A weak signal detection system and method for interference fingerprint divide-and-conquer compensation

Abstract

This invention discloses a weak signal detection system and method with interference fingerprint divide-and-conquer compensation, belonging to the field of weak signal detection technology. The system includes a frequency band interference fingerprint sensing module for acquiring the broadband electromagnetic signal under test, outputting frequency band coding information, interference fingerprint information, and useful signal strength information, and determining the interference fingerprint stability coefficient; a divide-and-conquer decision module for performing divide-and-conquer processing on the broadband electromagnetic signal under test, generating anti-interference strategy parameters, attenuation prediction values, and compensation parameters, and recursively updating the mapping model parameters; an anti-interference compensation module for implementing anti-interference processing and compensation processing; and a signal detection module for outputting detection results and generating feedback data. This invention is applicable to weak signal detection in complex broadband electromagnetic interference environments.

Description

Technical Field

[0001] This invention relates to the field of weak signal detection technology, and in particular to a weak signal detection system and method with interference fingerprint divide-and-conquer compensation. Background Technology

[0002] Weak signal detection is widely used in communication sensing, spectrum monitoring, and target identification in complex electromagnetic environments. In broadband electromagnetic signal environments, the signals to be detected are usually of low strength and are easily affected by external interference. In particular, when the interference exhibits characteristics such as time-varying, non-stationarity, and uneven frequency domain distribution, it can easily lead to a decrease in the accuracy and stability of weak signal detection.

[0003] In existing technologies, interference suppression, signal enhancement, or compensation processing of the received signal is typically performed to improve subsequent detection performance. However, in complex broadband interference environments, the adaptability of existing methods to interference effects across different frequency ranges remains limited. This can easily lead to insufficient interference suppression or additional impacts on the useful signal, thereby affecting weak signal detection performance.

[0004] Existing technologies are often not accurate enough in judging the degree of interference to effective signals, which can easily lead to discrepancies between the intensity of subsequent processing and actual needs. Furthermore, the linkage between detection results and front-end processing is not fully utilized, making it difficult to correct and optimize the preceding processing in a timely manner based on the detection effect. Therefore, the adaptive adjustment capability still needs to be improved in complex and dynamic interference environments.

[0005] Therefore, how to improve the adaptability to interference effects in different frequency ranges under complex, dynamic, and non-stationary broadband electromagnetic interference environments, more accurately determine the needs for weak signal recovery or enhancement, and continuously optimize the pre-processing process based on the detection results, thereby improving the accuracy, stability, and environmental adaptability of weak signal detection, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] The purpose of this invention is to provide a weak signal detection system and method with interference fingerprint divide-and-conquer compensation, in order to solve the problems of insufficient adaptability to interference effects in different frequency ranges, inaccurate judgment of weak signal recovery or enhancement requirements, and difficulty in continuously optimizing the pre-processing process based on the detection results in complex, dynamic and non-stationary broadband electromagnetic interference environments, thereby improving the accuracy, stability and environmental adaptability of weak signal detection.

[0007] To achieve the above objectives, the present invention employs the following technical solution:

[0008] On one hand, the present invention provides a weak signal detection system with interference fingerprint divide-and-conquer compensation, comprising:

[0009] The frequency band interference fingerprint sensing module is used to acquire the broadband electromagnetic signal under test, output frequency band coding information, interference fingerprint information and useful signal strength information, and determine the interference fingerprint stability coefficient based on the interference fingerprint information in multiple time windows; the frequency band coding information is coding information used to indicate the frequency boundaries or frequency band indexes of multiple divide-and-conquer frequency bands;

[0010] The divide-and-conquer decision module is used to divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on frequency band coding information, and for each divide-and-conquer frequency band: determine anti-interference strategy parameters based on interference fingerprint information and useful signal strength information; generate attenuation prediction values ​​based on the anti-interference strategy parameters and interference fingerprint information through attenuation characteristic relationships, wherein the attenuation characteristic relationships are mapping relationships between anti-interference strategy parameters, interference fingerprint information, and attenuation prediction values, and are represented by mapping model parameters; generate compensation parameters based on the attenuation prediction values, the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters, useful signal strength information, and a preset target signal strength threshold, wherein the compensation parameters are jointly determined by the attenuation prediction values, the pre-stored inherent attenuation amount, and the strength difference between the target signal strength threshold and the useful signal strength information; and recursively update the mapping model parameters based on feedback data and interference fingerprint stability coefficients.

[0011] The anti-interference compensation module is used to perform anti-interference processing on the divided frequency band signal segment of each divided frequency according to the anti-interference strategy parameters, and to perform compensation processing on the divided frequency band signal after anti-interference processing according to the compensation parameters, output the compensated divided frequency band signal, and reassemble multiple compensated divided frequency band signals into the processed signal.

[0012] The signal detection module is used to perform signal detection based on the processed signal and output the detection result, and generate feedback data based on the detection result, and output the feedback data to the divide-and-conquer decision module.

[0013] As a preferred embodiment of the present invention, the frequency band interference fingerprint sensing module includes:

[0014] The time-frequency representation selection unit is used to generate two candidate time-frequency representations with different window lengths for the signal components in the broadband electromagnetic signal under test that correspond to the same candidate analysis frequency band, and to construct a dynamic index based on the power change over time and the spectrum occupancy change over time obtained statistically from the candidate time-frequency representations; when the dynamic index is higher than a preset threshold, the short-window candidate time-frequency representation is selected as the target time-frequency representation; when the dynamic index is lower than or equal to the preset threshold, the long-window candidate time-frequency representation is selected as the target time-frequency representation.

[0015] The fingerprint generation unit is used to generate a binary interference fingerprint code based on the target time-frequency representation. The binary interference fingerprint code serves as the interference fingerprint information and includes a center frequency encoding field and a bandwidth encoding field, which are used to decode the interference center frequency and interference bandwidth.

[0016] As a preferred embodiment of the present invention, the frequency band interference fingerprint sensing module further includes:

[0017] Stability evaluation unit, used to evaluate the stability of the same divide-and-conquer frequency band in continuous The binary interference fingerprint obtained within a time window Calculate the stability coefficient of the interference fingerprint The following equation is satisfied:

[0018] ;

[0019] In the formula, To interfere with the fingerprint stability coefficient; The number of stability assessment time windows greater than 1; The time window number, and ; For the first The binary interference fingerprint code obtained from the time window; For the first The binary interference fingerprint code obtained from the time window; Let Hamming distance function be used.

[0020] Suppression control parameter update unit, used to update the stability coefficient of the interference fingerprint Determine the suppression control parameters and adaptively update the gain. The following equation is satisfied:

[0021] ;

[0022] In the formula, To suppress adaptive gain updates of control parameters; To suppress adaptive updates of the lower limit of the gain by controlling parameters; To suppress adaptive updates of the gain upper limit by controlling parameters;

[0023] The frequency domain suppression weight generation unit is used to decode the interference center frequency and interference bandwidth based on the center frequency encoding field and bandwidth encoding field, and construct the instantaneous frequency domain suppression weight matrix. and based on Frequency domain suppression weight matrix at the previous update time and We perform weighted fusion to obtain the frequency domain suppression weight matrix at the current update time. ,satisfy:

[0024] ;

[0025] In the formula, To update the time sequence number;

[0026] The frequency domain suppression weight matrix at the current update time The frequency domain suppression weight parameter, which is part of the anti-interference strategy parameters, is output to the anti-interference compensation module.

[0027] As a preferred embodiment of the present invention, the frequency band interference fingerprint sensing module further includes:

[0028] Interference suppression unit, used to suppress weight matrix in the frequency domain based on the current update time. The signal component in the corresponding candidate analysis frequency band of the broadband electromagnetic signal under test is subjected to frequency domain weighted suppression processing to generate a residual signal; the intensity estimation unit is used to form a power sample sequence based on the residual signal, and to perform statistical estimation on the power sample sequence using a preset robust statistical estimation method to obtain the power estimate and generate useful signal intensity information.

