A target signal detection method, device and electronic equipment based on internal and external interaction features

Abstract

The application provides a target signal detection method and device based on internal and external interaction features and an electronic device. The method comprises: modulating an ACARS signal collected to obtain an internal modulation signal and an external modulation signal of the ACARS signal; performing time-frequency domain conversion on the internal modulation signal and the external modulation signal to obtain an internal modulation time-frequency spectrum and an external modulation time-frequency spectrum, and performing point-by-point multiplication operation on the amplitude spectrum of the internal modulation time-frequency spectrum and the amplitude spectrum of the external modulation time-frequency spectrum to obtain an internal and external interaction feature map corresponding to the ACARS signal; processing the internal and external interaction feature map through a target detection model to obtain a target detection result; mapping the boundary box coordinates of the target rectangular frame to obtain an estimated value of the carrier frequency and an estimated value of the bandwidth of the ACARS signal, and performing frequency shift correction and band pass filtering on the ACARS signal based on the estimated value of the carrier frequency and the estimated value of the bandwidth to separate out the target signal. The technical scheme of the application stably detects and separates out the ACARS target signal from a complex electromagnetic environment.

Description

Technical Field

[0001] This invention relates to the field of deep learning technology, and specifically to a target signal detection method, apparatus, and electronic device based on internal and external interaction features. Background Technology

[0002] Parameter estimation of civil aviation communication signals is a crucial topic in the field of communication engineering. Taking ACARS (Aircraft Communications Addressing and Reporting System) data link signals as an example, the estimation accuracy of key parameters such as carrier frequency deviation, symbol timing deviation, burst start and end times, occupied bandwidth, and Doppler disturbance directly affects the receiver's synchronization, demodulation, and decoding performance, and determines link availability and message transmission reliability. Especially in weak signal, high-dynamic, and complex electromagnetic environments, parameter mismatch can easily lead to frame synchronization failure, increased bit error rate, and missed message detection. Therefore, conducting research on high-precision and robust parameter estimation for civil aviation communication signals is not only fundamental to improving physical layer receiver performance but also a vital technological foundation for supporting the safe operation and intelligent sensing of civil aviation communication systems.

[0003] Currently, relevant research approaches mainly include parameter estimation methods based on feature extraction and those based on deep learning. The former relies on spectral feature analysis and manual criteria, and is prone to estimation bias under conditions of low signal-to-noise ratio and strong interference; while the latter has strong feature representation and anti-interference capabilities, it still suffers from problems such as strong data dependence and insufficient generalization across scenarios.

[0004] In feature extraction-based parameter estimation methods, Zhu Hang et al., in their proposed method for estimating key parameters of Binary Phase Shift Keying (BPSK) signals based on spectral feature analysis, employed signal power spectrum, Chirp-Z transform spectrum of squared terms, and envelope spectrum analysis, combined with Singular Value Decomposition (SVD) to extract signal features for estimating the symbol rate, carrier frequency, roll-off factor, and delay of BPSK signals. However, this method, based on a simplified periodic modulation signal model, is only applicable to idealized signal conditions and fails to fully consider the complex modulation schemes used in ACARS signals. Especially in low signal-to-noise ratio and complex civil aviation communication environments, the estimation accuracy and robustness of this method are significantly limited, making it difficult to effectively address non-ideal factors such as frequency offset, multipath effects, and differences in modulation characteristics.

[0005] Liu Tengfei proposed a stepwise estimation method for modulation parameters based on cyclic spectrum and ZAM transform, which can estimate the carrier frequency and symbol parameters of composite modulation signals. However, under conditions of low signal-to-noise ratio, adjacent channel interference, frequency offset, and clock skew, this method is prone to spectral peak broadening, feature weakening, and peak aliasing, thus affecting the stability of carrier frequency estimation. At the same time, the criterion for negative peaks in phase transformation by ZAM transform is quite sensitive to noise disturbances, channel fading, multipath propagation, and amplitude and phase distortion. In the complex electromagnetic environment of civil aviation, its parameter estimation accuracy and robustness will significantly decrease.

[0006] Liu Yanzhe proposed a wavelet transform-based algorithm for estimating the bandwidth and carrier frequency of Orthogonal Frequency Division Multiplexing (OFDM) signals. This method introduces wavelet transform on the basis of Welch power spectrum estimation, thereby improving the accuracy of carrier frequency and bandwidth estimation for OFDM signals. However, the modeling premise and feature extraction mechanism of this method are mainly designed for broadband multi-carrier OFDM signals, and are highly dependent on the sharpness of spectral edges, spectral stability, and the distinguishability of boundary abrupt changes. ACARS signals belong to a narrowband composite modulation system, and their spectral structure is fundamentally different from that of OFDM. This makes the boundary detection based on wavelet difference prone to estimation errors under the influence of amplitude fluctuations, frequency offset, adjacent channel interference, and noise disturbances.

[0007] In summary, existing parameter estimation methods based on feature extraction generally suffer from low signal-to-noise ratio, insufficient estimation accuracy, and poor robustness when dealing with narrowband composite modulation signals such as ACARS. Due to inaccurate parameter estimation, subsequent frequency shift correction and bandpass filtering based on the estimated parameters are difficult to effectively separate the complete target signal, resulting in a serious decline in signal detection and extraction performance, which is difficult to meet the actual requirements of civil aviation communication systems for high-reliability parameter estimation. Summary of the Invention

[0008] In view of this, this application proposes a target signal detection method, apparatus and electronic device based on internal and external interaction characteristics, so as to robustly detect and separate ACARS target signals from complex electromagnetic environments.

