix; Represents matrix transpose.

[0013] As a further improvement of the technical solution, in S6.5, the process involved in constructing a rule base and guiding the feature matching and classification process through the Bayesian algorithm is: Handling uncertainty and ambiguity for partial matching and probabilistic classification; By using Bayes' theorem, calculate the posterior probability of each category: ; in, represents the posterior probability; to indicate evidence; represents the prior probability; represents the likelihood function.

[0014] As a further improvement of the technical solution, the specific steps involved in selecting the frequency and modulation mode to generate a new signal based on the classification result of the feature extraction unit are: S9.1. Obtain the classification result from the feature extraction unit and determine the category to which the signal belongs; S9.2. performing spectrum scanning to select a carrier frequency according to the classification result; S9.3. Select a modulation method according to the selected carrier frequency; Among them, modulation methods include frequency modulation, phase modulation and amplitude modulation; S9.4. Generate a new modulation signal according to the selected carrier frequency and modulation method.

[0015] As a further improvement of the technical solution, in S9.4, the process of generating a new modulation signal according to the selected carrier frequency and modulation mode is as follows: Frequency modulation: Transmitting information by changing the frequency of a carrier signal; ; in, represents the output signal after frequency modulation; Indicates the carrier amplitude; Indicates the carrier frequency; represents the modulation index; represents the modulation signal; Indicates time; Phase modulation: information is transmitted by changing the phase of the carrier signal; ; in, represents the output signal after phase modulation; Represents the phase change caused by the modulating signal; Amplitude modulation: information is transmitted by changing the amplitude of the carrier signal; ; in, represents the output signal after amplitude modulation; represents the modulation coefficient; Orthogonal frequency division multiplexing: divides the signal into multiple subcarriers to transmit information; ; in, Represents the output signal after orthogonal frequency division multiplexing; Indicates the number of subcarriers; Indicates Modulation symbols on subcarriers; Indicates the subcarrier number; Indicates the subcarrier spacing frequency.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. This drone frequency system identification system based on AI intelligent analysis uses a high-performance RF receiver module that can capture weak signals in a low signal-to-noise ratio environment. It can quickly adjust to different frequencies to improve the efficiency and response speed of signal capture. Through advanced digital signal processing technology, it can effectively remove background noise and improve the signal-to-noise ratio. It can quickly process large amounts of data, realize real-time monitoring and analysis, and meet the needs of immediate response.

[0017] 2. In this drone frequency system identification system based on AI intelligent analysis, the expert system rule base is used to achieve high-precision classification of drone communication signals and reduce false positives and negatives. It can adapt to different types of drone signals and support multiple modulation methods and communication protocols. It can accurately reproduce the identified drone communication signals for verification and testing system performance. It supports multiple modulation methods and is suitable for different application scenarios and technical requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is the overall flow chart of the present invention; The meaning of each number in the figure is: 1. Signal acquisition unit; 11. RF receiver module; 2. Information processing unit; 3. Feature extraction unit; 4. Frequency modulation unit. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Example

[0021] See also Figure 1 As shown, a UAV frequency system identification system based on AI intelligent analysis is provided, including a signal acquisition unit 1, which captures potential radio signals in the environment based on a radio frequency receiver module 11; In this example, the signal acquisition unit 1 includes a radio frequency receiver module 11 , and the radio frequency receiver module 11 is used to capture potential radio signals in the environment through multi-band scanning and fast tuning.

[0022] Specifically, in the drone frequency system identification system based on AI intelligent analysis, the signal acquisition unit 1, as the front-end sensor of the system, can efficiently capture various potential radio signals that may be used by drones in the environment by performing multi-band scanning and fast tuning through its built-in RF receiver module 11. This function is crucial because it provides the raw data basis for subsequent information processing, feature extraction, and final frequency and modulation mode identification, ensuring that the system can accurately monitor and identify the communication activities of drones, thereby achieving effective supervision and safety control of drones in the airspace.

[0023] Among them, potential radio signals include frequency, frequency band and power.