[0029] As a preferred embodiment of the present invention, the divide-and-conquer decision module includes:

[0030] The divide-and-conquer frequency band generation unit is used to divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on the frequency band coding information and generate divide-and-conquer frequency band identifiers; the strategy parameter determination unit is used to determine the anti-interference strategy parameters for the divide-and-conquer frequency bands corresponding to the divide-and-conquer frequency bands based on interference fingerprint information and useful signal strength information; the attenuation prediction unit is used to generate attenuation prediction values ​​for the divide-and-conquer frequency bands corresponding to the divide-and-conquer frequency bands based on the attenuation characteristic relationship, using the anti-interference strategy parameters and interference fingerprint information. ;

[0031] The compensation parameter generation unit is used to generate compensation parameters based on the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters for the divided frequency bands identified by the divided frequency band identifier. Attenuation prediction value Target signal strength threshold and useful signal strength information Calculate the compensation parameters to satisfy the following formula:

[0032] ;

[0033] In the formula, For dividing and conquering frequency band identification; For compensation parameters; This is a function for calculating the maximum value.

[0034] As a preferred embodiment of the present invention, the divide-and-conquer decision module further includes:

[0035] The attenuation prediction model library unit stores multiple mapping models representing attenuation characteristic relationships; the stability acquisition unit acquires the interference fingerprint stability coefficients output by the frequency band interference fingerprint sensing module; and the model selection unit selects a mapping model for the divide-and-conquer frequency band corresponding to the divide-and-conquer frequency band identifier based on the interference fingerprint stability coefficients, and calls the selected mapping model using the divide-and-conquer frequency band identifier as an index to determine the attenuation characteristic relationship of the corresponding divide-and-conquer frequency band and output the attenuation prediction value.

[0036] As a preferred embodiment of the present invention, the divide-and-conquer decision module further includes a mapping model parameter update unit, which includes: performing recursive updates on the mapping model parameters of the attenuation characteristic relationship based on feedback data; determining the measured intensity change based on the difference between the useful signal strength information before and after anti-interference processing, and determining the prediction error based on the difference between the measured intensity change and the attenuation prediction value; determining the recursive update gain lower limit and recursive update gain upper limit based on the signal-to-noise ratio change and the signal detection success rate index according to a preset gain constraint mapping relationship; and determining the recursive update gain based on the interference fingerprint stability coefficient, satisfying the following formula:

[0037] ;

[0038] In the formula, To interfere with the fingerprint stability coefficient; For the first Recursive update gain for each divide-and-conquer frequency band; To recursively update the lower bound of the gain; To recursively update the upper limit of gain;

[0039] The signal detection success rate index is an index obtained by statistics within a preset statistical window. The useful signal strength information before anti-interference processing is output by the frequency band interference fingerprint sensing module, and the useful signal strength information after anti-interference processing is output by the anti-interference compensation module.

[0040] As a preferred embodiment of the present invention, the anti-interference compensation module includes:

[0041] The frequency band processing execution unit is used to perform weighted suppression on the frequency domain components of the divided frequency band signal within the interference frequency range based on the frequency domain suppression weight matrix, which serves as an anti-interference strategy parameter, and output the divided frequency band signal after anti-interference processing.

[0042] The compensation control unit is used to obtain the interference fingerprint stability coefficient output by the stability evaluation unit, and determine the compensation weight update gain and compensation gain limiting constraint based on the interference fingerprint stability coefficient according to the preset compensation control mapping relationship.

[0043] The compensation weight generation unit is used to suppress the weight moments in the frequency domain based on the current update time. Determine the compensation scope, and define the compensation scope in relation to the frequency domain suppression weight moments at the current update time. The corresponding suppression domains are consistent, and the compensation weight matrix is ​​generated by updating the gain and compensation gain limiting constraint based on the compensation parameters and compensation weights.

[0044] The compensation execution unit is used to apply a compensation weight matrix to the divided-band signal after anti-interference processing and output the compensated divided-band signal.

[0045] As a preferred embodiment of the present invention, the signal detection module includes:

[0046] The detection processing unit is used to receive the processed signal output by the anti-interference compensation module, generate detection statistics for the divide-and-conquer frequency band signal corresponding to the divide-and-conquer frequency band identifier in the processed signal, and the generation of the detection statistics includes correlation matching operation, coherent accumulation operation or non-coherent accumulation operation; and determine the detection threshold according to the preset false alarm probability constraint, and output the detection result according to the comparison result of the detection statistics and the detection threshold.

[0047] The feedback index generation unit is used to generate a signal detection success rate index within a preset statistical window based on the detection results, using the divide-and-conquer frequency band identifier as an index. It also determines the signal-to-noise ratio after processing based on the estimated signal power and noise power of the processed signal within the corresponding divide-and-conquer frequency band, determines the signal-to-noise ratio before processing based on the useful signal strength information output by the frequency band interference fingerprint sensing module, and generates the signal-to-noise ratio change from the processed signal-to-noise ratio to the signal-to-noise ratio before processing.

[0048] The feedback scheduling unit is used to obtain the interference fingerprint stability coefficient and determine the statistical window length or feedback output period corresponding to the signal detection success rate index and the signal-to-noise ratio change for the next statistical period based on the interference fingerprint stability coefficient.

[0049] The feedback interface unit is used to output at least one of the following: signal detection success rate index, signal-to-noise ratio change, statistical window length of the next statistical period or feedback output period, and corresponding divide-and-conquer frequency band identifier to the mapping model parameter update unit.

[0050] On the other hand, the present invention also provides a method for detecting weak signals with interference fingerprint divide-and-conquer compensation, applied to the interference fingerprint divide-and-conquer compensation weak signal detection system as described above, the method comprising:

[0051] Step 1: Acquire the broadband electromagnetic signal to be tested, determine the frequency band coding information, interference fingerprint information and useful signal strength information, and determine the interference fingerprint stability coefficient based on the interference fingerprint information in multiple time windows; the frequency band coding information is the coding information used to indicate the frequency boundaries or frequency band index of multiple divide-and-conquer frequency bands;

[0052] Step 2: Divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on the frequency band coding information, and determine anti-interference strategy parameters for each divide-and-conquer frequency band based on interference fingerprint information and useful signal strength information; generate attenuation prediction values ​​based on the anti-interference strategy parameters and interference fingerprint information through attenuation characteristic relationships, whereby the attenuation characteristic relationships are the mapping relationships between the anti-interference strategy parameters, interference fingerprint information, and attenuation prediction values, and are represented by mapping model parameters; generate compensation parameters based on the attenuation prediction values, the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters, the useful signal strength information, and the preset target signal strength threshold; the compensation parameters are jointly determined by the attenuation prediction values, the pre-stored inherent attenuation amount, and the strength difference between the target signal strength threshold and the useful signal strength information.

[0053] Step 3: For each divided frequency band, implement anti-interference processing according to the anti-interference strategy parameters, and implement compensation processing according to the compensation parameters to obtain the processed signal of the divided frequency band.

[0054] Step 4: Perform signal detection based on the processed signal and output the detection result, and generate feedback data based on the detection result; recursively update the mapping model parameters according to the feedback data and the interference fingerprint stability coefficient.

