Short-wave signal modulation mode and protocol type combined identification method and system, and medium

Abstract

The application discloses a short-wave signal modulation mode and protocol type joint identification method and system and a medium, and belongs to the technical field of communication, and comprises the following steps: signal preprocessing and multi-modal fusion feature construction; shared feature representation and task branch construction; cross-branch feature association and bidirectional feature compensation; performing hierarchical constraints on the protocol type identification process; executing joint output decision and outputting the checked joint identification result. The short-wave signal modulation mode and protocol type joint identification method and system and the medium in the application can form a closed-loop processing link from signal input, feature construction, identification decision, legality verification to joint output by utilizing the combination design of each step and each module, realize synchronous determination of the short-wave signal modulation mode and the protocol type, improve the consistency and engineering usability of the identification result by utilizing a legal combination constraint mechanism, and provide support and guarantee for the application of the short-wave communication system in a complex communication environment.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, specifically relating to a method, system, and medium for joint identification of shortwave signal modulation mode and protocol type, which is particularly suitable for application scenarios such as signal detection, identification, and link adaptation in shortwave communication. Background Technology

[0002] Shortwave communication, as an important long-distance wireless communication method, relies on high-frequency electromagnetic waves to achieve beyond-line-of-sight propagation through ionospheric reflection. It has the characteristics of strong diffraction capability, long communication distance, low deployment cost, and low dependence on ground backbone network. It is widely used in scenarios such as maritime communication, emergency rescue, aviation telemetry and control, and remote data transmission.

[0003] In practical applications, shortwave communication links are affected by factors such as ionospheric refraction, multipath propagation, Rayleigh fading, random noise, and spectrum congestion, resulting in complex changes in the amplitude, phase, carrier frequency, and spectrum occupancy structure of the received signal. In addition, shortwave links are characterized by rapid fluctuations in channel conditions with time, geography, and solar activity cycles, making the detection of shortwave signals, modulation method identification, and protocol type identification quite challenging.

[0004] Currently, the identification of shortwave signals mainly includes modulation scheme identification and protocol type identification. Modulation scheme identification is mostly based on the amplitude, phase, and symbol variation patterns of the IQ baseband signal to determine the modulation format of the physical layer; protocol type identification is mostly based on frame format, pulse structure, frequency occupancy mode, or message characteristics to determine the type of upper-layer communication protocol.

[0005] In existing signal recognition systems, modulation scheme identification and protocol type identification are usually handled separately. However, due to the correspondence, mapping relationship, or joint occurrence pattern between modulation schemes and protocol types, the separation of the identification process leads to significant engineering bottlenecks in existing identification technologies. On the one hand, the lack of mutual constraints and information sharing at the feature level between modulation scheme identification and protocol type identification can easily result in inconsistent identification results or invalid "modulation-protocol combinations," making it difficult to directly use the identification results for link configuration, device identification, or communication strategy formulation. On the other hand, in complex electromagnetic environments such as ionospheric disturbances, short-term fading, or interference shielding, a single identification task is highly sensitive to noise, frequency drift, frequency offset, or sudden bit errors, resulting in insufficient overall identification robustness and thus affecting the engineering usability of the signal recognition system. Summary of the Invention

[0006] In response to one or more of the above-mentioned defects or improvement needs of the existing technology, the present invention provides a method, system and medium for joint identification of shortwave signal modulation mode and protocol type, which can simultaneously complete the identification of modulation mode and protocol type in complex electromagnetic environment, and perform consistency constraints and legality verification on the identification results, thereby meeting the actual needs of shortwave communication systems in link establishment, device identification and communication strategy selection.

[0007] To achieve the above objectives, one aspect of the present invention provides a method for jointly identifying shortwave signal modulation scheme and protocol type, comprising the following steps: S1: The acquired shortwave signal is preprocessed to construct multimodal features including time domain, spectrum domain and time-frequency domain, and weighted fusion is performed according to the quality index of each mode to obtain multimodal fusion features; S2: Construct a shared feature representation based on multimodal fusion features, and construct task-specific recognition branches for modulation scheme identification and protocol type identification, and output the task-specific features of the two recognition branches respectively; S3: Perform cross-branch feature association calculation and bidirectional feature compensation on the task-specific features output by the two recognition branches; wherein, the bidirectional feature compensation includes: compensating for modulation features using the temporal structure information of the protocol features, and compensating for protocol features using the physical attribute information of the modulation features; S4: Obtain modulation mode identification information and map it to auxiliary constraint information for protocol type identification, construct a legality correlation matrix between modulation mode and protocol type, and execute hierarchical constraints; wherein, the hierarchical constraints include: according to the legality correlation matrix, forcibly setting the predicted probability weight of the illegal protocol candidate class corresponding to the identified modulation mode to zero, and constructing an effective output space for protocol type identification; S5: Perform a joint output decision on the modulation scheme identification result and the protocol type identification result, and output the joint identification result that has passed the legality check and reliability evaluation.

[0008] As a further improvement to the present invention, the following steps are also included: S6: Perform joint credibility evaluation and legality verification on the joint identification results to obtain joint credibility and perform dynamic balance adjustment on the output results.

[0009] As a further improvement of the present invention, step S1 specifically includes: The acquired shortwave signal is subjected to data segmentation, amplitude normalization, and DC removal processing. The time-domain waveform features, frequency-domain power spectrum features, and impulse response time-frequency domain features of the processed signal are extracted as multimodal features. By calculating the energy concentration and signal-to-noise ratio estimates of each modal feature within a preset observation window, the quality index of the corresponding modality is generated. The SoftMax function is used to normalize the quality indices of each modality to obtain the corresponding modal weight parameters, and then the multimodal features are weighted and fused accordingly.