[0024] The drone frequency system identification system based on AI intelligent analysis also includes an information processing unit 2, which uses digital signal processing technology to process data based on the potential radio signals collected by the signal collection unit 1; In this example, based on the potential radio signals collected by the signal collection unit 1, the specific steps involved in data processing using digital signal processing technology are as follows: S3.1. Apply a bandpass filter to remove unnecessary frequency components and retain the signal within the target frequency band to reduce noise interference; Among them, the expression involved in applying a bandpass filter to remove unnecessary frequency components and retaining the signal within the target frequency band is: ; in, represents the coefficients of the filter; Indicates the length of the filter; represents a discrete time index variable; represents the input signal; represents discrete time; Indicates input signal In time index The value at S3.2, dynamically adjust the gain according to the received signal strength to ensure that the signal is within the strength range to avoid overload and distortion; The expression involved in dynamically adjusting the gain is: ; in, Represents the power of the input signal; represents the target power level; Indicates gain; S3.3, converting the time domain signal into the frequency domain signal by means of a fast Fourier transform algorithm; Among them, the expression involved in converting the time domain signal into the frequency domain signal through the fast Fourier transform algorithm is: ; in, Indicates the number of signal samples; represents a frequency domain signal; represents an imaginary unit; S3.4, extracting amplitude spectrum and phase spectrum from frequency domain signal; In this example, the process of extracting the amplitude spectrum and phase spectrum from the frequency domain signal is: Extract the magnitude spectrum: ; in, Represents frequency domain signal The amplitude spectrum of express The real part of express The imaginary part of Extract the phase spectrum: ; in, Represents frequency domain signal The phase spectrum.

[0025] Specifically, the extraction of amplitude spectrum and phase spectrum not only provides key data support for subsequent feature extraction and pattern recognition, but also enhances the overall performance of the system, enabling it to more accurately and reliably monitor and analyze the communication activities of drones and ensure the safety and order of the airspace.

[0026] S3.5. Use the Welch method to estimate the power spectral density of the signal to evaluate the energy distribution and noise level of the signal; Among them, the expression involved in estimating the power spectral density of the signal using the Welch method is: ; in, represents the estimated power spectral density; Indicates the number of segments; Indicates the paragraph number; Indicates Periodogram of the segment; S3.6, using an adaptive filter to suppress noise on the estimated signal; Among them, the expression involved in using the adaptive filter to suppress the noise of the estimated signal is: ; in, represents the step size parameter; represents the weight vector of the adaptive filter; represents the error signal; Represents the weight vector of the adaptive filter at time index The value at S3.7. Convert the noise-suppressed frequency domain signal back to a time domain signal using an inverse fast Fourier transform algorithm.

[0027] In this example, the expression involved in converting the noise-suppressed frequency domain signal back to the time domain signal through the inverse fast Fourier transform algorithm is: ; in, represents the time domain signal after inverse fast Fourier transform; Represents the frequency domain signal after noise suppression.

[0028] Specifically, the IFFT step restores the denoised and optimized frequency domain signal to a time domain signal, allowing the signal to restore its original time characteristics while retaining the signal quality enhanced during the noise suppression process. This is crucial for subsequent signal analysis, demodulation, and feature extraction, as it ensures that the processed signal maintains temporal coherence while eliminating unnecessary interference. This step not only helps to improve the accuracy and reliability of UAV communication signal identification, but also provides high-quality data support for real-time monitoring and response of the system, thereby enhancing the adaptability and performance of the entire system in complex electromagnetic environments. In addition, the time domain signal can be directly used for interaction or output with other system components, such as for further decoding, display, or recording, ensuring the integrity and accuracy of information transmission.

[0029] Apply bandpass filters to remove noise, dynamically adjust gain to optimize signal strength, use fast Fourier transform (FFT) to convert time domain signals to frequency domain signals, extract amplitude spectrum and phase spectrum, use Welch method to estimate power spectrum density to evaluate signal energy distribution and noise level, and use adaptive filters for noise suppression, and finally convert the processed frequency domain signal back to time domain signal through inverse fast Fourier transform (IFFT). This series of processing not only significantly improves the quality of the signal and reduces the impact of background noise, but also provides a clearer and more accurate data basis for subsequent feature extraction and pattern recognition, thereby enhancing the system's ability to detect, classify and identify drone communication signals, ensuring the effectiveness and accuracy of airspace safety monitoring.