[0055] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention can more effectively address interference effects in complex, dynamic, and non-stationary broadband electromagnetic interference environments across different frequency ranges, thereby improving adaptability to complex interference environments and alleviating the problems of insufficient interference suppression or excessive impact on useful signals caused by the uniform processing methods in existing technologies. This invention can determine the need for weak signal recovery or enhancement based on the signal interference situation, thus making subsequent processing intensity more aligned with actual needs, reducing the adverse effects of accumulated processing deviations on detection results, and improving the stability of weak signal detection. This invention can provide feedback correction and continuous adjustment to the preceding processing process based on the detection results, thereby maintaining good adaptive optimization capabilities when the interference environment changes, and improving the accuracy and robustness of weak signal detection under continuous operation conditions. Therefore, this invention can effectively improve the accuracy, stability, and environmental adaptability of weak signal detection in complex broadband electromagnetic interference environments. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0057] Figure 1 This is a schematic diagram of the modular structure of the system of the present invention;

[0058] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0060] like Figure 1 As shown, this embodiment provides a weak signal detection system with interference fingerprint divide-and-conquer compensation. The system mainly includes a frequency band interference fingerprint sensing module, a divide-and-conquer decision module, an anti-interference compensation module, and a signal detection module. The frequency band interference fingerprint sensing module acquires the broadband electromagnetic signal under test and outputs frequency band coding information, interference fingerprint information, useful signal strength information, and interference fingerprint stability coefficient. The divide-and-conquer decision module performs divide-and-conquer processing on the broadband electromagnetic signal under test based on the frequency band coding information and determines the anti-interference strategy parameters, attenuation prediction values, and compensation parameters corresponding to each divide-and-conquer frequency band. The anti-interference compensation module first performs anti-interference processing on each divide-and-conquer frequency band, then performs compensation processing, and outputs the processed signal. The signal detection module performs signal detection based on the processed signal and outputs the detection results, while simultaneously generating feedback data. The feedback data is output to the divide-and-conquer decision module to achieve closed-loop updating of the mapping model parameters.

[0061] (1) Frequency band interference fingerprint sensing module

[0062] The frequency band interference fingerprint sensing module is used to acquire the broadband electromagnetic signal under test, output frequency band coding information, interference fingerprint information and useful signal strength information, and determine the interference fingerprint stability coefficient based on the interference fingerprint information in multiple time windows; the frequency band coding information is used to indicate the frequency boundaries or frequency band indexes of multiple divide-and-conquer frequency bands.

[0063] In this embodiment, the frequency band interference fingerprint sensing module consists of a time-frequency representation selection unit, a fingerprint generation unit, a stability evaluation unit, a suppression control parameter update unit, a frequency domain suppression weight generation unit, an interference suppression unit, and an intensity estimation unit. Its basic data flow is as follows: the broadband electromagnetic signal under test sequentially undergoes time-frequency representation selection, interference fingerprint extraction, interference stability evaluation, suppression weight update, frequency domain suppression processing, and intensity estimation processing, outputting frequency band coding information, interference fingerprint information, useful signal strength information, and interference fingerprint stability coefficient.

[0064] The time-frequency representation selection unit is used to generate two candidate time-frequency representations with different window lengths for the signal components of the broadband electromagnetic signal under test within the same candidate analysis frequency band, and to construct a dynamic index based on the power change over time and the spectrum occupancy change over time obtained statistically from the candidate time-frequency representations; when the dynamic index is higher than a preset threshold, the short-window candidate time-frequency representation is selected as the target time-frequency representation; when the dynamic index is lower than or equal to the preset threshold, the long-window candidate time-frequency representation is selected as the target time-frequency representation.

[0065] Specifically, the broadband electromagnetic signal under test is first divided into several candidate analysis frequency bands. For each candidate analysis frequency band, two candidate time-frequency representations are generated: a short-window representation and a long-window representation. The short-window candidate time-frequency representation has higher time resolution and is suitable for characterizing rapidly changing interference; the long-window candidate time-frequency representation has higher frequency resolution and is suitable for characterizing relatively stable or narrowband interference. Subsequently, a dynamic index is constructed based on the degree of power variation and spectrum occupancy variation within the candidate analysis frequency band, and a target time-frequency representation that is more suitable for the current interference state is automatically selected accordingly.

[0066] The fingerprint generation unit is used to generate a binary interference fingerprint code based on the target time-frequency representation. The binary interference fingerprint code serves as the interference fingerprint information and includes a center frequency encoding field and a bandwidth encoding field, which are used to decode the interference center frequency and interference bandwidth.

[0067] Specifically, the fingerprint generation unit extracts high-energy regions from the target time-frequency representation to identify the main interference regions within the current candidate analysis frequency band. Subsequently, it calculates the interference center frequency and interference bandwidth corresponding to each interference region, and encodes them respectively into a center frequency encoded field and a bandwidth encoded field, thereby forming a binary interference fingerprint code. Simultaneously, frequency band encoding information can be generated based on the frequency boundaries of the interference regions, which is then used by the subsequent divide-and-conquer decision module to perform divide-and-conquer processing on the broadband electromagnetic signal under test.

[0068] The stability evaluation unit is used to evaluate the stability of the same divide-and-conquer frequency band in continuous frequency ranges. The binary interference fingerprint sequence obtained within a time window Calculate the stability coefficient of the interference fingerprint The following equation is satisfied:

[0069] ;

[0070] In the formula, To interfere with the fingerprint stability coefficient; The number of stability assessment time windows greater than 1; The time window number, and ; For the first The binary interference fingerprint code obtained from the time window; For the first The binary interference fingerprint code obtained from the time window; Let Hamming distance function be used.

[0071] In this embodiment, the stability evaluation unit associates binary interference fingerprint codes belonging to the same divided frequency band within multiple consecutive time windows based on the divided frequency band identifier or frequency proximity relationship, forming corresponding code sequences. The Hamming distance between interference fingerprint codes in adjacent time windows is calculated to measure the degree of temporal change in the interference state within the current divided frequency band. A smaller Hamming distance indicates a more stable interference state and a larger corresponding interference fingerprint stability coefficient; conversely, a larger distance indicates faster interference changes and a smaller corresponding interference fingerprint stability coefficient.

[0072] The suppression control parameter update unit is used to update the interference fingerprint stability coefficient. Determine the suppression control parameters and adaptively update the gain. The following equation is satisfied:

[0073] ;

[0074] In the formula, To suppress adaptive gain updates of control parameters; To suppress adaptive updates of the lower limit of the gain by controlling parameters; To suppress adaptive updates of the gain upper limit by controlling parameters;

[0075] The frequency domain suppression weight generation unit is used to decode the interference center frequency and interference bandwidth based on the center frequency encoding field and the bandwidth encoding field, and construct an instantaneous frequency domain suppression weight matrix. and based on Frequency domain suppression weight matrix at the previous update time and We perform weighted fusion to obtain the frequency domain suppression weight matrix at the current update time. The following equation is satisfied:

[0076] ;

[0077] In the formula, To update the time sequence number; the frequency domain suppression weight matrix of the current update time is... The frequency domain suppression weight parameter, which is part of the anti-interference strategy parameters, is output to the anti-interference compensation module.

[0078] In this embodiment, the frequency domain suppression weight generation unit recovers the current interference center frequency and interference bandwidth based on the binary interference fingerprint code, and determines the suppression domain accordingly, constructing an instantaneous frequency domain suppression weight matrix. Subsequently, the instantaneous frequency domain suppression weight matrix is ​​fused with the frequency domain suppression weight matrix of the previous update time to obtain the frequency domain suppression weight matrix of the current update time, so as to take into account both historical stability and current dynamic response capability.

[0079] The interference suppression unit is used to perform frequency domain weighted suppression processing on the signal components of the broadband electromagnetic signal under test in the corresponding candidate analysis frequency band based on the frequency domain suppression weight matrix at the current update time, and generate residual signals.

[0080] Specifically, the interference suppression unit transforms the broadband electromagnetic signal under test to the frequency domain, and then uses the frequency domain suppression weight matrix at the current update time to perform weighted suppression on the frequency domain components within the corresponding candidate analysis frequency band. After inverse transformation, a residual signal is generated. The main interference components in the residual signal have been weakened, which is beneficial for subsequent estimation of the useful signal strength.