[0009] Specifically, this application is implemented through the following technical solution: According to a first aspect of the embodiments of this specification, a target signal detection method based on internal and external interaction features is provided, comprising: Step S1: Modulate the acquired ACARS signal to obtain the inner modulation signal and the outer modulation signal of the ACARS signal; Step S2: Perform time-frequency domain conversion on the internal modulation signal and the external modulation signal to obtain the internal modulation time spectrum and the external modulation time spectrum, and perform point-by-point multiplication operation on the amplitude spectrum of the internal modulation time spectrum and the amplitude spectrum of the external modulation time spectrum to obtain the internal and external interaction feature map corresponding to the ACARS signal. Step S3: The internal and external interaction feature maps are processed by a pre-built target detection model to obtain the target detection result. The target detection model includes a backbone network, a feature fusion network, and a detection head. The backbone network introduces an ODConv module to enhance feature modeling of weakly structured regions. The feature fusion network introduces an SCSA attention module to enhance the effective region response related to signal carrier frequency and bandwidth estimation. The detection head is used to output the bounding box coordinates of the target rectangle. Step S4: Map the bounding box coordinates of the target rectangle based on the frequency axis calibration relationship to obtain the carrier frequency estimate and bandwidth estimate of the ACARS signal, and perform frequency shift correction and bandpass filtering on the ACARS signal based on the carrier frequency estimate and bandwidth estimate to separate the target signal.

[0010] According to a second aspect of the embodiments of this specification, a target signal detection device based on internal and external interaction features is provided, comprising: The signal adjustment unit is used to modulate the acquired ACARS signal to obtain the internal modulation signal and the external modulation signal of the ACARS signal. The mutual feature calculation unit is used to perform time-frequency domain conversion on the inner modulation signal and the outer modulation signal to obtain the inner modulation time spectrum and the outer modulation time spectrum, and to perform point-by-point multiplication operation on the amplitude spectrum of the inner modulation time spectrum and the amplitude spectrum of the outer modulation time spectrum to obtain the inner and outer interaction feature map corresponding to the ACARS signal. The target detection unit is used to process the internal and external interaction feature maps through a pre-built target detection model to obtain target detection results. The target detection model includes a backbone network, a feature fusion network, and a detection head. The backbone network introduces an ODConv module to enhance feature modeling of weakly structured regions. The feature fusion network introduces an SCSA attention module to enhance the effective region response related to signal carrier frequency and bandwidth estimation. The detection head is used to output the bounding box coordinates of the target rectangle. The signal processing unit is used to map the bounding box coordinates of the target rectangle based on the frequency axis calibration relationship to obtain the carrier frequency estimate and bandwidth estimate of the ACARS signal, and to perform frequency shift correction and bandpass filtering on the ACARS signal based on the carrier frequency estimate and bandwidth estimate to separate the target signal.

[0011] According to a third aspect of the embodiments of this specification, an electronic device is provided, comprising: a processor; and a computer-readable storage medium storing computer program instructions that, when executed by the processor, cause the processor to perform the method as described in the first aspect.

[0012] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, characterized in that a computer program is stored on the computer-readable storage medium, the computer program being executed by a processor as described in the first aspect.

[0013] The embodiments of this application have at least the following technical effects: First, it enhances the accuracy and robustness of parameter estimation in weak signals and complex environments. By constructing an internal and external interactive feature map and introducing the ODConv module and SCSA attention mechanism, the embodiments of this application can still maintain high-precision estimation of the carrier frequency and bandwidth of ACARS signals under non-ideal conditions such as low signal-to-noise ratio, strong interference, frequency offset drift and adjacent channel interference, which is significantly better than traditional methods and mainstream YOLO series models; Second, it suppresses noise and spurious peak interference, improving the discriminative power of feature representation. The internal and external interactive feature maps in this application strengthen the consistent response region of the internal and external modulation spectra through point-by-point multiplication operations, effectively suppressing noise fluctuations, isolated spurious peaks and local interference, making the model more focused on the stable time-frequency structure related to parameter estimation, thereby improving the separability and stability of detection; Third, it supports adaptive modeling and feature recalibration in dynamic environments. The ODConv module enables the convolution kernel to dynamically adjust with the input content, enhancing the modeling ability for fuzzy boundaries and multi-scale targets; the SCSA module strengthens the response of effective regions and suppresses background interference through a spatial and channel-based collaborative attention mechanism, further improving the model's generalization ability and robustness. Attached Figure Description

[0014] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Some specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings in an exemplary and non-limiting manner. The same reference numerals in the drawings indicate the same or similar parts or components. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a simplified schematic diagram illustrating a target signal detection process according to an exemplary embodiment of this application; Figure 2 This is a schematic flowchart illustrating a target signal detection method based on internal and external interaction features, as shown in an exemplary embodiment of this application. Figure 3 This is a comparative schematic diagram of various special sample drawings shown in an exemplary embodiment of this application; Figure 4 This is a schematic diagram of the architecture of an object detection model shown in an exemplary embodiment of this application; Figure 5 This is a schematic diagram of the structure of an ODConv module shown in an exemplary embodiment of this application; Figure 6 This is a schematic diagram illustrating the structure of an SCSA attention module according to an exemplary embodiment of this application; Figure 7 This is a schematic diagram comparing feature heatmaps under different network models, as illustrated in an exemplary embodiment of this application. Figure 8 This is a comparison chart showing the variation of download frequency estimation error with signal-to-noise ratio for different feature extraction methods illustrated in an exemplary embodiment of this application; Figure 9 This is a comparison chart showing the bandwidth estimation error as a function of signal-to-noise ratio under different feature extraction methods according to an exemplary embodiment of this application; Figure 10 This is a comparison chart showing the variation of download frequency estimation error with signal-to-noise ratio for different network models, as illustrated in an exemplary embodiment of this application. Figure 11 This is a comparison chart showing the bandwidth estimation error as a function of signal-to-noise ratio under different network models, as illustrated in an exemplary embodiment of this application. Figure 12 This is a structural block diagram of an electronic device illustrated in an exemplary embodiment of this application; Figure 13 This is a block diagram of a target signal detection device based on internal and external interaction features, as illustrated in an exemplary embodiment of this application. Detailed Implementation

[0015] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0016] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0017] like Figure 1 As shown, the technical solution of this application takes ACARS signals as the research object and improves the estimation accuracy of key parameters such as carrier frequency and bandwidth in an open and complex electromagnetic environment, so as to effectively separate the complete target signal.

