Anti-interference method and device for high-frequency coaxial cable

Abstract

The invention relates to the technical field of cable anti-interference, and discloses a high-frequency coaxial cable anti-interference method and device.The method comprises the steps that a radial progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer is constructed, and a triangular supporting frame structure is arranged between the inner shielding layer and the outer shielding layer; an electromagnetic induction probe is installed on the triangular supporting frame structure, and differential interference signals between the inner shielding layer and the outer shielding layer are collected; performing time-frequency domain conjoint analysis on the differential interference signal to obtain an interference source feature mapping map; and performing multi-mode interference identification and suppression according to the interference source characteristic mapping map to generate an interference offset signal, and performing real-time elimination and feedback on electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference offset signal to obtain an interference offset feedback effect. And signal reflection and crosstalk among the structures of all layers in the cable are effectively compensated.

Description

Technical Field

[0001] The present invention relates to the technical field of cable anti-interference, and particularly to a method and device for anti-interference of high-frequency coaxial cables. Background Art

[0002] Traditional coaxial cables adopt a single-material shielding layer structure. This shielding method shows obvious deficiencies in high-frequency application scenarios of 0.1 - 18 GHz. Especially when the cable needs to transmit high-frequency signals and charging currents simultaneously, a single shielding layer cannot meet the differential shielding requirements of signals in different frequency bands.

[0003] In the prior art, cable anti-interference mainly relies on increasing the shielding layer thickness or using high-cost materials, which not only increases the cable diameter and weight but also significantly raises the manufacturing cost. In addition, physical shielding and signal processing technologies are disjointed, lacking a unified collaborative anti-interference solution. When the cable is affected by external high-frequency interference, it cannot adjust the shielding performance in real time, resulting in a serious decline in signal integrity, especially obvious in application scenarios where charging and data transmission are carried out simultaneously. Summary of the Invention

[0004] The present invention provides a method and device for anti-interference of high-frequency coaxial cables, which realizes efficient shielding in the full frequency band and effectively compensates for signal reflection and crosstalk between internal layers of the cable.

[0005] In a first aspect, the present invention provides a method for anti-interference of high-frequency coaxial cables, and the method for anti-interference of high-frequency coaxial cables includes: Construct a radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and set a triangular support frame structure between the inner shielding layer and the outer shielding layer; Install an electromagnetic induction probe on the triangular support frame structure to collect the differential interference signal between the inner shielding layer and the outer shielding layer; Perform joint time-frequency domain analysis on the differential interference signal to obtain an interference source feature mapping spectrum; Execute multi-modal interference recognition and suppression according to the interference source feature mapping spectrum, generate an interference cancellation signal, and perform real-time elimination and feedback of electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference cancellation signal to obtain an interference cancellation feedback effect.

[0006] In a second aspect, the present invention provides a device for anti-interference of high-frequency coaxial cables, and the device for anti-interference of high-frequency coaxial cables includes: A construction module for constructing a radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and setting a triangular support frame structure between the inner shielding layer and the outer shielding layer; A collection module, configured to install an electromagnetic induction probe on the triangular support frame structure and collect the differential interference signal between the inner shielding layer and the outer shielding layer; A joint analysis module, configured to perform joint time-frequency domain analysis on the differential interference signal to obtain an interference source feature mapping spectrum; A feedback module, configured to perform multi-modal interference recognition and suppression according to the interference source feature mapping spectrum, generate an interference cancellation signal, and perform real-time cancellation and feedback on the electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference cancellation signal to obtain an interference cancellation feedback effect.

[0007] The third aspect of the present invention provides a high-frequency coaxial cable anti-interference device, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the high-frequency coaxial cable anti-interference device executes the above-mentioned high-frequency coaxial cable anti-interference method.

[0008] The fourth aspect of the present invention provides a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when the instructions are run on a computer, the computer is made to execute the above-mentioned high-frequency coaxial cable anti-interference method.