[0030] The drone frequency system identification system based on AI intelligent analysis also includes a feature extraction unit 3, which performs feature matching and classification based on the potential radio signal processed by the information processing unit 2 through a Bayesian algorithm; In this example, the steps involved in feature matching and classification by Bayesian algorithm based on the potential radio signal processed by the information processing unit 2 are: S6.1, selecting potential radio signal features from the processed signal; Specifically, potential radio signal characteristics are selected including: Statistical characteristics: such as mean, variance, peak, kurtosis, etc.

[0031] Time domain characteristics: such as zero-crossing rate, energy, autocorrelation function, etc.

[0032] Frequency domain characteristics: such as power spectral density (PSD), frequency band energy, main frequency, etc.

[0033] Time-frequency domain features: such as wavelet transform coefficients, short-time Fourier transform (STFT) coefficients, etc.

[0034] S6.2, extracting key parameters from selected potential radio signal features; Among them, key parameters include frequency and power; Specifically, the expressions involved in extracting key parameters from the selected potential radio signal features are: ; ; in, represents the eigenvector of the covariance matrix; represents the covariance matrix; Indicates the importance of the corresponding eigenvector; represents the transpose of the feature matrix; represents the feature matrix; S6.3, normalize features of different scales to the same range; Specifically, the expression involved in normalizing features of different scales to the same range is: ; in, represents the normalized eigenvalue; Indicates the minimum value of the feature; Indicates the maximum value of the feature; S6.4, match the extracted features with the predefined rule base through the SVM algorithm to find the most similar category; In this example, the extracted features are matched with the predefined rule base through the SVM algorithm, and the matching process involved in finding the most similar category is: Compute the Euclidean distance between two eigenvectors: ; in, and represents two eigenvectors; Represents the feature vector In the The values ​​in the dimensions; Represents the feature vector In the The values ​​in the dimensions; Indicates the number of dimensions of the feature vector; Represents the feature vector and The Euclidean distance between Represents dimension index variable; Measure the cosine of the angle between two eigenvectors: ; in, Represents the feature vector and The cosine value between ; Represents the feature vector The Euclidean norm of ; Represents the feature vector The Euclidean norm of ; Represents the feature vector and The dot product of Consider the covariance structure between features: ; in, Represents the feature vector and The Mahalanobis distance between represents the covariance matrix; Represents matrix transpose.

[0035] Specifically, the Euclidean distance It is used to measure the straight-line distance between two feature vectors in multidimensional space. It helps the system quickly filter out the known category that is geometrically closest to the unknown signal and is suitable for situations where the feature distribution is relatively uniform. By calculating the angle between two eigenvectors, it reflects their directional consistency rather than their absolute position. This helps to identify signals with similar patterns but different scales, especially for processing situations with large amplitude variations, such as the same type of signal at different power levels. The covariance structure between features is taken into account, which can adapt to the correlation and variance differences between different feature dimensions. This is particularly important for processing complex and changeable drone communication signals, because it can effectively deal with scenes with uneven feature distribution or strong correlation, and provide more robust classification results.

[0036] S6.5. Build a rule base to guide the feature matching and classification process through the Bayesian algorithm.

[0037] In this example, the process of building a rule base and guiding the feature matching and classification process through the Bayesian algorithm is as follows: Specifically, the IF-THEN rule is a logical structure used to express the relationship between conditions and results, and is widely used in expert systems and rule engines. Its basic form is "If (IF) a certain condition is true, then (THEN) a specific action is performed or a certain conclusion is drawn". This rule clearly defines the actions to be taken or the results to be inferred when specific conditions are met, allowing the system to automatically make decisions or inferences based on input data. For example, in the drone frequency system identification system, the IF-THEN rule can be used to express judgment logic such as "If the spectral characteristics of the signal conform to a specific pattern, then the signal is likely to come from a certain model of drone", thereby guiding the system's feature matching and classification process.

[0038] Handling uncertainty and ambiguity for partial matching and probabilistic classification; Specifically, the expert's judgment logic is converted into IF-THEN rules. For example: Rule 1: If feature A > threshold X and feature B < threshold Y, then classify as category 1.