[0018] Traditional parameter estimation methods often suffer from problems such as blurred spectral boundaries, weakened key structures, and unstable estimation under conditions of low signal-to-noise ratio, frequency offset drift, and adjacent channel interference. This application first preprocesses the ACARS signal using an AM (Amplitude Modulation)-MSK (Minimum Shift Keying) composite modulation scheme. The external modulation information mainly reflects the overall envelope characteristics and energy distribution of the signal, while the internal modulation information reflects the fine-grained changes of the minimum shift keying structure in the frequency and phase domains. Furthermore, by performing logarithmic normalization and point-by-point interactive operations on the two types of time-frequency spectra, a grayscale image of the internal and external interactive features is generated. This explicitly enhances the consistency between the overall distribution of the external modulation and the detailed structure of the internal modulation in the time-frequency domain, suppressing the influence of noise fluctuations, isolated spurious peaks, and local interference on parameter characterization, providing a more stable and discriminative input representation for subsequent detection-based estimation.

[0019] Regarding the target detection model, this application constructs an improved ODSC_YOLOv13 network. Based on the YOLOv13 target detection framework, this network transforms the traditional parameter estimation problem into a target localization and regression problem in the time-frequency plane, and achieves structured inversion of parameters such as carrier frequency and bandwidth through the spatial location and geometric scale of the detection box.

[0020] To enhance feature extraction and target perception capabilities in complex backgrounds, this application introduces an Omni-Dimensional Dynamic Convolution (ODConv) module into the backbone network and detection head. This allows the convolution kernel to adaptively adjust according to the input content, thereby improving the feature representation capability of weakly structured regions, such as blurred boundaries. Furthermore, in the multi-scale feature fusion stage, a Spatial and Channel Synergistic Attention (SCSA) module is introduced. The SCSA attention mechanism achieves cross-layer feature enhancement and key region recalibration, highlighting effective time-frequency structures related to parameter estimation while suppressing background noise and irrelevant responses.

[0021] Thus, the technical solution of this application can maintain high parameter estimation accuracy and stability under non-ideal conditions such as low signal-to-noise ratio, strong interference and frequency offset disturbance, so as to effectively separate the complete target signal and provide effective technical support for intelligent monitoring and safety perception of civil aviation communication signals.

[0022] The embodiments described in this specification will now be described in detail.

[0023] This application provides a target signal detection method based on internal and external interaction features. Figure 2 This is a flowchart illustrating an exemplary embodiment of a target signal detection method based on internal and external interaction features, as shown in this application. Figure 2 As shown, the target signal detection method includes the following steps: Step S1: Modulate the acquired ACARS signal to obtain the inner modulation signal and the outer modulation signal of the ACARS signal; Step S2: Perform time-frequency domain conversion on the internal modulation signal and the external modulation signal to obtain the internal modulation time spectrum and the external modulation time spectrum, and perform point-by-point multiplication operation on the amplitude spectrum of the internal modulation time spectrum and the amplitude spectrum of the external modulation time spectrum to obtain the internal and external interaction feature map corresponding to the ACARS signal. Step S3: The internal and external interaction feature maps are processed by a pre-built target detection model to obtain the target detection result. The target detection model includes a backbone network, a feature fusion network, and a detection head. The backbone network introduces an ODConv module to enhance feature modeling of weakly structured regions. The feature fusion network introduces an SCSA attention module to enhance the effective region response related to signal carrier frequency and bandwidth estimation. The detection head is used to output the bounding box coordinates of the target rectangle. Step S4: Map the bounding box coordinates of the target rectangle based on the frequency axis calibration relationship to obtain the carrier frequency estimate and bandwidth estimate of the ACARS signal, and perform frequency shift correction and bandpass filtering on the ACARS signal based on the carrier frequency estimate and bandwidth estimate to separate the target signal.

[0024] In some embodiments, the internal modulation signal in step S1 is a minimum shift keying (MSK) signal; the external modulation signal is an amplitude modulation (AM) signal.

[0025] MSK is a modulation method that achieves signal orthogonality under minimum modulation index conditions, and is a special form of Continuous Phase Frequency Shift Keying (CPFSK). In minimum frequency shift keying modulation, the carrier phase between adjacent symbols changes continuously. Specifically, each symbol causes a phase shift of ±90° in the signal, but since the modulation index of MSK is 0.5, the equivalent result is that the modulated signal maintains strict phase continuity between two adjacent symbols, and the phase accumulation is equivalent to a difference of 180°. Therefore, the signal phase at a certain moment is determined not only by the value of the current symbol, but also by the symbol polarity of the previous symbol and its accumulated phase state.

[0026] From a waveform structure perspective, MSK belongs to a form of Continuous Phase Modulation (CPM). Because its pulse shaping function is a half-cosine pulse, the instantaneous frequency of each symbol affects the phase within two symbol periods. This means that the baseband waveform corresponding to the current symbol contains not only information about the current symbol, but the phase is cumulative throughout the modulation process; therefore, previous or even earlier symbols contribute to the current phase. Due to the phase memory property of MSK, the receiver cannot make the optimal decision based solely on observations within one symbol period corresponding to the current symbol. Instead, it needs to consider the current symbol and several preceding symbols together to reconstruct a continuous phase trajectory. By extending the observation time and combining it with the correlation between phases, the reliability of symbol decision can be effectively improved, and the probability of bit errors can be reduced.

[0027] Assume the quadrature modulation formula for the MSK signal is as follows: (1) In equation (1), A is the signal amplitude. For carrier phase accumulation term, The width of the symbol. This indicates the current nth input symbol. This is the cumulative phase constant.