[0010] As a further improvement of the present invention, in step S2, the process of constructing the shared feature representation includes: Feature extraction processes are performed on time-domain features, spectral features, and time-frequency structure features respectively to obtain initial modal feature representations; fusion processing is then performed on each initial modal feature to form a shared feature representation; wherein: The fusion process is implemented using at least one of the following methods: feature splicing, modal weighted superposition, modal mapping fusion, attention-based modal fusion, or modal mapping fusion based on projection or encoder.

[0011] As a further improvement of the present invention, in step S3: The cross-branch feature association calculation is determined based on at least one of the following: feature similarity of two task-specific features, temporal structure correspondence, frequency occupancy correspondence, and signal symbol structure association; and The bidirectional feature compensation includes: using the periodic features of the synchronization header extracted from the protocol branch as a time reference to align and compensate the symbol decision boundary of the modulation branch; and using the instantaneous frequency jump features extracted from the modulation branch as a physical layer trigger threshold to silence invalid data segments of the protocol branch.

[0012] As a further improvement of the present invention, in step S4: The modulation method identification information includes modulation method category label, modulation method probability distribution, or modulation method confidence level; and / or The effective output space is constructed based on at least one of the following: the protocol set corresponding to the modulation mode, the compatibility relationship between the protocol and the modulation mode, the communication link configuration rules, the communication standard specification, or the protocol family division.

[0013] As a further improvement of the present invention, in step S5: The joint output decision is determined based on at least one of the following: a credibility weighting strategy, a candidate set voting strategy, a hierarchical restriction strategy, a protocol set screening strategy, and a protocol weighting strategy based on modulation auxiliary information; The legality verification process is as follows: construct a legal combination relationship matrix between modulation method and protocol type to represent legal modulation-protocol combination methods, and perform legal combination verification on the joint identification results.

[0014] Another aspect of the present invention provides a joint identification system for shortwave signal modulation mode and protocol type, comprising: The signal input module is used to acquire shortwave signals and perform sampling, segmentation, and amplitude normalization processing. A multimodal feature construction module is used to extract time-domain, spectral-domain, and time-frequency-domain features and perform weighted fusion based on quality assessment. The shared feature fusion module is used to perform fusion processing on multimodal features to form a shared feature representation; The task specialization identification module includes a modulation scheme identification submodule and a protocol type identification submodule, which are used to perform modulation scheme identification and protocol type identification respectively based on shared feature representation, and output the task specialization features of modulation scheme and protocol type; The cross-branch association processing module is used to perform association calculations and bidirectional feature compensation on the two types of task-specific features; The hierarchical association and constraint module is used to construct an effective output space for protocol type identification based on modulation identification information and to perform hierarchical constraints on the protocol type identification process. The joint decision-making and legality verification module is used to perform output decision-making and legal combination verification on the modulation identification result and the protocol type identification result; The output module is used to output the joint recognition result that meets the requirements of credibility and legitimacy.

[0015] As a further improvement of the present invention, the identification system further includes: The credibility evaluation and output balancing module is used to evaluate the credibility of the recognition results and fuse the joint credibility, and to perform output balancing adjustment according to communication requirements. The performance feedback module is used to monitor the operating status and channel conditions of the identification system and to adaptively adjust the identification parameters.

[0016] In another aspect, the present invention also provides a storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the aforementioned method for jointly identifying shortwave signal modulation mode and protocol type.

[0017] The aforementioned improved technical features can be combined with each other as long as they do not conflict with each other.

[0018] In summary, the beneficial effects of the above-described technical solutions conceived by this invention compared with the prior art include: The present invention discloses a method for jointly identifying shortwave signal modulation schemes and protocol types, comprising the following steps: signal preprocessing and multimodal fusion feature construction; shared feature representation and task branch construction; cross-branch feature association and bidirectional feature compensation; applying hierarchical constraints to the protocol type identification process; and executing joint output decisions and outputting the verified joint identification results. By combining these steps, a closed-loop processing chain is formed from signal input, feature construction, identification decision, legality verification to joint output, enabling synchronous determination of shortwave signal modulation schemes and protocol types. Furthermore, the legal combination constraint mechanism improves the consistency and engineering usability of the identification results, providing support and assurance for the application of shortwave communication systems in complex communication environments. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in 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.

[0020] Figure 1 This is a flowchart of the joint identification method of shortwave signal modulation mode and protocol type in Embodiment 1 of the present invention; Figure 2 This is a flowchart illustrating the signal preprocessing and multimodal fusion feature construction process in Embodiment 1 of the present invention. Figure 3 This is an architecture diagram of the joint identification system for shortwave signal modulation mode and protocol type in Embodiment 2 of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0022] In the description of this invention, it should be understood that, unless otherwise expressly specified and limited, the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," "circumferential," etc., indicating the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0023] Furthermore, unless otherwise expressly defined, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise expressly and specifically defined.

[0024] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0025] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can mean that the first and second features are in direct contact, or that they are in indirect contact through an intermediate medium. Furthermore, "above," "over," and "on top" of the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0026] Below, for reference Figures 1-3 This invention describes a method, system, and storage medium for jointly identifying shortwave signal modulation scheme and protocol type according to a preferred embodiment of the present invention.

[0027] Specifically, the core technical solution of the present invention is illustrated through the following three embodiments in the preferred embodiments.

[0028] Example 1: In this embodiment, a method for jointly identifying shortwave signal modulation scheme and protocol type is proposed, and its preferred steps are as follows: Figure 1 As shown, and specifically includes the following steps: S1: Signal preprocessing and multimodal fusion feature construction This step mainly involves the following processes: The acquired shortwave signal is preprocessed to construct multimodal features including time domain, spectrum domain, and time-frequency domain. The features are then weighted and fused according to the quality index of each mode to obtain multimodal fused features.

[0029] In practice, the preferred preprocessing method for shortwave signals is to acquire the baseband or intermediate frequency data of the shortwave signal, and then perform data segmentation, amplitude normalization, and DC removal processing on the data to obtain signal segments.