[0039] Rule 2: If feature C = threshold Z, then classify as category 2.

[0040] For uncertain situations, fuzzy logic or probabilistic rules can be used. For example: Fuzzy rules: If feature A is close to threshold X, then the signal may belong to class 1 or class 2, depending on other features.

[0041] Probability rule: If the value of feature A is x, then the probability of belonging to class 1 is P(class 1|x).

[0042] Set priorities for different rules to ensure that the system can select the most appropriate classification when multiple rules are triggered at the same time.

[0043] By using Bayes' theorem, calculate the posterior probability of each category: ; in, represents the posterior probability; to indicate evidence; represents the prior probability; represents the likelihood function.

[0044] Specifically, we collect expertise about signal features and classification rules from domain experts, and transform this knowledge and judgment logic into specific IF-THEN rules to form a structured rule base. This enables the system to make decisions based on proven best practices. By introducing fuzzy logic and probabilistic models to handle partial matches and uncertainty, the system is allowed to make reasonable inferences when faced with incomplete or ambiguous information, improving the adaptability and flexibility of the system. We use Bayes' theorem to calculate the posterior probability of each category, taking into account the match between signal features and known categories, and combining prior knowledge to provide more accurate and reliable classification results.

[0045] By integrating expert knowledge and automated algorithms, the system not only improves the accuracy and reliability of recognition, but also copes with complex and changing electromagnetic environments and enhances the ability to process unknown or abnormal signals. This provides solid technical support for the monitoring, management and safety control of drones, ensuring the safety and order of the airspace. In addition, intelligent recognition based on expert systems also gives the system greater adaptability and flexibility, enabling it to maintain efficient operation in ever-changing application scenarios.

[0046] The unmanned aerial vehicle frequency system identification system based on AI intelligent analysis also includes a frequency modulation unit 4, which selects a frequency and a modulation method to generate a new signal based on the classification result of the feature extraction unit 3.

[0047] In this example, based on the classification result of the feature extraction unit 3, the specific steps involved in selecting the frequency and modulation mode to generate a new signal are: S9.1, obtaining the classification result from the feature extraction unit 3, and determining the category to which the signal belongs; S9.2. performing spectrum scanning to select a carrier frequency according to the classification result; S9.3. Select a modulation method according to the selected carrier frequency; Among them, modulation methods include frequency modulation, phase modulation and amplitude modulation; S9.4. Generate a new modulation signal according to the selected carrier frequency and modulation method.

[0048] In this example, the process involved in generating a new modulated signal according to the selected carrier frequency and modulation method is: Frequency modulation: Transmitting information by changing the frequency of a carrier signal; ; in, represents the output signal after frequency modulation; Indicates the carrier amplitude; Indicates the carrier frequency; represents the modulation index; represents the modulation signal; Indicates time; Phase modulation: information is transmitted by changing the phase of the carrier signal; ; in, represents the output signal after phase modulation; Represents the phase change caused by the modulating signal; Amplitude modulation: Transmits information by changing the amplitude of the carrier signal; ; in, represents the output signal after amplitude modulation; represents the modulation coefficient; Orthogonal frequency division multiplexing: divides the signal into multiple subcarriers to transmit information; ; in, Represents the output signal after orthogonal frequency division multiplexing; Indicates the number of subcarriers; Indicates Modulation symbols on subcarriers; Indicates the subcarrier number; Indicates the subcarrier spacing frequency.

[0049] Specifically, frequency modulation transmits information by changing the frequency of the carrier signal. This modulation method is suitable for communication scenarios that require high anti-interference capabilities and can effectively simulate and test FM communication signals used by drones. Phase modulation transmits information by changing the phase of the carrier signal. Phase modulation is commonly used in digital communications and can provide higher data transmission rates and reliability, which helps to accurately reproduce the PM communication characteristics of drones. Amplitude modulation transmits information by changing the amplitude of the carrier signal. Amplitude modulation is simple and intuitive, suitable for simulating traditional radio communication signals, and helps the system understand and respond to drone equipment using AM. Orthogonal frequency division multiplexing divides the signal into multiple subcarriers for transmission. OFDM technology improves spectrum utilization and resistance to multipath fading. It is widely used in modern wireless communication systems and can simulate complex drone communication protocols such as Wi-Fi, LTE, etc.