[0028] , The value is Therefore, formula (1) can be equivalent to: (2) Furthermore, due to ,therefore, Therefore, the expression for the MSK signal can be derived as follows: (3) in, and These are the waveforms of the I and Q baseband signals, respectively.

[0029] The representation of ACARS signals can be obtained based on the modulation principle of ACARS signals: (4) Formula (4) can be expanded into the following formula (5): (5) In formulas (4) and (5), To meet DC bias, For the MSK signal amplitude, For MSK carrier angular frequency, This is the preset frequency offset. This refers to the VHF radio frequency carrier angular frequency.

[0030] The composite signal after AM adjustment is: (6) In some embodiments, the VHF-band civil aviation ACARS signal received by the antenna is down-converted, quadrature demodulated, and analog-to-digital converted by a ground receiving station to obtain complex baseband I / Q data (i.e., in-phase and quadrature component data) in zero-IF or low-IF form. This I / Q data completely preserves the amplitude and phase information of the original ACARS signal.

[0031] In this embodiment, the acquired ACARS signal is adjusted and decomposed using the above formula (3) to obtain the inner modulation signal, and the outer modulation signal is obtained using the above formula (6).

[0032] In other embodiments, to expand the training dataset or verify the algorithm's performance under ideal conditions, the MSK-modulated signal can also be generated through simulation. Specifically, a random binary sequence of 0s and 1s is generated, and after serial-to-parallel conversion, two orthogonal signals are obtained, where the in-phase branch and the quadrature branch are offset in time by a period T. Then, the waveforms of the in-phase branch and the quadrature branch are respectively... and Weighted summation yields the I and Q baseband signals; these are then compared with the carrier wave. and Multiply the two results; finally, add the two results together to obtain the modulated MSK signal, which serves as the inner modulation signal. The outer modulation signal can be generated by amplitude modulation of the MSK signal using single-tone or multi-tone signals to simulate the AM-MSK composite modulation characteristics of actual ACARS signals.

[0033] In some embodiments, step S2 generates an internal and external interaction feature map through the following steps: Logarithmic domain transformations are performed on the amplitude spectra of the internal modulation time spectrum and the amplitude spectra of the external modulation time spectrum, respectively, to obtain the logarithmic domain transformation results of the internal modulation amplitude spectrum and the external modulation amplitude spectrum. The logarithmic domain transformation results of the inner modulation amplitude spectrum and the logarithmic domain transformation results of the outer modulation amplitude spectrum are normalized respectively to obtain the normalized inner modulation amplitude spectrum and the normalized outer modulation amplitude spectrum. The normalized internal modulation amplitude spectrum and the normalized external modulation amplitude spectrum are multiplied point by point to obtain the internal and external interaction feature map.

[0034] To address the problem that traditional single time-frequency representations in ACARS signal parameter estimation cannot simultaneously reflect both external modulation envelope information and internal modulation structure information, this embodiment proposes an interactive feature representation method based on the formation mechanism of AM-MSK composite modulation signals. The core idea of ​​this embodiment is to enhance the consistency of the two by extracting the overall time-frequency distribution of external modulation and the fine-grained structural information of internal modulation through time-frequency domain interactive operations. This highlights the effective structural regions related to parameters such as carrier frequency and bandwidth, and suppresses the influence of noise, spurious peaks, and local interference on parameter representation.

[0035] Specifically, to characterize the time-frequency distribution characteristics of ACARS signals, this embodiment performs Short-Time Fourier Transform (STFT) on both the internal and external modulation signals. Assume the analysis window function is... Then the spectrum during internal modulation With external modulation time spectrum The expressions are as follows: (7) (8) in, For integration variables, The analysis window is centered at t. For frequency variables, These are Fourier basis functions.

[0036] The window function selection in this embodiment A Hamming window with a length of 256 and an overlap length of 255 was used to obtain higher time-frequency sampling density and more continuous time-frequency evolution trajectory.

[0037] The spectrum when the internal modulation is obtained With external modulation time spectrum Then, the corresponding amplitude spectrum can be obtained: , (9) In formula (9), The amplitude spectrum of the spectrum during internal modulation. This is the amplitude spectrum of the externally modulated spectrum.

[0038] Considering the large dynamic range of the time spectrum, in order to enhance the visibility of the weak structure region and compress the amplitude span, logarithmic domain transformations are performed on both: (10) (11) In formulas (10) and (11), This is the result of the logarithmic domain transformation of the internal modulation amplitude spectrum. This is the result of the logarithmic domain transformation of the external modulation amplitude spectrum. It is a very small positive number, used to avoid singularities in logarithmic operations when the amplitude is zero.

[0039] because Emphasizing the overall energy distribution of external modulation, while This approach emphasizes the internal modulation frequency and bit structure, which are significantly complementary in parameter estimation. To enhance the consistency of the internal and external modulation representations, this embodiment first normalizes the two types of time-spectrum data: (12) (13) In formulas (12) and (13), For the normalized internal modulation amplitude spectrum, For the normalized external modulation amplitude spectrum, , These represent the maximum and minimum values ​​in the logarithmic domain transform of the external modulation amplitude spectrum. , These are the maximum and minimum values ​​in the logarithmic domain transformation result of the internal modulation amplitude spectrum.

[0040] Therefore, the internal and external interaction feature map of this embodiment This can be expressed by the following formula: (14) In formula (14), the pointwise multiplication operation ensures that only regions with strong responses in both the inner and outer modulation spectra are further enhanced in the interactive feature map; while noise floor undulations, isolated spurious peaks, and local interferences appearing only in the unilateral spectra are significantly weakened in the interactive map due to the lack of information support from the other side. Therefore, the inner and outer interactive feature maps... It can better highlight the stable structural regions related to carrier frequency location and occupied bandwidth, thereby improving the separability and robustness of parameter estimation.