[0030] More specifically, the actual acquisition process of the shortwave signal preferably includes: S11: Acquire shortwave communication signals through shortwave communication receiving equipment; wherein, the shortwave communication signals are represented in the form of in-phase components (I) and quadrature components (Q) (i.e., IQ complex baseband signals), or are intermediate frequency sampling signals (which are subsequently down-converted to IQ form).

[0031] S12: The acquired signal is segmented according to a preset sampling rate and time window to generate multiple independent signal segment samples; each segment is organized into a two-dimensional data matrix, where the rows of the data matrix represent the sampling point sequence and the columns represent the signal channels or components.

[0032] S13: Perform amplitude normalization processing on the data matrix of each segment to adjust the signal amplitude to a uniform range, avoid the subsequent feature extraction offset caused by amplitude differences, and finally obtain the preprocessed signal segment.

[0033] Based on the acquisition of signal segments, the time-domain waveform features, frequency spectrum power spectrum features, and impulse response time-frequency domain features of the processed signal are extracted as multimodal features.

[0034] The specific operation process is as follows: Constructing time-domain feature inputs: (1) Extract time-domain signal features from the preprocessed signal segments. These time-domain signal features include at least amplitude change information and phase change information. (2) Combine the amplitude change and phase change information to form a time domain feature vector, and finally obtain the time domain feature input for subsequent recognition.

[0035] In more detail, when extracting features from time-domain signals, additional features such as envelope features, root mean square energy, and symbol rate features can be extracted as needed.

[0036] Constructing the spectral feature input: (1) Perform frequency domain transformation analysis on the preprocessed signal segments to obtain the corresponding spectrum data; For example, in a specific preferred embodiment, a fast Fourier transform (FFT) is used to obtain a frequency domain amplitude or power spectrum sequence.

[0037] (2) Extract spectral structure features based on spectral data; Among them, the further optimization of spectral structure characteristics includes characteristic data such as spectral energy distribution, bandwidth occupancy range, main frequency offset, and frequency component variation.

[0038] (3) The spectral features are normalized in amplitude and resampled in size, and finally the spectral features with uniform size are obtained.

[0039] Constructing time-frequency structural feature inputs: (1) Perform time-frequency analysis on the preprocessed signal segments to obtain time-frequency structure data describing the joint change relationship of the signal in the time dimension and frequency dimension; In practice, the aforementioned time-frequency analysis and processing methods include, but are not limited to, Short-Time Fourier Transform (STFT), Continuous Wavelet Transform (CWT), and Wigner-Ville Distribution (WVD).

[0040] (2) Extract time-frequency features based on time-frequency structure data; wherein, the preferred time-frequency features include time-frequency energy distribution, frequency jump behavior, duration features and symbol structure.

[0041] (3) The time-frequency features are adjusted in size and normalized in intensity, and finally the time-frequency structural features with uniform size are obtained.

[0042] After acquiring the aforementioned multimodal features, a further modal unification process and modal quality assessment and fusion process are performed. Specifically: The preferred modal unification process includes: (1) The input specifications of time domain, spectrum domain and frequency domain features are uniformly processed; among them, the preferred input specifications include signal segment time length, number of sampling points, number of feature channels and data format, etc.

[0043] (2) Perform amplitude normalization or standardization on different modal features to make their amplitudes fall within a uniform feature scale range.

[0044] The preferred modal quality assessment and fusion process includes: (1) By calculating the energy concentration and signal-to-noise ratio estimates of each modal feature within the preset observation window, the quality index of the corresponding modality is generated.

[0045] (2) Calculate the corresponding modal weights based on each modal quality index to form the weight parameters for modal fusion.

[0046] More specifically, the preferred method for determining the modal weights in weighted fusion is as follows: by calculating the energy envelope smoothness, spectral smoothness, and Rayleigh entropy of the preprocessed signal in the time domain, the SoftMax function is used to normalize the above three indicators, and the result is the feature weight of the corresponding modality.

[0047] (3) Weighted fusion of different modal features according to modal weights to obtain multimodal fusion features.

[0048] Using the above process design, signal segments can be obtained based on the acquired shortwave signals. Based on the signal quality assessment that comprehensively considers the time domain features, time-frequency domain features, and spectrum domain features, a weighted fusion is performed to obtain multimodal feature inputs for use by subsequent modulation mode identification and protocol type identification modules.

[0049] S2: Shared Feature Representation and Task Branch Construction This step mainly involves the following processes: Based on multimodal fusion features, a shared feature representation is constructed, and task-specific recognition branches for modulation scheme identification and protocol type identification are constructed, outputting the task-specific features of the two recognition branches respectively.

[0050] Specifically, a shared feature representation is obtained by deep encoding the multimodal fusion features, and its preferred method includes the following process: (1) Perform feature extraction processing on the time domain features, spectral features and time-frequency structure features respectively to obtain the initial modal feature representation; The feature extraction process here is mainly used to characterize the differences and complementarities of the signal in terms of time structure, frequency structure, and time-frequency joint structure. The extracted features are further optimized to include features such as instantaneous amplitude changes, spectral energy distribution, frequency hopping trajectory, and symbol timing structure.

[0051] More specifically, the feature extraction process in the preferred embodiment preferably includes processes such as encoding, convolution, temporal analysis, statistical analysis, or attention-weighted processing.

[0052] Furthermore, the "initial modal features" obtained through the aforementioned process serve as the basic input for subsequent shared feature construction and task-specific recognition branches.

[0053] (2) Perform fusion processing on each initial modal feature and form a shared feature representation.

[0054] Specifically, the fused shared feature representation serves as the common input for the specialized recognition branch of subsequent tasks. It is used to store the general physical layer features and protocol layer structure features of the signal, thereby simultaneously supporting two different recognition tasks: modulation method recognition and protocol type recognition. This can improve the degree of feature reuse, reduce the resource overhead caused by repeatedly extracting the same information, and improve the efficiency of model resource utilization.