[0050] This process plays a vital role in the drone frequency system identification system. It can not only accurately reconstruct and simulate the identified drone communication signals, but also provide technical support for the system's testing, verification and countermeasures. By accurately generating new signals that match the original signals, the system can be used to simulate drone communication activities in real scenarios, helping researchers and engineers evaluate the system's detection capabilities and response strategies. In addition, this function can also be used to develop and test new defense mechanisms to ensure that the airspace security monitoring system can effectively respond to various types of drone threats. In short, this step enhances the versatility and practicality of the system, enabling it to maintain efficient operation in a complex and changing electromagnetic environment, and provides strong technical support for the management and safety control of drones.

[0051] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only preferred examples of the present invention, and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. The UAV frequency system identification system based on AI intelligent analysis is characterized by: include: A signal acquisition unit (1), wherein the signal acquisition unit (1) captures potential radio signals in the environment based on a radio frequency receiver module (11); An information processing unit (2), wherein the information processing unit (2) performs data processing using digital signal processing technology based on the potential radio signal collected by the signal collection unit (1); A feature extraction unit (3), wherein the feature extraction unit (3) performs feature matching and classification based on the potential radio signal processed by the information processing unit (2) by using a Bayesian algorithm; A frequency modulation unit (4) is configured to select a frequency and a modulation method to generate a new signal based on the classification result of the feature extraction unit (3).

2. The UAV frequency system identification system based on AI intelligent analysis according to claim 1 is characterized by: The signal acquisition unit (1) comprises a radio frequency receiver module (11), and the radio frequency receiver module (11) is used to capture potential radio signals in the environment through multi-band scanning and fast tuning; Among them, potential radio signals include frequency, frequency band and power.

3. The UAV frequency system identification system based on AI intelligent analysis according to claim 1 is characterized by: The specific steps involved in performing data processing using digital signal processing technology based on the potential radio signal collected by the signal collection unit (1) are as follows: S3.

1. Apply a bandpass filter to remove unnecessary frequency components and retain the signal within the target frequency band to reduce noise interference; Among them, the expression involved in applying a bandpass filter to remove unnecessary frequency components and retaining the signal within the target frequency band is: ； in, represents the coefficients of the filter; Indicates the length of the filter; represents a discrete time index variable; represents the input signal; represents discrete time; Indicates input signal In time index The value at S3.2, dynamically adjust the gain according to the received signal strength to ensure that the signal is within the strength range to avoid overload and distortion; The expression involved in the dynamic adjustment gain is: ； in, Represents the power of the input signal; Indicates the target power level; Indicates gain; S3.3, converting the time domain signal into the frequency domain signal by means of a fast Fourier transform algorithm; Among them, the expression involved in converting the time domain signal into the frequency domain signal through the fast Fourier transform algorithm is: ； in, Indicates the number of signal samples; represents a frequency domain signal; represents an imaginary unit; S3.4, extracting amplitude spectrum and phase spectrum from frequency domain signal; S3.

5. Use the Welch method to estimate the power spectral density of the signal to evaluate the energy distribution and noise level of the signal; Among them, the expression involved in estimating the power spectral density of the signal using the Welch method is: ； in, represents the estimated power spectral density; Indicates the number of segments; Indicates the paragraph number; Indicates Periodogram of the segment; S3.6, using an adaptive filter to suppress noise on the estimated signal; Among them, the expression involved in using the adaptive filter to suppress the noise of the estimated signal is: ； in, represents the step size parameter; represents the weight vector of the adaptive filter; represents the error signal; Represents the weight vector of the adaptive filter at time index The value at S3.

7. Convert the noise-suppressed frequency domain signal back to a time domain signal using an inverse fast Fourier transform algorithm.

4. The UAV frequency system identification system based on AI intelligent analysis according to claim 3 is characterized by: In S3.4, the process of extracting the amplitude spectrum and phase spectrum from the frequency domain signal is: Extract the magnitude spectrum: ； in, Represents frequency domain signal The amplitude spectrum of express The real part of express The imaginary part of Extract the phase spectrum: ； in, Represents frequency domain signal The phase spectrum.