[0041] To adapt to the input of the target detection network, this embodiment also includes internal and external interaction feature maps. By uniformly scaling the image size and converting it to grayscale, compared to pseudo-color representation, grayscale processing can preserve the time-frequency intensity distribution, texture boundaries, and geometric structure information while avoiding additional visual redundancy and irrelevant perturbations introduced by color mapping. This allows the network to focus more on structural features directly related to parameter estimation. Simultaneously, grayscale representation reduces the input dimensionality and model computational overhead, which is beneficial for improving training stability and engineering deployment efficiency, ultimately generating an internally and externally modulated interactive grayscale image.

[0042] Figure 3 Examples of STFT time-frequency diagrams, external modulation grayscale diagrams, internal modulation grayscale diagrams, and interactive grayscale diagrams are provided, based on... Figure 3 As can be seen from the various sample images shown, the internal and external interaction feature maps can better highlight the stable structural regions related to carrier frequency location and occupied bandwidth.

[0043] In some embodiments, step S3 of this embodiment detects the internal and external interaction feature maps using a pre-built target detection model. The specific detection process is as follows: The target detection model in this embodiment is the improved YOLOv13 network, abbreviated as ODSC_YOLOv13 network model.

[0044] like Figure 4 As shown, the network model consists of three parts: a backbone network, a feature fusion network (Neck), and a detection head (Head). This embodiment uses the constructed internal and external interaction feature grayscale image as model input to achieve detection-based estimation of the carrier frequency and bandwidth of ACARS signals.

[0045] First, the backbone network progressively extracts and downsamples features through convolution, depthwise separable convolution, and DS-C3k2 modules, forming three main feature layers (B3, B4, and B5) to characterize time-frequency structural information at different scales. This embodiment introduces the ODConv module into the backbone network, using an input-driven dynamic convolution kernel aggregation mechanism to enhance the network's adaptive modeling of weakly structured regions, thereby improving the extraction accuracy of parameter-related features.

[0046] The input processing procedure of the ODConv module in this embodiment can be referred to Figure 5 As shown, the feature extraction process can be described in detail in this embodiment, and those skilled in the art can refer to relevant solutions.

[0047] In the feature fusion stage, the network takes the three backbone features B3, B4, and B5 as input and performs cross-scale correlation enhancement through the HyperACE+SCSA module to generate enhanced features H3, H4, and H5. The HyperACE module is responsible for modeling higher-order correlations between multi-scale features, while the SCSA module further performs spatial channel co-calibration on the fused features, strengthening the effective region response directly related to signal carrier frequency and bandwidth estimation, and suppressing background noise and irrelevant interference features.

[0048] The input processing procedure of the SCSA module in this embodiment can be referred to Figure 6 As shown, the attention fusion process can be described in detail in this embodiment, and those skilled in the art can refer to relevant solutions.

[0049] Subsequently, the network adaptively distributes enhanced features to branches of different scales through FullPAD Tunnel, and combines upsampling, splicing and DS-C3k2 modules to complete bidirectional feature fusion from top to bottom and bottom to top; at the same time, the ODConv module is introduced again in the detection head to enhance the structural modeling ability of mid-to-high-level features during scale backpropagation.

[0050] Finally, the features extracted by each level of the network are standardized through a feature preprocessing module. Non-maximum suppression is then applied, and the outputs from the three detectors are regressed through a fully connected layer to obtain the bounding box coordinates (x, y) of the signal rectangular region. min y min (x) max y max ).

[0051] After obtaining the bounding box coordinates of the signal rectangular region, the center and width of the rectangle can be further obtained, and the carrier frequency estimate can be calculated based on the frequency axis calibration relationship. Compared with bandwidth estimates .

[0052] The ODSC_YOLOv13 network in this embodiment effectively improves the robustness and accuracy of ACARS signal parameter estimation in complex electromagnetic environments through the collaborative design of the two key modules, ODConv and SCSA.

[0053] Obtaining carrier frequency estimates Compared with bandwidth estimates Subsequently, these two parameters drive the subsequent frequency shift correction and bandpass filtering processes to separate the target ACARS signal from the original received signal. Frequency shift correction down-converts the signal to near zero frequency, compensating for frequency drift during actual reception. After this processing, the center frequency of the target signal is shifted to the baseband, facilitating subsequent filtering. Then, bandpass filtering is applied, setting the passband range based on the estimated bandwidth. Since the target signal has been down-converted to zero frequency, the filter's passband is set using either a finite impulse response (FIR) or infinite impulse response (IOR) filter. Its passband cutoff frequency, stopband attenuation, and other parameters are adaptively determined based on the estimated bandwidth. This bandpass filtering effectively suppresses adjacent channel interference and out-of-band noise propagation, achieving accurate extraction of the target signal components.

[0054] Since the parameters for frequency shift correction and bandpass filtering are determined by the estimated values ​​output by the detection network, they can be adaptively adjusted according to the actual time-frequency distribution of the signal, thus effectively avoiding the over-truncation and under-truncation problems easily introduced by traditional fixed window length or manual threshold methods. Subsequent experiments show that this embodiment can maintain high signal separation integrity and consistency under different signal-to-noise ratios and complex electromagnetic environments, providing reliable signal input for subsequent message parsing, situational awareness, and device tracing.

[0055] In some embodiments, Figure 2 The target signal detection method shown also includes the following steps: Step S5: The separated target signal, the carrier frequency estimate, and the bandwidth estimate are associated and stored in the database for visualization.

[0056] This embodiment executes a target signal detection method through a signal processing system. When the system is running online, the continuously acquired I / Q data is divided into frames according to time windows to construct an internal and external interactive feature map. The detection network is called to complete the target box localization and estimate the carrier frequency and bandwidth. Then, based on the estimated parameters, frequency shifting and bandpass filtering are performed to separate the target signal. Finally, the identification results, key parameters and associated screenshots are written into the database and displayed on the front-end interface, providing key basis for subsequent situation analysis, equipment tracing and anomaly alarm.

[0057] Next, the performance of the target signal detection scheme of this application will be verified through experiments.