[0055] More specifically, the aforementioned fusion process is preferably implemented using at least one of the following methods: feature concatenation, modal weighted superposition, modal mapping fusion, attention-based modal fusion, or projection- or encoder-based modal mapping fusion.

[0056] Furthermore, based on the acquired shared feature representations, modulation scheme identification branches and protocol type identification branches are constructed, and modulation identification results and protocol identification results are output respectively. Wherein: The modulation scheme identification branch emphasizes the sensitivity to physical layer information such as amplitude variation, phase variation, carrier frequency offset, symbol energy distribution, and frequency jump trajectory. It is used to output the modulation scheme identification result and can characterize and classify physical layer signal attributes such as amplitude modulation characteristics, phase modulation characteristics, and carrier frequency variation characteristics.

[0057] The protocol type identification branch emphasizes sensitivity to protocol layer information such as symbol timing structure, frame structure, synchronization header, pilot sequence, spread spectrum mode, and channel occupancy strategy. It is used to output protocol type identification results and can characterize and classify protocol layer signal structures such as frame structure, pulse structure, symbol timing mode, and spread spectrum structure.

[0058] The two identification branches enhance the ability to identify physical layer and protocol layer information through task specialization, thereby enhancing the distinguishability and independence between modulation method identification and protocol type identification.

[0059] More specifically, the two constructed recognition branches are further subjected to feature enhancement and importance highlighting processes, namely: Feature enhancement modules are set up in both recognition branches to perform weighted processing on shared features or task-specific features to highlight important feature components.

[0060] In actual configuration, the feature enhancement module in the preferred embodiment preferably assigns feature weights based on the importance, discriminative ability, or credibility information of the features, and the configuration strategy of the feature enhancement module in the preferred embodiment is preferably at least one of the following strategies: (a) Region weighting strategy; (b) Temporal weighting strategy; (c) Modality weighting strategy; (d) Feature importance ranking strategy; (e) Attention weighting strategy.

[0061] Feature enhancement processing can further improve the robustness of modulation scheme identification and protocol type identification tasks under complex channel conditions (including noise interference, multipath fading, frequency offset, and masking).

[0062] In step S2, by constructing a sharing and specialization collaboration mechanism, feature reuse can be effectively realized, module resources can be optimized, each task branch can be relatively independent, feature confusion can be avoided, the overall recognition efficiency can be effectively improved, and higher quality feature input can be provided for the downstream correlation compensation module.

[0063] Step S3: Cross-branch feature association and bidirectional feature compensation This step mainly involves the following processes: Perform cross-branch feature association calculation and bidirectional feature compensation on the task-specific features output from the two recognition branches; wherein, the bidirectional feature compensation includes: compensating for modulation features using the temporal structure information of the protocol features, and compensating for protocol features using the physical attribute information of the modulation features; Specifically, step S3 includes the following process: (1) Construct feature association information; (1.1) Take the task-specific features of the modulation mode identification branch output as the first type of feature input, and take the task-specific features of the protocol type identification branch output as the second type of feature input; (1.2) Perform correlation calculation on the first type of features and the second type of features to characterize the correlation between the modulation recognition task features and the protocol type recognition task features at the feature level.

[0064] In practice, the aforementioned "degree of association" is preferably determined based on at least one of the following: (a) feature similarity; (b) temporal structure correspondence; (c) frequency occupancy correspondence; and (d) signal symbol structure association. Furthermore, the preferred method for calculating the degree of association in the preferred embodiment includes a distance metric model, an attention weighting mechanism, statistical correlation analysis, or other means that can perform association analysis.

[0065] (2) Two-way feature compensation processing is performed to obtain enhanced modulation features and protocol features; Based on the acquisition of feature association information, the following bidirectional compensation processes are performed respectively: First-direction compensation: Based on the second type of feature (protocol feature), compensate the first type of feature (modulation feature), and use the time dimension feature (time structure information) of the protocol type branch to reconstruct the symbol boundary information of the modulation branch to compensate for the timing alignment deviation of the modulation feature.

[0066] Second-direction compensation: Based on the first type of feature (modulation feature), compensate the second type of feature (protocol feature). Utilize the signal envelope and instantaneous frequency features (physical attribute information) of the modulation mode branch to perform weighted correction on the feature channels of the protocol type branch in order to suppress background noise interference from non-communication protocol types.

[0067] In the process of compensating modulation features for protocol features, since protocol features typically contain obvious periodic structures in the frame synchronization header or pilot sequence, this periodic structure information can be used to help the modulation branch more accurately locate symbol jump points, thereby solving the symbol synchronization ambiguity problem caused by shortwave multipath effects. Conversely, in the process of compensating protocol features for modulation features, the instantaneous frequency jitter or phase jump features extracted by the modulation branch can serve as physical layer features, helping the protocol branch quickly eliminate pulse interference that, although similar to protocol packet structures in the time domain, does not conform to communication specifications in the modulation domain.

[0068] Therefore, in practice, bidirectional feature compensation includes the following process: using the periodic features of the synchronization header extracted from the protocol branch as a time reference, the symbol decision boundary of the modulation branch is aligned and compensated; at the same time, the instantaneous frequency jump features extracted from the modulation branch are used as the physical layer trigger threshold to silence invalid data segments of the protocol branch.

[0069] After the aforementioned compensation process, enhanced modulation feature representation and enhanced protocol feature representation can be formed, thereby improving the degree of information interaction and complementarity between the two tasks.

[0070] More specifically, the compensation method used in the aforementioned bidirectional feature compensation process is preferably at least one of the following processing methods: (a) feature weighted superposition; (b) feature splicing mapping; (c) feature selection and pruning; (d) attention enhancement compensation.