[0081] The intensity estimation unit is used to form a power sample sequence based on the residual signal, and to perform statistical estimation on the power sample sequence using a preset robust statistical estimation method to obtain power estimates and generate useful signal intensity information.

[0082] Specifically, the intensity estimation unit calculates the power of the residual signal within a statistical window, forming a corresponding power sample sequence. It then uses a robust statistical estimation method to obtain power estimates, and generates useful signal strength information based on these estimates. The robust statistical estimation method can employ median estimation, truncated mean estimation, or other conventional robust statistical methods to reduce the impact of outliers on the estimation results. Finally, it outputs the useful signal strength information corresponding to the current divide-and-conquer frequency band for use by the divide-and-conquer decision module.

[0083] In summary, the working process of the frequency band interference fingerprint sensing module is as follows: First, candidate time-frequency representations are generated and the target time-frequency representation is selected; then, the interference region is extracted, and a binary interference fingerprint code and frequency band coding information are formed; subsequently, the interference fingerprint stability coefficient is calculated based on the interference fingerprint code of continuous time windows, and the frequency domain suppression weight is updated accordingly; finally, frequency domain suppression is performed and the useful signal strength is estimated. This completes the joint generation of frequency band coding information, interference fingerprint information, useful signal strength information, and interference fingerprint stability coefficient.

[0084] (2) Divide and conquer decision-making module

[0085] The divide-and-conquer decision module is used to divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on the frequency band coding information, and for each divide-and-conquer frequency band: determine anti-interference strategy parameters based on interference fingerprint information and useful signal strength information; generate attenuation prediction values ​​based on the anti-interference strategy parameters and interference fingerprint information through attenuation characteristic relationships, wherein the attenuation characteristic relationships are the mapping relationship between the anti-interference strategy parameters and interference fingerprint information to the attenuation prediction values, and are represented by mapping model parameters; generate compensation parameters based on the attenuation prediction values, the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters, useful signal strength information, and preset target signal strength threshold; and recursively update the mapping model parameters based on feedback data and interference fingerprint stability coefficients.

[0086] In this embodiment, the divide-and-conquer decision-making module consists of a divide-and-conquer frequency band generation unit, a strategy parameter determination unit, an attenuation prediction unit, a compensation parameter generation unit, an attenuation prediction model library unit, a stability acquisition unit, a model selection unit, and a mapping model parameter update unit. Its basic data flow is as follows: frequency band coding information first forms each divide-and-conquer frequency band and its identifier; then, interference fingerprint information and useful signal strength information are combined to form anti-interference strategy parameters; next, an applicable mapping model is selected from the attenuation prediction model library based on the interference fingerprint stability coefficient, and the attenuation prediction value is output; then, compensation parameters are generated based on the pre-stored inherent attenuation amount, attenuation prediction value, target signal strength threshold, and useful signal strength information; finally, the current mapping model parameters are recursively updated based on feedback data and the interference fingerprint stability coefficient.

[0087] The divide-and-conquer frequency band generation unit is used to divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on the frequency band coding information and generate divide-and-conquer frequency band identifiers.

[0088] In this embodiment, the divide-and-conquer frequency band generation unit receives the frequency band coding information output by the frequency band interference fingerprint sensing module, and divides the broadband electromagnetic signal under test according to the frequency boundary or frequency band index to form multiple continuous and non-overlapping divide-and-conquer frequency bands. Subsequently, a unique divide-and-conquer frequency band identifier is assigned to each divide-and-conquer frequency band in frequency order for use in subsequent strategy parameter determination, attenuation prediction, and compensation parameter generation.

[0089] The strategy parameter determination unit is used to determine anti-interference strategy parameters for the divide-and-conquer frequency band corresponding to the divide-and-conquer frequency band identifier, based on interference fingerprint information and useful signal strength information.

[0090] In this embodiment, the anti-interference strategy parameters include at least the frequency domain suppression weight parameters, strategy mode identifier, suppression depth adjustment parameters, and compensation gain limiting parameters corresponding to the current divided frequency band. The strategy parameter determination unit determines the applicable processing mode and parameter configuration for the current divided frequency band based on the interference center frequency, interference bandwidth, and useful signal strength information within the current divided frequency band. When the interference within the current divided frequency band is narrow and the useful signal strength is high, a stronger suppression mode can be adopted; when the interference is wide or the useful signal strength is low, a suppression mode can be adopted to reduce the additional impact on the useful signal.

[0091] The attenuation prediction unit is used to input anti-interference strategy parameters and interference fingerprint information into the attenuation characteristic relationship for the divided frequency band corresponding to the divided frequency band identifier, and output attenuation prediction value.

[0092] The attenuation prediction model library unit is used to store multiple mapping models for representing the relationship of attenuation characteristics.

[0093] The stability acquisition unit is used to acquire the interference fingerprint stability coefficient output by the frequency band interference fingerprint sensing module.

[0094] The model selection unit is used to select a mapping model for the divide-and-conquer frequency band corresponding to the divide-and-conquer frequency band identifier based on the interference fingerprint stability coefficient, and call the selected mapping model with the divide-and-conquer frequency band identifier as an index to determine the attenuation characteristic relationship of the corresponding divide-and-conquer frequency band and output the attenuation prediction value.

[0095] In this embodiment, the attenuation characteristic relationship is implemented using a parameterized mapping model. This mapping model takes the interference state information, anti-interference strategy parameters, and useful signal strength information corresponding to the current divide-and-conquer frequency band as input, and outputs the corresponding attenuation prediction value. Before deployment, multiple mapping models are constructed using historical sample data and stored in the attenuation prediction model library according to the interference fingerprint stability interval. During online operation, the model selection unit selects an applicable mapping model from the model library based on the interference fingerprint stability coefficient corresponding to the current divide-and-conquer frequency band, and then the attenuation prediction unit outputs the attenuation prediction value for the current divide-and-conquer frequency band.

[0096] The compensation parameter generation unit is used to generate compensation parameters based on the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters for the divided frequency bands identified by the divided frequency band identifier. Attenuation prediction value Target signal strength threshold and useful signal strength information Calculate the compensation parameters to satisfy the following formula:

[0097] ;

[0098] In the formula, For dividing and conquering frequency band identification; For the corresponding compensation parameters; This is for calculating the maximum value.

[0099] In this embodiment, the pre-stored inherent attenuation represents the basic attenuation introduced by the current anti-interference strategy itself, and the attenuation prediction value represents the additional attenuation predicted based on the current interference state. The formula... This indicates the additional compensation required when the current useful signal strength is lower than the target signal strength threshold. Therefore, the compensation parameter generation unit can comprehensively consider the impact of the current strategy, environmental factors, and target strength requirements to provide corresponding divide-and-conquer frequency band compensation parameters for subsequent anti-interference compensation modules.

[0100] The mapping model parameter update unit is used to: recursively update the mapping model parameters of the attenuation characteristic relationship based on feedback data; determine the measured intensity change based on the difference between the useful signal strength information before and after anti-interference processing, and determine the prediction error based on the difference between the measured intensity change and the attenuation prediction value; determine the recursive update gain lower limit and recursive update gain upper limit based on the signal-to-noise ratio change and the signal detection success rate index according to the preset gain constraint mapping relationship; and determine the recursive update gain based on the interference fingerprint stability coefficient.

[0101] In this embodiment, the mapping model parameter update unit first determines the measured intensity change based on the difference between the useful signal strength information before and after the anti-interference processing. The following equation is satisfied: ;

[0102] The prediction error is determined by the difference between the measured intensity change and the predicted attenuation value, satisfying the following formula:

[0103] ;

[0104] In the formula, For the first Useful signal strength information before anti-interference processing of each frequency band; For the first Useful signal strength information after anti-interference processing of each frequency band; This corresponds to the prediction error.