[0058] (1) The dataset used in this embodiment.

[0059] The communication signal samples used in this embodiment are ACARS signals from civil aircraft collected using a satellite antenna. A total of 10,000 signal samples were collected, with 8,000 used as the training set, 1,000 as the validation set, and 1,000 as the test set. Specific parameter settings were: carrier frequency of 2~6.5kHz and sampling rate of 16kHz. To compare the estimation performance of the algorithm under different signal-to-noise ratios (SNRs), additive white Gaussian noise was added to simulate noise environments at different SNRs using simulation software. The dataset SNR range was {-20dB, -15dB, -10dB, -5dB, 0dB}.

[0060] (2) Experimental environment and network configuration in this embodiment.

[0061] The GPU used in this embodiment is an NVIDIA GeForce RTX 3090, the neural network training framework is PyTorch, and the input image resolution is 640×640, including 224×224 STFT time-frequency plots, internal modulation grayscale plots, external modulation grayscale plots, and internal-external interactive grayscale plots. Figure 3 Four types of images were used. The learning rate was fixed at 2e-4, the batch size was 16, and the training was performed for 200 rounds. The ODSC_YOLOv13 network model was used to train and test the carrier frequency and bandwidth estimation of these four types of images.

[0062] (3) Performance simulation.

[0063] 3.1 Visual comparison of feature maps of different network models.

[0064] To further analyze the differences in the focus of different detection networks on key time-frequency structures in the civil aviation ACARS signal parameter estimation task, this paper visualizes and compares the feature heatmaps of ODSC_YOLOv13, YOLOv13, YOLOv12, YOLOv11, and YOLOv8 on the same internal and external interactive grayscale image sample. The results are as follows: Figure 7 As shown.

[0065] from Figure 7It can be seen that the high-response region of ODSC_YOLOv13 is mainly concentrated in the main energy band of the signal and its boundary, with a continuous and sufficient overall response distribution, while the background region maintains a low activation level. This indicates that the model of this embodiment can relatively stably focus on the effective structure directly related to carrier frequency localization and bandwidth measurement. The reason for this is that this embodiment introduces two key modules, ODConv and SCSA, into the network: the former enhances the adaptive modeling capability for weakly structured regions, while the latter further strengthens the response of parameter-related regions and suppresses irrelevant background interference through spatial channel collaborative recalibration. Therefore, the ODSC_YOLOv13 network model of this embodiment can not only more accurately depict the overall outline of the signal rectangular region, but also more effectively focus on boundary transitions and local detail changes, thereby providing more reliable feature support for subsequent detection box localization and geometric inversion.

[0066] In comparison, other comparison networks each have their shortcomings. While YOLOv13 can generate a certain response near the main energy band of the signal, its high-response region still exhibits local dispersion and offset phenomena, and its coverage of frequency band boundaries and weak texture regions is insufficient. YOLOv12's thermal response shows obvious dispersion and noise characteristics, with many irrelevant activations in the background region, indicating that its ability to distinguish between the effective target region and background interference is relatively insufficient. Although YOLOv11 can form local hotspots in some main energy regions, its overall response range is small and its focusing is not stable enough, and its representation of the overall structure of the rectangular frequency band is still insufficient. YOLOv8, on the other hand, shows more local isolated hotspots and edge region activations, with relatively limited attention to the continuous structure inside the main energy band, and strong activations in the background region, indicating that its ability to model the overall time-frequency morphology related to parameters is still insufficient.

[0067] In summary, the thermal distribution of the ODSC_YOLOv13 network model in this embodiment shows higher consistency with the region related to the actual signal parameters, indicating that the proposed model has stronger feature selection ability, boundary awareness ability and robust parameter estimation ability in complex electromagnetic environments.

[0068] 3.2 Impact of different feature representations on the estimation performance of key parameters of ACARS signals To evaluate the impact of different feature representations on the performance of ACARS signal key parameter estimation, this paper conducts comparative experiments using internal and external interactive grayscale images, internal modulation grayscale images, external modulation grayscale images, and STFT time-frequency images as inputs to the ODSC_YOLOv13 network under the same training strategy and network configuration. The average relative error between carrier frequency and bandwidth is statistically analyzed under different signal-to-noise ratios to verify the improvement effect of the constructed features on the accuracy and robustness of parameter estimation. Figure 8 , Figure 9 As shown.

[0069] As shown in the figure. Figure 8 It can be seen that, under the same experimental settings, as the SNR increases from -20dB to 0dB, the carrier frequency estimation error of each feature extraction method generally decreases. However, the internal-external interactive grayscale image consistently maintains the lowest error across the entire SNR range, and its advantage is more significant under low signal-to-noise ratio (SNR) conditions. Under a SNR of -20dB, the carrier frequency estimation error of the internal-external interactive grayscale image is reduced by 13.85%, 14.35%, and 15.85% compared to the internal modulation grayscale image, external modulation grayscale image, and STFT time-frequency image methods, respectively. This indicates that the internal-external interactive representation can more stably highlight the carrier frequency-related structure in a strong noise background, thereby significantly improving the carrier frequency localization accuracy.

[0070] Depend on Figure 9 As can be seen, the bandwidth estimation follows a consistent pattern. With increasing SNR, the bandwidth estimation error of each method gradually decreases, but the internal-external interactive grayscale image consistently maintains the smallest error and exhibits stronger robustness. Under a signal-to-noise ratio of -10dB, the bandwidth estimation error of the internal-external interactive grayscale image is reduced by 8.07%, 9.72%, and 10.55% compared to the internal modulation grayscale image, external modulation grayscale image, and STFT time-frequency image methods, respectively. This result indicates that the internal-external interactive feature can effectively enhance the saliency of the target rectangular frequency band boundary, reduce the impact of boundary ambiguity and local breaks on bandwidth measurement, thereby improving the stability of bandwidth estimation.