[0071] (3) Perform fusion processing on the enhanced features and the original task-specific features to obtain fused features; By utilizing the fusion process, cross-task related information and task-specific feature information can be integrated. The resulting fused features contain both cross-task related information and maintain the specificity of each task.

[0072] More specifically, in actual fusion processing, the fusion ratio between the original features and the enhanced features can be adaptively adjusted according to the importance of the features, the feature discrimination ability, or the feature contribution.

[0073] In addition, the preferred embodiment uses at least one of the following methods for fusion processing: (1) weighted fusion; (2) proportional weighting; (3) feature filtering; (4) region mapping fusion.

[0074] (4) Enhanced consistency between the fused modulation features and protocol features.

[0075] By performing a consistency check on the fused modulation features and protocol features, conflicting information at the feature level can be avoided, thereby improving the stability and effectiveness of subsequent identification stages.

[0076] In practice, the aforementioned consistency check is preferably performed based on at least one of the following: (a) structural consistency; (b) timing consistency; (c) signal format consistency; (d) symbol mapping consistency.

[0077] S4: Implement hierarchical constraints for the protocol type identification process. This step mainly involves the following processes: The modulation scheme identification information is obtained and mapped to auxiliary constraint information for protocol type identification. A legality correlation matrix between the modulation scheme and the protocol type is constructed, and hierarchical constraints are executed. The hierarchical constraints include: according to the legality correlation matrix, the predicted probability weights of the illegal protocol candidate classes corresponding to the identified modulation schemes are forced to zero, and an effective output space for protocol type identification is constructed. Specifically, based on obtaining the enhanced modulation features and enhanced protocol features in step S3, step S4 establishes a hierarchical relationship between the modulation method and the protocol type, and uses modulation identification information to effectively constrain the protocol type identification process.

[0078] More specifically, step S4 in the preferred embodiment further preferably includes the following process: (1) Perform modulation information mapping processing: Obtain the modulation identification information output from the modulation method identification branch, and map the modulation identification information into auxiliary information (protocol compatibility mask vector) for protocol type identification.

[0079] Among them, the auxiliary information represents the protocol selection range or protocol feature correlation corresponding to the modulation mode. It is preferred to act on the output layer of the protocol type determination branch to determine the prediction probability weight of the protocol candidate class, thereby affecting the joint determination initial value of the output.

[0080] Meanwhile, the modulation method identification information output by the modulation method identification branch preferably includes modulation method category label, modulation method probability distribution, or modulation method credibility.

[0081] Meanwhile, the aforementioned mapping process is preferably implemented through table-driven mapping, rule-based mapping, or a training-based mapping network.

[0082] (2) Protocol type identification hierarchy association: In the protocol type identification branch, modulation auxiliary information is introduced to construct a legality correlation matrix between modulation mode and protocol type, and the protocol candidate set or protocol feature range is adjusted according to the legality correlation matrix.

[0083] By introducing modulation auxiliary information, the protocol type identification process can be hierarchically determined based on the modulation method, fully reflecting the hierarchical dependency between the modulation method and the protocol type.

[0084] More specifically, in practice, the aforementioned hierarchical association is preferably implemented through at least one of the following methods: (a) protocol feature selection strategy; (b) protocol candidate set pruning; (c) protocol probability distribution weighting; (d) protocol output confidence adjustment.

[0085] (3) Effective output space constraints: Based on the preset correspondence between modulation schemes and protocol types (i.e., the legality correlation matrix), the predicted probability weights of the illegal protocol candidate classes corresponding to the identified modulation schemes are forcibly set to zero, thus constructing an effective output space for protocol type identification.

[0086] By utilizing the constraints of the effective output space, the protocol type identification branch is made to perform classification or selection only within the effective output space, thereby avoiding the generation of modulation-protocol combination relationships that do not conform to the actual communication system.

[0087] More preferably, the aforementioned effective output space is further preferably constructed based on at least one of the following: (a) the protocol set corresponding to the modulation method; (b) the compatibility relationship between the protocol and the modulation method; (c) the communication link configuration rules; and (d) the communication standard specification or protocol family division.

[0088] Meanwhile, the effective output space is further optimized to be represented as a set, matrix, or mapping table structure.

[0089] (4) Consistency check of joint features: A consistency check is performed on the modulation identification result and the protocol type identification result to determine whether the combination conforms to the preset hierarchical association relationship, thereby realizing the hierarchical constraint on the protocol type identification process within the effective output space.

[0090] The consistency check preferably includes at least one of the following: (a) configuration consistency check; (b) symbol structure consistency check; (c) frame format consistency check; (d) communication standard consistency check.

[0091] Meanwhile, for cases where the consistency check fails, at least one of the following processes is preferred to be executed: (a) mark the abnormal identification; (b) correct the protocol type identification result; (c) output the candidate protocol set; (d) trigger the upper-layer link policy adjustment.

[0092] S5: Perform the joint output decision and output the verified joint recognition result. This step mainly involves the following processes: Perform a joint output decision on the modulation scheme identification result and the protocol type identification result, and output the joint identification result that passes the legality check and reliability evaluation.

[0093] Specifically, this step is further preferably included in the following specific processes: (1) Independent branch output processing: The modulation scheme identification information and the protocol type identification information are output independently through two identification branches; wherein, the modulation scheme identification information preferably includes a modulation scheme category label, a modulation scheme probability distribution or confidence information; the protocol type identification information preferably includes a protocol type category label, a protocol type probability distribution or confidence information.

[0094] By maintaining structural independence between the two types of recognition information in the output stage, the hierarchical features of each recognition task can be preserved.

[0095] (2) Joint decision-making Construct a joint output decision and perform a joint decision on modulation mode identification information and protocol type identification information.

[0096] Among them, the joint output decision is used to ensure the coordination and consistency between the modulation scheme identification result and the protocol type identification result at the output level.