[0105] In this embodiment, the lower and upper limits of the recursive update gain are determined based on the signal-to-noise ratio change and signal detection success rate in the feedback data, through a preset gain constraint mapping relationship, and are limited to a preset range. Subsequently, the recursive update gain is determined based on the interference fingerprint stability coefficient, satisfying the following formula:

[0106] ;

[0107] In the formula, To interfere with the fingerprint stability coefficient; For the first Each divide-and-conquer frequency band identifier corresponds to a recursive update gain; To recursively update the lower bound of the gain; To recursively update the upper limit of gain;

[0108] The signal detection success rate index is an index obtained by statistics within a preset statistical window. The useful signal strength information before anti-interference processing is output by the frequency band interference fingerprint sensing module, and the useful signal strength information after anti-interference processing is output by the anti-interference compensation module.

[0109] Therefore, when the interference state changes rapidly within the current divide-and-conquer frequency band, the recursive update gain is closer to the upper limit to enhance the model's update responsiveness; when the interference state is relatively stable, the recursive update gain is closer to the lower limit to maintain smooth parameter updates. Subsequently, an online recursive method using the recursive update gain as the step size and the prediction error as the basis can be adopted to update the parameters of the currently used mapping model.

[0110] In summary, the divide-and-conquer decision-making module works as follows: First, multiple divide-and-conquer frequency bands are formed based on the frequency band coding information, and corresponding divide-and-conquer frequency band identifiers are generated; then, the anti-interference strategy parameters corresponding to each divide-and-conquer frequency band are determined by combining the interference fingerprint information and the useful signal strength information; subsequently, an applicable mapping model is selected based on the interference fingerprint stability coefficient, and the attenuation prediction value is output; then, compensation parameters are generated based on the pre-stored inherent attenuation amount, the attenuation prediction value, the target signal strength threshold, and the useful signal strength information; finally, the mapping model parameters are recursively updated using feedback data, prediction error, and interference fingerprint stability coefficient, thereby realizing a complete closed-loop decision-making process.

[0111] (3) Disturbance compensation module

[0112] The anti-interference compensation module is used to perform anti-interference processing for each divided frequency band according to the anti-interference strategy parameters, and to perform compensation processing on the divided frequency band signal after anti-interference processing according to the compensation parameters, and output the processed signal.

[0113] In this embodiment, the anti-interference compensation module consists of a frequency band processing execution unit, a compensation control unit, a compensation weight generation unit, and a compensation execution unit. Its basic data flow is as follows: the frequency band processing execution unit, in conjunction with the frequency domain suppression weight parameters in the anti-interference strategy parameters, performs anti-interference processing on the current divided-band signal and outputs the processed divided-band signal; the compensation control unit determines the compensation weight update gain and compensation gain limiting constraint based on the interference fingerprint stability coefficient; the compensation weight generation unit determines the compensation scope based on the frequency domain suppression weight matrix at the current update time and generates a compensation weight matrix in conjunction with the compensation parameters; the compensation execution unit applies the compensation weight matrix to the processed divided-band signal and outputs the compensated divided-band signal; multiple compensated divided-band signals are recombined to form the processed signal.

[0114] The frequency band processing execution unit is used to perform weighted suppression on the frequency domain components of the divided-and-conquer frequency band signal within the interference frequency range, based on the frequency domain suppression weight matrix which serves as an anti-interference strategy parameter, and output the divided-and-conquer frequency band signal after anti-interference processing.

[0115] In this embodiment, the frequency band processing execution unit transforms the current divide-and-conquer frequency band signal to the frequency domain, then uses the frequency domain suppression weight matrix corresponding to the current divide-and-conquer frequency band to perform weighted suppression on the frequency domain components within the interference frequency range, and then performs an inverse transformation to restore it to the time domain, thus obtaining the anti-interference processed divide-and-conquer frequency band signal. This processing can weaken the main interference components while preserving as many useful signal components as possible.

[0116] The compensation control unit is used to obtain the interference fingerprint stability coefficient output by the stability evaluation unit, and determine the compensation weight update gain and compensation gain limit constraint based on the interference fingerprint stability coefficient according to the preset compensation control mapping relationship.

[0117] In this embodiment, the compensation control unit adjusts the update speed and gain range of the compensation processing based on the interference fingerprint stability coefficient corresponding to the current divide-and-conquer frequency band. When the interference state is relatively stable, a smoother compensation update method can be adopted; when the interference state changes rapidly, a more conservative gain limiting constraint can be adopted to reduce the risk of noise amplification and distortion caused by over-compensation.

[0118] The compensation weight generation unit is used to determine the compensation scope based on the frequency domain suppression weight matrix at the current update time. The compensation scope is consistent with the suppression scope corresponding to the frequency domain suppression weight matrix at the current update time. The compensation weight matrix is ​​generated based on the compensation parameters, the compensation weight update gain, and the compensation gain limiting constraint.

[0119] In this embodiment, the compensation weight generation unit first determines the suppression domain based on the frequency domain suppression weight matrix at the current update time, and uses it as the compensation domain, so that the compensation mainly applies to the frequency region where useful signals may be attenuated after anti-interference processing. Subsequently, it generates a target compensation distribution by combining the compensation parameters corresponding to the current divide-and-conquer frequency band, and merges the compensation weight matrix of the previous time and the current target compensation distribution according to the compensation weight update gain. Then, it limits the compensation weights of each frequency point according to the compensation gain limiting constraint to obtain the compensation weight matrix at the current time.

[0120] The compensation execution unit is used to apply a compensation weight matrix to the divided-band signal after anti-interference processing and output the compensated divided-band signal.

[0121] In this embodiment, the compensation execution unit transforms the anti-interference processed divide-and-conquer frequency band signal to the frequency domain, and then performs weighted enhancement on the frequency domain components within the compensation domain according to the compensation weight matrix. Frequency domain components outside the compensation domain retain their original values ​​or approximately retain their original values. The compensated divide-and-conquer frequency band signal is then obtained through inverse transformation. The compensated divide-and-conquer frequency band signals corresponding to multiple divide-and-conquer frequency bands can be reassembled in the original frequency position order to form a processed signal, which is then output to the signal detection module.

[0122] In summary, the anti-interference compensation module works as follows: First, frequency domain weighted suppression is performed on the signals of each divided frequency band according to the anti-interference strategy parameters to obtain the divided frequency band signals after anti-interference processing; then, the compensation control quantity is determined according to the interference fingerprint stability coefficient; subsequently, a compensation weight matrix is ​​generated according to the frequency domain suppression weight matrix, compensation parameters, and compensation control quantity at the current update time; finally, the compensation weight matrix is ​​applied to the divided frequency band signals after anti-interference processing, the compensated divided frequency band signals are output, and the processed signals are used for subsequent detection.

[0123] (4) Signal detection module

[0124] The signal detection module is used to perform signal detection based on the processed signal and output the detection result, and generate feedback data based on the detection result. The feedback data is output to the mapping model parameter update unit in the divide-and-conquer decision module.

[0125] In this embodiment, the signal detection module consists of a detection processing unit, a feedback index generation unit, a feedback scheduling unit, and a feedback interface unit. Its basic data flow is as follows: the detection processing unit generates detection statistics for the divided frequency band signals corresponding to each divided frequency band identifier in the processed signal and outputs the detection results; the feedback index generation unit generates a signal detection success rate index and a signal-to-noise ratio change based on the detection results; the feedback scheduling unit determines the statistical window length or feedback output period corresponding to the next statistical period based on the interference fingerprint stability coefficient; and the feedback interface unit outputs at least one of the signal detection success rate index, signal-to-noise ratio change, the statistical window length of the next statistical period, or the feedback output period, along with the corresponding divided frequency band identifier, to the mapping model parameter update unit.