[0071] 3.3 Comparison of parameter estimation performance of different network models To evaluate the effectiveness of the ODSC_YOLOv13 network model in the civil aviation ACARS signal parameter estimation task, it was compared with mainstream one-stage detection networks such as YOLOv13, YOLOv12, YOLOv11, and YOLOv8. Under the same training set, test set, input features, and training strategy, the average percentage of carrier frequency error and the average percentage of bandwidth error under different signal-to-noise ratios were statistically analyzed to quantitatively evaluate the parameter estimation performance of each model. The results are as follows: Figure 10 , Figure 11 As shown.

[0072] (1) Comparison of carrier frequency estimation performance: Depend on Figure 10As can be seen, the carrier frequency estimation error of each model decreases as the signal-to-noise ratio (SNR) increases from -20dB to 0dB. Among them, the ODSC_YOLOv13 network model in this embodiment maintains the lowest error throughout the entire SNR range, demonstrating stronger robustness and stability. Taking a -15dB SNR as an example, the carrier frequency estimation error of the ODSC_YOLOv13 network model is reduced by 5.51%, 6.42%, 15.66%, and 17.84% compared to the YOLOv13, YOLOv12, YOLOv11, and YOLOv8 network models, respectively. This result indicates that the network model in this embodiment has a more reliable ability to locate carrier frequency-related structures under strong noise conditions.

[0073] (2) Bandwidth estimation performance comparison: Depend on Figure 11 As can be seen, the bandwidth estimation performance exhibits a consistent pattern. The ODSC_YOLOv13 network model in this embodiment achieves the minimum bandwidth error under all signal-to-noise ratio conditions. Taking a -10dB signal-to-noise ratio as an example, its bandwidth estimation error is reduced by 4.78%, 5.14%, 11.15%, and 11.49% compared to the YOLOv13, YOLOv12, YOLOv11, and YOLOv8 network models, respectively. This indicates that even under conditions of blurred boundaries and decreased saliency of weak structures, the network model in this embodiment can still maintain relatively stable bandwidth fitting ability.

[0074] The comparative experiments above demonstrate that, under the same data and training configuration, the ODSC_YOLOv13 network model of this application significantly outperforms the comparative model in estimating both carrier frequency and bandwidth, especially exhibiting stronger noise resistance and robust estimation performance under low signal-to-noise ratio conditions. The experimental results verify the effectiveness and engineering applicability of the network model of this application for estimating civil aviation signal parameters in complex electromagnetic environments.

[0075] In summary, the embodiments of this application have the characteristics of high recognition rate and strong robustness, and still have the advantage of high recognition rate under low signal-to-noise ratio.

[0076] Figure 12 This is a schematic diagram of an electronic device illustrated in this specification according to an exemplary embodiment. Please refer to... Figure 12At the hardware level, the device includes a processor 102, an internal bus 104, a network interface 106, memory 108, a hardware acceleration device 110, and non-volatile memory 112, and may also include other hardware required for its functions. One or more embodiments of this application can be implemented in software, for example, the processor 102 reads the corresponding computer program from the non-volatile memory 112 into memory 108 and then runs it. Of course, in addition to software implementation, one or more embodiments of this application do not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the above processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0077] Figure 13 This is a structural block diagram of a target signal detection device based on internal and external interaction features, as illustrated in an exemplary embodiment of this application. The target signal detection device can be applied to, for example... Figure 12 The electronic device shown implements the technical solution of this application. The target signal detection device includes: a signal adjustment unit 210, a cross-feature calculation unit 220, a target detection unit 230, and a signal processing unit 240, wherein: The signal adjustment unit 210 is used to modulate the acquired ACARS signal to obtain the inner modulation signal and the outer modulation signal of the ACARS signal. The mutual feature calculation unit 220 is used to perform time-frequency domain conversion on the inner modulation signal and the outer modulation signal to obtain the inner modulation time spectrum and the outer modulation time spectrum, and to perform point-by-point multiplication operation on the amplitude spectrum of the inner modulation time spectrum and the amplitude spectrum of the outer modulation time spectrum to obtain the inner and outer interaction feature map corresponding to the ACARS signal. The target detection unit 230 is used to process the internal and external interaction feature map through a pre-built target detection model to obtain the target detection result. The target detection model includes a backbone network, a feature fusion network, and a detection head. The backbone network introduces an ODConv module to enhance feature modeling of weakly structured regions. The feature fusion network introduces an SCSA attention module to enhance the effective region response related to signal carrier frequency and bandwidth estimation. The detection head is used to output the bounding box coordinates of the target rectangle. The signal processing unit 240 is used to map the bounding box coordinates of the target rectangle based on the frequency axis calibration relationship to obtain the carrier frequency estimate and bandwidth estimate of the ACARS signal, and to perform frequency shift correction and bandpass filtering on the ACARS signal based on the carrier frequency estimate and bandwidth estimate to separate the target signal.

[0078] In some embodiments, the mutual feature calculation unit 220 is configured to perform logarithmic domain transformation on the amplitude spectrum of the internal modulation time spectrum and the amplitude spectrum of the external modulation time spectrum, respectively, to obtain the logarithmic domain transformation result of the internal modulation amplitude spectrum and the logarithmic domain transformation result of the external modulation amplitude spectrum; to normalize the logarithmic domain transformation result of the internal modulation amplitude spectrum and the logarithmic domain transformation result of the external modulation amplitude spectrum, respectively, to obtain the normalized internal modulation amplitude spectrum and the normalized external modulation amplitude spectrum; and to perform point-by-point multiplication operation on the normalized internal modulation amplitude spectrum and the normalized external modulation amplitude spectrum to obtain the internal and external interaction feature map.

[0079] In some embodiments, the target signal detection device further includes a post-processing unit for associating and storing the separated target signal, the carrier frequency estimate, and the bandwidth estimate in a database for visualization.

[0080] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0081] Accordingly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the methods described in any of the above embodiments.