[0097] Meanwhile, in the preferred embodiment, the joint output decision is preferably determined based on at least one of the following strategies: (a) a credibility weighting strategy; (b) a candidate set voting strategy; (c) a hierarchy restriction strategy; (d) a protocol set screening strategy; and (e) a protocol weighting strategy based on modulation auxiliary information.

[0098] (3) Perform reliability evaluation processing: The credibility evaluation index is calculated for modulation method identification information and protocol type identification information respectively.

[0099] In a preferred embodiment, the aforementioned credibility evaluation index is preferably determined based on at least one of the following: (a) the maximum credibility value; (b) the credibility difference; (c) the credibility entropy distribution; (d) the confidence interval index; and (e) the kurtosis or skewness characteristics of the credibility distribution.

[0100] More specifically, in the aforementioned joint output decision-making, it is further optimized to perform reliability labeling, result screening, or reliability compensation on the joint identification results based on credibility evaluation indicators.

[0101] (4) Perform valid combination verification: Construct a matrix of legal combination relationships between modulation schemes and protocol types to represent legal modulation-protocol combination schemes, and perform legal combination verification on the joint identification results; if the joint identification results belong to legal combinations, output the joint identification results and their reliability indicators.

[0102] In practice, the aforementioned legal combination relationship matrix is ​​preferably preset by communication standard specifications, protocol compatibility relationships, or link configuration rules.

[0103] For example, in one specific implementation, the legitimate combination relationship matrix is ​​constructed based on communication standards and historical reconnaissance data. This matrix is ​​an M×N binary mask table, where rows represent M modulation schemes and columns represent N protocol types. A matrix element of 1 indicates that the modulation scheme can carry the corresponding protocol type, while a matrix element of 0 indicates a mismatch between the physical layer and the link layer. More preferably, this matrix supports dynamic updates through online learning or manual rule injection.

[0104] In addition, if the joint identification result is an illegal combination, at least one of the following processing strategies is preferably executed: (a) result correction; (b) output candidate legal combinations; (c) output anomaly flag; (d) request the upper-layer module to re-identify; (e) output a separate identification result containing only the modulation method or protocol type.

[0105] By setting the corresponding steps S1 to S5, the synchronous determination of shortwave signal modulation mode and protocol type can be achieved, and the consistency and engineering usability of the identification results can be improved by using the legal combination constraint mechanism, thereby meeting the application requirements of shortwave communication link establishment, equipment identification and communication strategy selection in complex electromagnetic environments.

[0106] More specifically, to further improve the reliability of shortwave signal recognition in complex communication environments, the method in the preferred embodiment further performs the following steps before the final output of the joint output result: S6: Joint Credibility Evaluation and Output Balancing Process This step mainly involves the following processes: The credibility fusion of joint credibility evaluation and legality verification is performed on the joint identification results to obtain joint credibility, and dynamic balance adjustment is performed on the output results.

[0107] Specifically, based on the joint identification result obtained in step S5, step S6 further optimizes the performance of credibility evaluation and legality fusion judgment on the modulation mode identification output and protocol type identification output, and balances the joint output result to improve the identification reliability in complex communication environments.

[0108] Step S6 is further preferably included in the following sub-processes: (1) Evaluation of credibility: The credibility evaluation indexes of modulation mode identification results and protocol type identification results are calculated separately to obtain modulation credibility and protocol credibility.

[0109] The credibility evaluation index is used for subsequent steps such as legality determination, credibility fusion and output balancing. It is preferably determined based on at least one of the following methods: (a) maximum output value; (b) output difference; (c) output distribution concentration; (d) credibility entropy; (e) historical decision consistency; (f) category confidence interval.

[0110] (2) Determination of the legality of the combination: Construct a table of legal combinations between modulation methods and protocol types, and determine the legality of the modulation method identification results and protocol type identification results based on the table of legal combinations to determine whether the identification results belong to legal combinations.

[0111] The valid combination table is used to represent the modulation-protocol combination methods allowed in the communication system.

[0112] If the identification result does not belong to a legitimate combination, at least one of the following processing strategies can be executed: (a) result correction; (b) outputting a set of candidate legitimate combinations; (c) adjusting the confidence weight; (d) marking anomalies; (e) reporting to the upper-level module or decision system.

[0113] (3) Perform joint credibility fusion on legitimate combinations: The credibility of modulation credibility and protocol credibility are fused to obtain a joint credibility evaluation result.

[0114] The fused joint credibility is used to measure the reliability of the joint identification results and is used for subsequent output balancing strategies.

[0115] Meanwhile, the credibility fusion method preferably adopts at least one of the following methods: (a) weighted average; (b) dynamic proportion allocation; (c) hierarchical priority rule; (d) joint scoring mechanism; (e) Bayes fusion mechanism; (f) soft voting or hard voting fusion.

[0116] (4) Output dynamic balancing strategy: Based on the reliability requirements of the joint identification task, output balancing adjustment is performed on the modulation scheme identification result and the protocol type identification result.

[0117] The dynamic balancing strategy is used to maintain the availability of output results and the continuity of decision-making under different channel conditions (such as noise interference, multipath, frequency offset changes), and the balancing strategy in the preferred embodiment is preferably constructed based on at least one of the following: (a) the difference in credibility between the two; (b) the stability of historical decisions; (c) the requirements of the communication scenario; (d) the link status index; (e) the target service type; and (f) the task priority parameter.

[0118] (5) Final result determination and output If the joint identification result satisfies both the credibility requirement and the legal combination relationship, then the joint identification result is output; which includes at least the modulation method identification result and the protocol type identification result.

[0119] If the joint identification result does not meet the credibility requirements, at least one of the following output strategies is preferred: (a) output credibility level; (b) output candidate result set; (c) request resampling or re-identification; (d) request the upper-layer module to make manual or rule-based judgment; (e) output only modulation identification result or only protocol type identification result.