[0126] The detection processing unit is used to receive the processed signal output by the anti-interference compensation module, generate detection statistics for the divide-and-conquer frequency band signal corresponding to the divide-and-conquer frequency band identifier in the processed signal, and the generation of the detection statistics includes correlation matching operation, coherent accumulation operation or non-coherent accumulation operation; and determine the detection threshold according to the preset false alarm probability constraint, and output the detection result according to the comparison result of the detection statistics and the detection threshold.

[0127] In this embodiment, the detection processing unit first extracts the corresponding divide-and-conquer frequency band signal based on the divide-and-conquer frequency band identifier, and then selects the detection statistic generation method according to the target signal structure. When the target template, synchronization sequence, or pilot structure is known, the correlation matching method can be preferred; when the target signal has a stable repeating structure in multiple time slots or periods, the coherent accumulation method can be further used to enhance the weak signal response; when the phase consistency is poor, the incoherent accumulation method can be used to improve the detection stability. The detection threshold is determined based on a preset false alarm probability constraint, and the detection result is output based on the comparison result between the detection statistic and the threshold.

[0128] The feedback index generation unit is used to generate a signal detection success rate index within a preset statistical window based on the detection results, using the divide-and-conquer frequency band identifier as an index. It also determines the signal-to-noise ratio (SNR) after processing based on the estimated signal power and noise power of the processed signal within the corresponding divide-and-conquer frequency band, determines the SNR before processing based on the useful signal strength information output by the frequency band interference fingerprint sensing module, and generates the SNR change from the SNR after processing to the SNR before processing.

[0129] In this embodiment, the feedback index generation unit statistically analyzes the detection results for each divided frequency band and uses the ratio of the number of successful detections within a preset statistical window to the total number of detections as the signal detection success rate index for the current divided frequency band. Simultaneously, the feedback index generation unit obtains the processed signal-to-noise ratio (SNR) based on the estimated signal power and noise power of the processed signal within the current divided frequency band. It then combines this with the useful signal strength information output by the frequency band interference fingerprint sensing module to determine the unprocessed SNR, and the difference between the two yields the SNR change. This SNR change reflects the degree to which the interference suppression compensation processing improves the signal quality of the current divided frequency band.

[0130] The feedback scheduling unit is used to obtain the interference fingerprint stability coefficient, and to determine the statistical window length or feedback output period corresponding to the signal detection success rate index and the signal-to-noise ratio change for the next statistical period based on the interference fingerprint stability coefficient.

[0131] In this embodiment, the feedback scheduling unit adjusts the feedback rhythm based on the interference fingerprint stability coefficient corresponding to the current divide-and-conquer frequency band. When the interference fingerprint stability coefficient is large, it indicates that the current interference state is relatively stable, and a longer statistical window length or a longer feedback output cycle can be used. When the interference fingerprint stability coefficient is small, it indicates that the current interference changes rapidly, and a shorter statistical window length or a shorter feedback output cycle can be used to improve the timeliness of feedback.

[0132] The feedback interface unit is used to output the signal detection success rate index, signal-to-noise ratio change, statistical window length or feedback output period of the next statistical period, and corresponding divide-and-conquer frequency band identifier to the mapping model parameter update unit.

[0133] In this embodiment, the feedback interface unit organizes and encapsulates the feedback indicators and feedback scheduling results corresponding to each divide-and-conquer frequency band, forming feedback data indexed by the divide-and-conquer frequency band identifier, and sends it to the mapping model parameter update unit in the divide-and-conquer decision module. If necessary, the feedback data may also include auxiliary information such as the current frame number, feedback timestamp, or probe confidence level to enhance the cross-module data correspondence.

[0134] In summary, the signal detection module works as follows: First, it extracts the corresponding divide-and-conquer frequency band signal based on the divide-and-conquer frequency band identifier, generates detection statistics, completes threshold decision, and outputs the detection results; then, it generates a signal detection success rate index based on the detection results, and obtains the change in signal-to-noise ratio by combining the signal-to-noise ratio before and after processing; subsequently, it determines the statistical window length or feedback output period for the next statistical period based on the interference fingerprint stability coefficient; finally, it outputs the feedback data corresponding to each divide-and-conquer frequency band to the mapping model parameter update unit in the divide-and-conquer decision module to support the closed-loop update of the mapping model parameters.

[0135] The candidate analysis frequency band division method, interference region extraction method, mapping model form, statistical window length, threshold determination method, compensation control mapping relationship, and feedback scheduling method involved in the above-described module implementation methods are all preferred examples in this embodiment. Those skilled in the art can make conventional substitutions or parameter adjustments based on specific signal types, receiving bandwidth, real-time requirements, and hardware resources without departing from the technical concept of this invention.

[0136] like Figure 2 As shown, another embodiment of the present invention provides a weak signal detection method with interference fingerprint divide-and-conquer compensation, applied to the aforementioned weak signal detection system with interference fingerprint divide-and-conquer compensation. The method can be implemented by a processor executing program instructions stored in a memory, or by a digital signal processor, a field-programmable gate array, an application-specific integrated circuit, or a hardware-software collaborative approach. The method corresponds to the aforementioned frequency band interference fingerprint sensing module, divide-and-conquer decision module, interference suppression compensation module, and signal detection module, and its execution process can be completed collaboratively by these modules.

[0137] The method includes the following steps:

[0138] Step 1: Obtain the broadband electromagnetic signal to be tested, determine the frequency band coding information, interference fingerprint information and useful signal strength information, and determine the interference fingerprint stability coefficient based on the interference fingerprint information in multiple time windows; the frequency band coding information is the coding information used to indicate the frequency boundaries or frequency band indexes of multiple divide-and-conquer frequency bands.

[0139] In this embodiment, step 1 can be performed by the frequency band interference fingerprint sensing module. Specifically, firstly, time-frequency analysis is performed on the broadband electromagnetic signal under test to extract interference features within the candidate analysis frequency band, forming interference fingerprint information and frequency band coding information; then, the stability of the interference fingerprint information in multiple consecutive time windows of the same divided frequency band is evaluated to obtain the interference fingerprint stability coefficient; simultaneously, the intensity of the broadband electromagnetic signal under test related to the useful signal is estimated to obtain the useful signal strength information.

[0140] Step 2: Divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on the frequency band coding information, and determine anti-interference strategy parameters for each divide-and-conquer frequency band based on interference fingerprint information and useful signal strength information; generate attenuation prediction values ​​based on the anti-interference strategy parameters and interference fingerprint information through attenuation characteristic relationships, whereby the attenuation characteristic relationships are the mapping relationships between the anti-interference strategy parameters, interference fingerprint information, and attenuation prediction values, and are represented by mapping model parameters; generate compensation parameters based on the attenuation prediction values, the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters, the useful signal strength information, and the preset target signal strength threshold; the compensation parameters are jointly determined by the attenuation prediction values, the pre-stored inherent attenuation amount, and the strength difference between the target signal strength threshold and the useful signal strength information.

[0141] In this embodiment, step 2 can be executed by the divide-and-conquer decision module. Specifically, firstly, the broadband electromagnetic signal under test is divided into multiple divide-and-conquer frequency bands according to the frequency band coding information, and a corresponding divide-and-conquer frequency band identifier is generated for each divide-and-conquer frequency band; then, the anti-interference strategy parameters for the current divide-and-conquer frequency band are determined by combining the interference fingerprint information and useful signal strength information corresponding to each divide-and-conquer frequency band; next, based on the anti-interference strategy parameters, interference fingerprint information, and interference fingerprint stability coefficient corresponding to the current divide-and-conquer frequency band, an applicable mapping model is selected from the pre-stored mapping models, and the attenuation prediction value corresponding to the current divide-and-conquer frequency band is output; finally, the compensation parameters corresponding to the current divide-and-conquer frequency band are generated by combining the pre-stored inherent attenuation, attenuation prediction value, target signal strength threshold, and useful signal strength information.