[0082] Accordingly, embodiments of this application also provide a computer program product configured to perform the methods described in any of the above embodiments.

[0083] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which can take the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email sending and receiving device, game console, tablet computer, wearable device, or any combination of these devices.

[0084] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0085] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0086] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random-access memory (SRAM), dynamic random-access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0087] While this specification contains numerous specific implementation details, these should not be construed as limiting the scope of any invention or the scope of the claims, but rather are primarily intended to describe features of specific embodiments of a particular invention. Certain features described in the various embodiments herein may also be implemented in combination in a single embodiment. Conversely, various features described in a single embodiment may also be implemented separately in various embodiments or in any suitable sub-combination. Furthermore, while features may function in certain combinations as described above and even initially claimed in this way, one or more features from a claimed combination may be removed from that combination in some cases, and a claimed combination may refer to a sub-combination or a variation thereof.

[0088] Similarly, although the operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system modules and components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0089] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims may be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings are not necessarily shown in a specific order or sequence to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0090] It should be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0091] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A target signal detection method based on internal and external interaction features, characterized in that, Includes the following steps: Step S1: Modulate the acquired ACARS signal to obtain the inner modulation signal and the outer modulation signal of the ACARS signal; Step S2: Perform time-frequency domain conversion on the internal modulation signal and the external modulation signal to obtain the internal modulation time spectrum and the external modulation time spectrum, and perform point-by-point multiplication operation on the amplitude spectrum of the internal modulation time spectrum and the amplitude spectrum of the external modulation time spectrum to obtain the internal and external interaction feature map corresponding to the ACARS signal. Step S3: The internal and external interaction feature maps are processed by a pre-built target detection model to obtain the target detection result. The target detection model includes a backbone network, a feature fusion network, and a detection head. The backbone network introduces an ODConv module to enhance feature modeling of weakly structured regions. The feature fusion network introduces an SCSA attention module to enhance the effective region response related to signal carrier frequency and bandwidth estimation. The detection head is used to output the bounding box coordinates of the target rectangle. Step S4: Map the bounding box coordinates of the target rectangle based on the frequency axis calibration relationship to obtain the carrier frequency estimate and bandwidth estimate of the ACARS signal, and perform frequency shift correction and bandpass filtering on the ACARS signal based on the carrier frequency estimate and bandwidth estimate to separate the target signal.

2. The method according to claim 1, characterized in that, Step S2 generates an internal and external interaction feature map through the following steps: Logarithmic domain transformations are performed on the amplitude spectra of the internal modulation time spectrum and the amplitude spectra of the external modulation time spectrum, respectively, to obtain the logarithmic domain transformation results of the internal modulation amplitude spectrum and the external modulation amplitude spectrum. The logarithmic domain transformation results of the inner modulation amplitude spectrum and the logarithmic domain transformation results of the outer modulation amplitude spectrum are normalized respectively to obtain the normalized inner modulation amplitude spectrum and the normalized outer modulation amplitude spectrum. The normalized internal modulation amplitude spectrum and the normalized external modulation amplitude spectrum are multiplied point by point to obtain the internal and external interaction feature map.

3. The method according to claim 2, characterized in that, The expression for the amplitude spectrum after frequency normalization during internal modulation is: ； The expression for the amplitude spectrum after the frequency spectrum normalization process during external modulation is as follows: ； in, For the normalized internal modulation amplitude spectrum, This is the result of the logarithmic domain transformation of the internal modulation amplitude spectrum. , These represent the maximum and minimum values ​​in the logarithmic domain transform of the internal modulation amplitude spectrum. For the normalized external modulation amplitude spectrum, This is the result of the logarithmic domain transformation of the external modulation amplitude spectrum. , These are the maximum and minimum values ​​in the logarithmic domain transformation result of the external modulation amplitude spectrum.

4. The method according to claim 1, characterized in that: The internal modulation signal is a minimum frequency shift keying signal; The external modulation signal is an amplitude modulation signal.

5. The method according to claim 1, characterized in that, The target detection model has a multi-scale detection head, and the target detection model is configured to regress the output features of each scale detection head through a fully connected layer to obtain the bounding box coordinates of the target rectangle.

6. The method according to any one of claims 1 to 5, characterized in that, It also includes the following steps: Step S5: The separated target signal, the carrier frequency estimate, and the bandwidth estimate are associated and stored in the database for visualization.

7. A target signal detection device based on internal and external interaction features, characterized in that, The signal adjustment unit is used to modulate the acquired ACARS signal to obtain the internal modulation signal and the external modulation signal of the ACARS signal. The mutual feature calculation unit is used to perform time-frequency domain conversion on the inner modulation signal and the outer modulation signal to obtain the inner modulation time spectrum and the outer modulation time spectrum, and to perform point-by-point multiplication operation on the amplitude spectrum of the inner modulation time spectrum and the amplitude spectrum of the outer modulation time spectrum to obtain the inner and outer interaction feature map corresponding to the ACARS signal. The target detection unit is used to process the internal and external interaction feature maps through a pre-built target detection model to obtain target detection results. The target detection model includes a backbone network, a feature fusion network, and a detection head. The backbone network introduces an ODConv module to enhance feature modeling of weakly structured regions. The feature fusion network introduces an SCSA attention module to enhance the effective region response related to signal carrier frequency and bandwidth estimation. The detection head is used to output the bounding box coordinates of the target rectangle. The signal processing unit is used to map the bounding box coordinates of the target rectangle based on the frequency axis calibration relationship to obtain the carrier frequency estimate and bandwidth estimate of the ACARS signal, and to perform frequency shift correction and bandpass filtering on the ACARS signal based on the carrier frequency estimate and bandwidth estimate to separate the target signal.

8. An electronic device, characterized in that, include: processor; as well as A computer-readable storage medium storing computer program instructions that, when executed by the processor, cause the processor to perform the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which is executed by a processor according to any one of claims 1 to 6.

Images