[0120] By setting the steps in step S6 above, the reliability of the joint identification results output can be further improved, ensuring that the method in the preferred embodiment can meet the usage requirements of different application scenarios and promote the application and development of shortwave communication technology.

[0121] Example 2: As another aspect of the present invention, a joint identification system for shortwave signal modulation mode and protocol type is also provided, wherein the system preferably includes: The signal input module is used to acquire shortwave signals and perform sampling, segmentation, and amplitude normalization processing. A multimodal feature construction module is used to extract time-domain, spectral-domain, and time-frequency-domain features and perform weighted fusion based on quality assessment. The shared feature fusion module is used to perform fusion processing on multimodal features to form a shared feature representation; The task specialization identification module includes a modulation scheme identification submodule and a protocol type identification submodule, which are used to perform modulation scheme identification and protocol type identification respectively based on shared feature representation, and output the task specialization features of modulation scheme and protocol type; The cross-branch association processing module is used to perform association calculations and bidirectional feature compensation on the two types of task-specific features; The hierarchical association and constraint module is used to construct an effective output space for protocol type identification based on modulation identification information and to perform hierarchical constraints on the protocol type identification process. The joint decision-making and legality verification module is used to perform output decision-making and legal combination verification on the modulation identification result and the protocol type identification result; The output module is used to output the joint recognition result that meets the requirements of credibility and legitimacy.

[0122] By utilizing the combined configuration of the aforementioned modules, the system in the preferred embodiment can achieve joint identification of shortwave signal modulation mode and protocol type, ensuring the reliability and accuracy of shortwave signal communication identification.

[0123] Furthermore, in the actual system configuration, the following configuration is preferred: The credibility evaluation and output balancing module is used to evaluate the credibility of the recognition results and fuse the joint credibility, and to perform output balancing adjustment according to communication requirements.

[0124] More preferably, it also includes an operating performance feedback module for monitoring the operating status and channel conditions of the identification system and performing adaptive adjustments to the identification parameters.

[0125] In actual configuration, the data flow relationship between the aforementioned modules follows the processing flow in steps S1 to S6, forming a closed-loop processing link from signal input, feature construction, identification decision, legality verification to joint output. This is effectively applicable to joint identification tasks of modulation mode and protocol type in actual shortwave communication equipment, communication monitoring equipment, or signal reconnaissance equipment.

[0126] Furthermore, the system in the preferred embodiment preferably includes the following process during actual deployment: (1) System initialization and parameter configuration The identification modules (i.e., modulation method identification submodule and protocol type identification submodule) of the modulation method identification branch and protocol type identification branch are deployed in the target operating environment, and other modules are configured accordingly; wherein, the operating environment preferably includes embedded devices, communication terminals, monitoring systems or server platforms.

[0127] Configure system operating parameters according to the processing capacity of communication equipment, channel conditions and service scenarios; among them, preferred operating parameters include input buffer size, data segment length, recognition period, number of running threads and caching strategy.

[0128] After completing the initialization of the recognition module, cache, and hardware driver, the system enters the standby state.

[0129] (2) Pre-deployment verification process Before official operation, a deployment verification process is performed to check whether the data processing link of the identification system is complete and whether the module output is normal.

[0130] The aforementioned deployment verification is preferably performed based on at least one of the following methods: (a) Identification and verification using known shortwave signal samples; (b) Check the output stability of the modulation recognition module; (c) Check the consistency of the output of the protocol type identification module; (d) Check the validity of the legality constraints on the modulation-protocol combination; (e) Check the compatibility of the buffer with the identification cycle configuration.

[0131] If the deployment verification passes, the system will proceed to the formal operation phase; otherwise, parameter adjustments or module reconfiguration will be performed.

[0132] In addition, for the deployed system, a runtime performance feedback mechanism is further configured in conjunction with the runtime performance feedback module settings. The specific process is as follows: (1) Configure a monitoring submodule with a performance feedback mechanism in the performance feedback module to monitor and identify the system's performance and status during system operation and perform performance feedback processing.

[0133] The operational performance feedback indicators are preferably constructed based on at least one of the following operational data: (a) identification success rate; (b) illegal combination occurrence rate; (c) credibility distribution characteristics; (d) identification time and resource overhead; (e) link quality or channel condition change trend.

[0134] Meanwhile, in the preferred embodiment, the performance feedback index is further preferably capable of performing at least one of the following actions by the upper-layer link control module or communication strategy module: (a) channel switching; (b) protocol switching; (c) transmission parameter adjustment; (d) sampling parameter adjustment; (e) identification period adjustment.

[0135] (2) Adjust the operating parameters of the recognition module according to the performance feedback; wherein the operating parameters may include input length, recognition period, filtering threshold, confidence threshold or buffer strategy.

[0136] If channel conditions change (including increased noise, enhanced multipath, intensified fading, or frequency offset), an adaptive adjustment process can be triggered to optimize the identification system's operating status.

[0137] The availability and stability of the identification system under complex channel conditions are improved by an adaptive adjustment mechanism for operating status.

[0138] Example 3: It is understood that the method steps in Embodiment 1 can be implemented by an electronic device executing program instructions, or by software, hardware, or a combination of both.

[0139] The program instructions can be stored in a computer-readable storage medium, and the electronic device involved includes a processor and a memory. The processor executes the program instructions in the memory to implement the method steps in Embodiment 1. Conversely, the software is preferably deployed on a server, embedded device, or other processing device with computing capabilities.

[0140] Therefore, as another aspect of the present invention, a storage medium is also provided, wherein a processor-executable program is stored, which, when executed by a processor, is used to perform the shortwave signal modulation mode and protocol type joint identification method described in Embodiment 1.