[0142] Step 3: For each divided frequency band, implement anti-interference processing according to the anti-interference strategy parameters, and implement compensation processing according to the compensation parameters to obtain the processed signal of the divided frequency band.

[0143] In this embodiment, step 3 can be performed by the anti-interference compensation module. Specifically, firstly, based on the anti-interference strategy parameters corresponding to each divided frequency band, anti-interference processing is performed on the signals of each divided frequency band to weaken the interference components within the corresponding interference frequency range, resulting in the divided frequency band signals after anti-interference processing; then, based on the compensation parameters corresponding to each divided frequency band, compensation processing is performed on the divided frequency band signals after anti-interference processing to restore the useful signal components that have been attenuated due to the anti-interference processing; after each divided frequency band is processed, it is reassembled in the original frequency position order to obtain the processed signal.

[0144] Step 4: Perform signal detection based on the processed signal and output the detection result, and generate feedback data based on the detection result; recursively update the mapping model parameters according to the feedback data and the interference fingerprint stability coefficient.

[0145] In this embodiment, step 4 can be executed collaboratively by the mapping model parameter update unit in the signal detection module and the divide-and-conquer decision module. Specifically, signal detection is first performed based on the processed signal, and the detection results corresponding to each divide-and-conquer frequency band are output; then, feedback data is generated based on the detection results. The feedback data includes at least the signal detection success rate index, signal-to-noise ratio change, and the statistical window length or feedback output period corresponding to the next statistical period obtained by statistical analysis of the divide-and-conquer frequency bands; finally, the feedback data is output to the mapping model parameter update unit, and the mapping model parameters are recursively updated in conjunction with the interference fingerprint stability coefficient corresponding to the current divide-and-conquer frequency band to improve the accuracy of subsequent attenuation prediction and compensation parameter generation.

[0146] In summary, the interference fingerprint divide-and-conquer compensation weak signal detection method provided in this embodiment first acquires the broadband electromagnetic signal to be tested and extracts frequency band coding information, interference fingerprint information, useful signal strength information, and interference fingerprint stability coefficient. Then, it performs divide-and-conquer processing on the broadband electromagnetic signal to be tested, and determines anti-interference strategy parameters, attenuation prediction values, and compensation parameters by combining the interference state and useful signal strength information. Subsequently, anti-interference processing and compensation processing are sequentially implemented for each divided frequency band. Based on the processed signal, signal detection, feedback data generation, and recursive updating of mapping model parameters are performed, thus forming a complete closed-loop optimization process. This method can improve the accuracy, stability, and environmental adaptability of weak signal detection in complex, dynamic, and non-stationary broadband electromagnetic interference environments.

[0147] The candidate analysis frequency band division method, interference fingerprint extraction method, anti-interference strategy parameter determination method, mapping model form, compensation parameter generation method, detection statistics generation method, threshold determination method, and feedback scheduling method involved in the above-described method implementation are all preferred examples in this embodiment. Those skilled in the art can make conventional substitutions or parameter adjustments based on specific signal types, receiving bandwidth, real-time requirements, and hardware resources without departing from the technical concept of this invention.

[0148] In summary, this invention achieves closed-loop processing for weak signal detection in complex broadband electromagnetic interference environments, and is well-suited for weak signal detection scenarios in complex electromagnetic environments. It should be noted that the above embodiments are merely preferred embodiments of this invention and are not intended to limit the scope of protection of this invention; any modifications, equivalent substitutions, or improvements made within the spirit and principles of this invention should fall within the scope of protection of this invention.

Claims

1. A weak signal detection system with interference fingerprint divide-and-conquer compensation, characterized in that, The system includes: The frequency band interference fingerprint sensing module is used to acquire the broadband electromagnetic signal under test, output frequency band coding information, interference fingerprint information and useful signal strength information, and determine the interference fingerprint stability coefficient based on the interference fingerprint information in multiple time windows; the frequency band coding information is coding information used to indicate the frequency boundaries or frequency band indexes of multiple divide-and-conquer frequency bands; The divide-and-conquer decision module is used to divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on frequency band coding information, and for each divide-and-conquer frequency band: determine anti-interference strategy parameters based on interference fingerprint information and useful signal strength information; generate attenuation prediction values ​​based on the anti-interference strategy parameters and interference fingerprint information through attenuation characteristic relationships, wherein the attenuation characteristic relationships are mapping relationships between anti-interference strategy parameters, interference fingerprint information, and attenuation prediction values, and are represented by mapping model parameters; generate compensation parameters based on the attenuation prediction values, the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters, useful signal strength information, and a preset target signal strength threshold, wherein the compensation parameters are jointly determined by the attenuation prediction values, the pre-stored inherent attenuation amount, and the strength difference between the target signal strength threshold and the useful signal strength information; and recursively update the mapping model parameters based on feedback data and interference fingerprint stability coefficients. The anti-interference compensation module is used to perform anti-interference processing on the divided frequency band signal of each divided frequency band according to the anti-interference strategy parameters, and to perform compensation processing on the divided frequency band signal after anti-interference processing according to the compensation parameters, output the compensated divided frequency band signal, and reassemble multiple compensated divided frequency band signals into the processed signal. The signal detection module is used to perform signal detection based on the processed signal and output the detection result, and generate feedback data based on the detection result, and output the feedback data to the divide-and-conquer decision module.

2. The interference fingerprint divide-and-conquer compensating weak signal detection system according to claim 1, wherein, The frequency band interference fingerprint sensing module includes: The time-frequency representation selection unit is used to generate two candidate time-frequency representations with different window lengths for signal components corresponding to the same candidate analysis frequency band in the broadband electromagnetic signal under test, and to construct a dynamic index based on the power change over time and the spectrum occupancy change over time obtained statistically from the candidate time-frequency representations; when the dynamic index is higher than a preset threshold, the short-window candidate time-frequency representation is selected as the target time-frequency representation; when the dynamic index is lower than or equal to the preset threshold, the long-window candidate time-frequency representation is selected as the target time-frequency representation. The fingerprint generation unit is used to generate a binary interference fingerprint code based on the target time-frequency representation. The binary interference fingerprint code serves as interference fingerprint information and includes a center frequency encoding field and a bandwidth encoding field, which are used to decode and obtain the interference center frequency and interference bandwidth.

3. The weak signal detection system for interference fingerprint divide-and-conquer compensation according to claim 2, characterized in that, The frequency band interference fingerprint sensing module also includes: Stability evaluation unit, used to evaluate the stability of the same divide-and-conquer frequency band in continuous The binary interference fingerprint obtained within a time window Calculate the stability coefficient of the interference fingerprint The following equation is satisfied: ； In the formula, To interfere with the fingerprint stability coefficient; The number of stability assessment time windows greater than 1; The time window number, and ; For the first The binary interference fingerprint code obtained from the time window; For the first The binary interference fingerprint code obtained from the time window; Let Hamming distance function be used. Suppression control parameter update unit, used to update the stability coefficient of the interference fingerprint Determine the suppression control parameters and adaptively update the gain. The following equation is satisfied: ； In the formula, To suppress adaptive gain updates of control parameters; To suppress adaptive updates of the lower limit of the gain by controlling parameters; To suppress adaptive updates of the gain upper limit by controlling parameters; The frequency domain suppression weight generation unit is used to decode the interference center frequency and interference bandwidth based on the center frequency encoding field and bandwidth encoding field, and construct the instantaneous frequency domain suppression weight matrix. and based on Frequency domain suppression weight matrix at the previous update time and We perform weighted fusion to obtain the frequency domain suppression weight matrix at the current update time. ,satisfy: ； In the formula, To update the time sequence number; The frequency domain suppression weight matrix at the current update time The frequency domain suppression weight parameter, which is part of the anti-interference strategy parameters, is output to the anti-interference compensation module.