5. The UAV frequency system identification system based on AI intelligent analysis according to claim 3 is characterized by: In S3.7, the expression involved in converting the noise-suppressed frequency domain signal back to the time domain signal by the inverse fast Fourier transform algorithm is: ； in, represents the time domain signal after inverse fast Fourier transform; Represents the frequency domain signal after noise suppression.

6. The UAV frequency system identification system based on AI intelligent analysis according to claim 1 is characterized by: The steps involved in performing feature matching and classification based on the potential radio signal processed by the information processing unit (2) using the Bayesian algorithm are as follows: S6.1, selecting potential radio signal features from the processed signal; S6.2, extracting key parameters from selected potential radio signal features; Among them, key parameters include frequency and power; S6.3, normalize features of different scales to the same range; S6.4, match the extracted features with the predefined rule base through the SVM algorithm to find the most similar category; S6.

5. Build a rule base to guide the feature matching and classification process through the Bayesian algorithm.

7. The UAV frequency system identification system based on AI intelligent analysis according to claim 6 is characterized by: In S6.4, the extracted features are matched with the predefined rule base by the SVM algorithm, and the matching process involved in finding the most similar category is: Compute the Euclidean distance between two eigenvectors: ； in, and represents two eigenvectors; Represents the feature vector In the The value of the dimension; Represents the feature vector In the The value of the dimension; Indicates the number of dimensions of the feature vector; Represents the feature vector and The Euclidean distance between Represents dimension index variable; Measure the cosine of the angle between two eigenvectors: ； in, Represents the feature vector and The cosine value between Represents the feature vector The Euclidean norm of ; Represents the feature vector The Euclidean norm of ; Represents the feature vector and The dot product of Consider the covariance structure between features: ； in, Represents the feature vector and The Mahalanobis distance between represents the covariance matrix; Represents matrix transpose.

8. The UAV frequency system identification system based on AI intelligent analysis according to claim 6 is characterized by: In S6.5, the process of constructing a rule base and guiding the feature matching and classification process through the Bayesian algorithm is as follows: Handling uncertainty and ambiguity for partial matching and probabilistic classification; By using Bayes' theorem, calculate the posterior probability of each category: ； in, represents the posterior probability; to indicate evidence; represents the prior probability; represents the likelihood function.

9. The UAV frequency system identification system based on AI intelligent analysis according to claim 1 is characterized by: The specific steps involved in selecting the frequency and modulation mode to generate a new signal based on the classification result of the feature extraction unit (3) are as follows: S9.1, obtaining the classification result from the feature extraction unit (3) and determining the category to which the signal belongs; S9.

2. performing spectrum scanning to select a carrier frequency according to the classification result; S9.

3. Select a modulation method according to the selected carrier frequency; Among them, modulation methods include frequency modulation, phase modulation and amplitude modulation; S9.

4. Generate a new modulation signal according to the selected carrier frequency and modulation method.

10. The UAV frequency system identification system based on AI intelligent analysis according to claim 9 is characterized by: In S9.4, the process of generating a new modulation signal according to the selected carrier frequency and modulation mode is as follows: Frequency modulation: Transmitting information by changing the frequency of a carrier signal; ； in, represents the output signal after frequency modulation; Indicates the carrier amplitude; Indicates the carrier frequency; represents the modulation index; represents the modulation signal; Indicates time; Phase modulation: information is transmitted by changing the phase of the carrier signal; ； in, represents the output signal after phase modulation; Represents the phase change caused by the modulating signal; Amplitude modulation: information is transmitted by changing the amplitude of the carrier signal; ； in, represents the output signal after amplitude modulation; represents the modulation coefficient; Orthogonal frequency division multiplexing: divides the signal into multiple subcarriers to transmit information; ； in, Represents the output signal after orthogonal frequency division multiplexing; Indicates the number of subcarriers; Indicates Modulation symbols on subcarriers; Indicates the subcarrier number; Indicates the subcarrier spacing frequency.

Images