[0141] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for jointly identifying shortwave signal modulation mode and protocol type, characterized in that, Includes the following steps: S1: The acquired shortwave signal is preprocessed to construct multimodal features including time domain, spectrum domain and time-frequency domain, and weighted fusion is performed according to the quality index of each mode to obtain multimodal fusion features; S2: Construct a shared feature representation based on multimodal fusion features, and construct task-specific recognition branches for modulation scheme identification and protocol type identification, and output the task-specific features of the two recognition branches respectively; S3: Perform cross-branch feature association calculation and bidirectional feature compensation on the task-specific features output by the two recognition branches; wherein, the bidirectional feature compensation includes: compensating for modulation features using the temporal structure information of the protocol features, and compensating for protocol features using the physical attribute information of the modulation features; S4: Obtain modulation mode identification information and map it to auxiliary constraint information for protocol type identification, construct a legality correlation matrix between modulation mode and protocol type, and execute hierarchical constraints; wherein, the hierarchical constraints include: according to the legality correlation matrix, forcibly setting the predicted probability weight of the illegal protocol candidate class corresponding to the identified modulation mode to zero, and constructing an effective output space for protocol type identification; S5: Perform a joint output decision on the modulation scheme identification result and the protocol type identification result, and output the joint identification result that has passed the legality check and reliability evaluation.

2. The method for jointly identifying shortwave signal modulation mode and protocol type according to claim 1, characterized in that, It also includes the following steps: S6: Perform joint credibility evaluation and legality verification on the joint identification results to obtain joint credibility and perform dynamic balance adjustment on the output results.

3. The method for jointly identifying shortwave signal modulation mode and protocol type according to claim 1 or 2, characterized in that, Step S1 specifically includes: The acquired shortwave signal is subjected to data segmentation, amplitude normalization, and DC removal processing. The time-domain waveform features, frequency-domain power spectrum features, and impulse response time-frequency domain features of the processed signal are extracted as multimodal features. By calculating the energy concentration and signal-to-noise ratio estimates of each modal feature within a preset observation window, the quality index of the corresponding modality is generated. The SoftMax function is used to normalize the quality indices of each modality to obtain the corresponding modal weight parameters, and then the multimodal features are weighted and fused accordingly.

4. The method for jointly identifying shortwave signal modulation mode and protocol type according to claim 1 or 2, characterized in that, In step S2, the process of constructing the shared feature representation includes: Feature extraction processes are performed on time-domain features, spectral features, and time-frequency structure features respectively to obtain initial modal feature representations; fusion processing is then performed on each initial modal feature to form a shared feature representation; wherein: The fusion process is implemented using at least one of the following methods: feature splicing, modal weighted superposition, modal mapping fusion, attention-based modal fusion, or modal mapping fusion based on projection or encoder.

5. The method for jointly identifying shortwave signal modulation mode and protocol type according to claim 1 or 2, characterized in that, In step S3: The cross-branch feature association calculation is determined based on at least one of the following: feature similarity of two task-specific features, temporal structure correspondence, frequency occupancy correspondence, and signal symbol structure association; and The bidirectional feature compensation includes: using the periodic features of the synchronization header extracted from the protocol branch as a time reference to align and compensate the symbol decision boundary of the modulation branch; and using the instantaneous frequency jump features extracted from the modulation branch as a physical layer trigger threshold to silence invalid data segments of the protocol branch.

6. The method for jointly identifying shortwave signal modulation mode and protocol type according to claim 1 or 2, characterized in that, In step S4: The modulation method identification information includes modulation method category label, modulation method probability distribution, or modulation method confidence level; and / or The effective output space is constructed based on at least one of the following: the protocol set corresponding to the modulation mode, the compatibility relationship between the protocol and the modulation mode, the communication link configuration rules, the communication standard specification, or the protocol family division.

7. The method for jointly identifying shortwave signal modulation mode and protocol type according to claim 1 or 2, characterized in that, In step S5: The joint output decision is determined based on at least one of the following: a credibility weighting strategy, a candidate set voting strategy, a hierarchical restriction strategy, a protocol set screening strategy, and a protocol weighting strategy based on modulation auxiliary information; The legality verification process is as follows: construct a legal combination relationship matrix between modulation method and protocol type to represent legal modulation-protocol combination methods, and perform legal combination verification on the joint identification results.

8. A system for jointly identifying shortwave signal modulation mode and protocol type, characterized in that, include: The signal input module is used to acquire shortwave signals and perform sampling, segmentation, and amplitude normalization processing. A multimodal feature construction module is used to extract time-domain, spectral-domain, and time-frequency-domain features and perform weighted fusion based on quality assessment. The shared feature fusion module is used to perform fusion processing on multimodal features to form a shared feature representation; The task specialization identification module includes a modulation scheme identification submodule and a protocol type identification submodule, which are used to perform modulation scheme identification and protocol type identification respectively based on shared feature representation, and output the task specialization features of modulation scheme and protocol type; The cross-branch association processing module is used to perform association calculations and bidirectional feature compensation on the two types of task-specific features; The hierarchical association and constraint module is used to construct an effective output space for protocol type identification based on modulation identification information and to perform hierarchical constraints on the protocol type identification process. The joint decision-making and legality verification module is used to perform output decision-making and legal combination verification on the modulation identification result and the protocol type identification result; The output module is used to output the joint recognition result that meets the requirements of credibility and legitimacy.

9. The shortwave signal modulation mode and protocol type joint identification system according to claim 8, characterized in that, The identification system also includes: The credibility evaluation and output balancing module is used to evaluate the credibility of the recognition results and fuse the joint credibility, and to perform output balancing adjustment according to communication requirements. The performance feedback module is used to monitor the operating status and channel conditions of the identification system and to adaptively adjust the identification parameters.

10. A storage medium storing a processor-executable program, characterized in that, When executed by a processor, the program is used to perform the joint identification method of shortwave signal modulation mode and protocol type as described in any one of claims 1 to 7.

Images