4. The weak signal detection system for interference fingerprint divide-and-conquer compensation according to claim 3, characterized in that, The frequency band interference fingerprint sensing module also includes: Interference suppression unit, used to suppress weight matrix in the frequency domain based on the current update time. Frequency domain weighted suppression processing is performed on the signal components of the corresponding candidate analysis frequency band in the broadband electromagnetic signal under test to generate residual signals; The intensity estimation unit is used to form a power sample sequence based on the residual signal, and to perform statistical estimation on the power sample sequence using a preset robust statistical estimation method to obtain the power estimate and generate useful signal intensity information.

5. The weak signal detection system for interference fingerprint divide-and-conquer compensation according to claim 1, characterized in that, The divide-and-conquer decision-making module includes: The divide-and-conquer frequency band generation unit is used to divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands and generate divide-and-conquer frequency band identifiers based on the frequency band coding information. The strategy parameter determination unit is used to determine the anti-interference strategy parameters for the divided frequency bands corresponding to the divided frequency band identifiers, based on the interference fingerprint information and the useful signal strength information. The attenuation prediction unit is used to generate attenuation prediction values ​​for the divided frequency bands corresponding to the divided frequency band identifier, by using the anti-interference strategy parameters and interference fingerprint information through the attenuation characteristic relationship. ; The compensation parameter generation unit is used to generate compensation parameters based on the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters for the divided frequency bands identified by the divided frequency band identifier. Attenuation prediction value Target signal strength threshold and useful signal strength information Calculate the compensation parameters to satisfy the following formula: ； In the formula: For dividing and conquering frequency band identification; For compensation parameters; This is a function for calculating the maximum value.

6. The weak signal detection system for interference fingerprint divide-and-conquer compensation according to claim 5, characterized in that, The divide-and-conquer decision-making module also includes: The attenuation prediction model library unit is used to store multiple mapping models for representing the relationship of attenuation characteristics; The stability acquisition unit is used to acquire the interference fingerprint stability coefficient output by the frequency band interference fingerprint sensing module; The model selection unit is used to select a mapping model for the divide-and-conquer frequency band corresponding to the divide-and-conquer frequency band identifier based on the interference fingerprint stability coefficient, and call the selected mapping model with the divide-and-conquer frequency band identifier as an index to determine the attenuation characteristic relationship of the corresponding divide-and-conquer frequency band and output the attenuation prediction value.

7. The weak signal detection system for interference fingerprint divide-and-conquer compensation according to claim 6, characterized in that, The divide-and-conquer decision module further includes a mapping model parameter update unit, which includes: The parameters of the mapping model for the attenuation characteristic relationship are recursively updated based on the feedback data. The measured intensity change is determined by the difference between the useful signal strength information before and after anti-interference processing, and the prediction error is determined by the difference between the measured intensity change and the attenuation prediction value. Based on the preset gain constraint mapping relationship, the recursive update lower limit and recursive update upper limit of gain are determined by the signal-to-noise ratio change and the signal detection success rate index; The recursive update gain is determined based on the interference fingerprint stability coefficient, satisfying the following formula: ； In the formula, To interfere with the fingerprint stability coefficient; For the first Recursive update gain for each divide-and-conquer frequency band; To recursively update the lower bound of the gain; To recursively update the upper limit of gain; The signal detection success rate index is an index obtained by statistics within a preset statistical window. Before anti-interference processing, the useful signal strength information is output by the frequency band interference fingerprint sensing module, and after anti-interference processing, the useful signal strength information is output by the anti-interference compensation module.

8. The interference fingerprint divide-and-conquer compensation weak signal detection system according to claim 1, characterized in that, The anti-interference compensation module includes: The frequency band processing execution unit is used to perform weighted suppression on the frequency domain components of the divided frequency band signal within the interference frequency range based on the frequency domain suppression weight matrix, which serves as an anti-interference strategy parameter, and output the divided frequency band signal after anti-interference processing. The compensation control unit is used to obtain the interference fingerprint stability coefficient output by the stability evaluation unit, and determine the compensation weight update gain and compensation gain limiting constraint based on the interference fingerprint stability coefficient according to the preset compensation control mapping relationship. The compensation weight generation unit is used to suppress the weight moments in the frequency domain based on the current update time. Determine the compensation scope, which is related to the frequency domain suppression weight moments at the current update time. The corresponding suppression domains are consistent, and the compensation weight matrix is ​​generated by updating the gain and compensation gain limiting constraint based on the compensation parameters and compensation weights. The compensation execution unit is used to apply a compensation weight matrix to the divided-band signal after anti-interference processing and output the compensated divided-band signal.

9. The interference fingerprint divide-and-conquer compensation weak signal detection system according to claim 7, characterized in that, The signal detection module includes: The detection processing unit is used to receive the processed signal output by the anti-interference compensation module, generate detection statistics for the divide-and-conquer frequency band signal corresponding to the divide-and-conquer frequency band identifier in the processed signal, and the generation of the detection statistics includes correlation matching operation, coherent accumulation operation or non-coherent accumulation operation; and determine the detection threshold according to the preset false alarm probability constraint, and output the detection result according to the comparison result of the detection statistics and the detection threshold. The feedback index generation unit is used to generate a signal detection success rate index within a preset statistical window based on the detection results, using the divide-and-conquer frequency band identifier as an index. It also determines the signal-to-noise ratio after processing based on the estimated signal power and noise power of the processed signal within the corresponding divide-and-conquer frequency band, determines the signal-to-noise ratio before processing based on the useful signal strength information output by the frequency band interference fingerprint sensing module, and generates the signal-to-noise ratio change from the processed signal-to-noise ratio to the signal-to-noise ratio before processing. The feedback scheduling unit is used to obtain the interference fingerprint stability coefficient and determine the statistical window length or feedback output period corresponding to the signal detection success rate index and the signal-to-noise ratio change for the next statistical period based on the interference fingerprint stability coefficient. The feedback interface unit is used to output at least one of the following: signal detection success rate index, signal-to-noise ratio change, statistical window length of the next statistical period or feedback output period, and corresponding divide-and-conquer frequency band identifier to the mapping model parameter update unit.

10. A method for detecting weak signals with interference fingerprint divide-and-conquer compensation, applied to the weak signal detection system with interference fingerprint divide-and-conquer compensation as described in any one of claims 1-9, characterized in that, The method includes: Step 1: Acquire the broadband electromagnetic signal to be tested, determine the frequency band coding information, interference fingerprint information and useful signal strength information, and determine the interference fingerprint stability coefficient based on the interference fingerprint information in multiple time windows; the frequency band coding information is the coding information used to indicate the frequency boundaries or frequency band index of multiple divide-and-conquer frequency bands; Step 2: Divide the broadband electromagnetic signal under test into multiple divide-and-conquer frequency bands based on the frequency band coding information, and determine anti-interference strategy parameters for each divide-and-conquer frequency band based on interference fingerprint information and useful signal strength information; generate attenuation prediction values ​​based on the anti-interference strategy parameters and interference fingerprint information through attenuation characteristic relationships, whereby the attenuation characteristic relationships are the mapping relationships between the anti-interference strategy parameters, interference fingerprint information, and attenuation prediction values, and are represented by mapping model parameters; generate compensation parameters based on the attenuation prediction values, the pre-stored inherent attenuation amount corresponding to the anti-interference strategy parameters, the useful signal strength information, and the preset target signal strength threshold; the compensation parameters are jointly determined by the attenuation prediction values, the pre-stored inherent attenuation amount, and the strength difference between the target signal strength threshold and the useful signal strength information. Step 3: For each divided frequency band, implement anti-interference processing according to the anti-interference strategy parameters, and implement compensation processing according to the compensation parameters to obtain the processed signal of the divided frequency band. Step 4: Perform signal detection based on the processed signal and output the detection result, and generate feedback data based on the detection result; recursively update the mapping model parameters according to the feedback data and the interference fingerprint stability coefficient.

Images