[0009] In the technical solution provided by the present invention, the radial progressive impedance gradient double-layer shielding structure design realizes the functional separation of the inner and outer shielding layers. The inner shielding layer focuses on protecting the transmission stability of the signal line, and the outer shielding layer focuses on protecting the charging wire from interference, enabling the cable to significantly improve the anti-interference ability without increasing the overall diameter. The setting of the triangular support frame structure not only improves the compressive strength of the cable but also creates an ideal geometric space between the inner and outer shielding layers, enabling the shielding efficiency to remain stable even when the cable is compressed and deformed, achieving high-efficiency shielding across the entire frequency band. The shielding structure optimization algorithm based on the field strength attenuation coefficient precisely controls the material composition and geometric parameters of the inner and outer shielding layers, ensuring the maximization of shielding efficiency under the given thickness limit and effectively reducing the cable manufacturing cost. The real-time acquisition system for differential signals of the inner and outer shielding layers enables the cable to have the ability to perceive the electromagnetic environment, be able to monitor the characteristics of interference sources in real time, and provide accurate data support for subsequent interference suppression. The interference feature extraction method of joint time-frequency domain analysis realizes the precise identification and positioning of interference sources. Through the collaborative analysis of frequency domain and time domain characteristics, different types and sources of interference signals can be distinguished. The multi-modal adaptive interference identification and elimination algorithm is optimized for different frequency bands and interference characteristics, significantly improving the signal transmission quality without changing the physical structure and effectively compensating for signal reflection and crosstalk between the internal layers of the cable. The double-layer heterogeneous recurrent neural network interference prediction system separates the influence of inner high-frequency interference and outer low-frequency interference through hierarchical decoupling technology, solving the problem of cross-influence of interference sources in traditional methods and performing outstandingly in an environment with complex electromagnetic interference.

[0010] Other features and advantages of the present invention will be described in the following specification, and part of them will be obvious from the specification or understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification, claims, and drawings.

[0011] To make the above objectives, features, and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given in conjunction with the accompanying drawings and described in detail as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic diagram of an embodiment of the anti-interference method for a high-frequency coaxial cable in an embodiment of the present invention; Figure 2 It is a schematic diagram of an embodiment of the anti-interference device for a high-frequency coaxial cable in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0014] As used in the embodiments of the present invention, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes other steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0015] For ease of understanding of this embodiment, first, a high-frequency coaxial cable anti-interference method disclosed in the embodiments of the present invention will be introduced in detail. As Figure 1 shown, this method includes the following steps: 101. Construct a radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and set a triangular support frame structure between the inner shielding layer and the outer shielding layer; It can be understood that the execution subject of the present invention can be a high-frequency coaxial cable anti-interference device, or a terminal or a server. Specifically, no limitation is made here. The embodiments of the present invention will be described by taking the server as the execution subject as an example.

[0016] Specifically, the conductor part of the cable is centered cylindrically to ensure that the subsequent layer structures are coaxially symmetrically distributed, minimizing the transmission characteristic fluctuations caused by structural asymmetry. A high-purity copper core conductor is selected as the main channel for signal transmission, and a polytetrafluoroethylene (PTFE) insulating material is coated on the surface of the copper core conductor with a constant thickness through a mechanical extrusion device. PTFE has excellent high-frequency dielectric properties and extremely low loss factors. After the insulation layer is formed, a high-density grid braiding of silver-copper alloy wires is performed on its outer surface. While ensuring high conductivity, silver-copper alloy has good high-frequency shielding performance. During braiding, the wire diameter, braiding density, and coverage rate are controlled to make the overall impedance and shielding ability of the inner shielding layer meet the predetermined requirements. After the inner shielding layer is formed, to avoid electrical short circuits and further isolate the interference effects between the inner and outer shielding layers, a layer of fluorinated ethylene propylene material is coated on its surface to obtain an insulating isolation layer. The insulating isolation layer is subjected to a tin-copper alloy wire grid braiding treatment. Tin-copper alloy has a lower cost than silver-copper alloy but also has good conductivity and high-frequency anti-interference ability. By adjusting the braiding parameters of the tin-copper alloy wires, the outer shielding layer has shielding characteristics complementary to those of the inner shielding layer. The inner and outer shielding layers form a radially progressive gradient structure in terms of material properties and impedance distribution. This structure helps the incident electromagnetic wave to gradually dissipate during multi-layer attenuation, effectively improving the overall shielding efficiency. Between the inner and outer shielding layers, to achieve precise spatial separation and form an ideal electromagnetic isolation zone, a triangular support frame structure made of reinforced nylon material is introduced. The support frame is precisely manufactured by a mold according to the design requirements to ensure that the geometric dimensions, distribution angles, and spacing distances of each support column are highly consistent. The triangular support frames are arranged equidistantly along the axial direction of the cable, with each group of support columns separated by 120 degrees in the circumferential direction and having an equilateral triangle cross-section, ensuring a fixed spacing between the shielding layers, effectively preventing spatial changes caused by mechanical pressure or bending deformation, and enhancing the overall compressive resistance and structural strength of the cable, ultimately obtaining an electromagnetic isolation space between the inner shielding layer and the outer shielding layer.

[0017] In this embodiment, a mathematical model of the electromagnetic field distribution is established for the overall structure and physical characteristics of the high-frequency coaxial cable. This process takes into account the cylindrical symmetry of the coaxial cable and the electromagnetic properties of each layer of material. By applying classical electromagnetic theory, the cross-section of the cable is regarded as a multi-layer concentric cylinder, and the variation relationship of the electromagnetic field strength E(r) with the radial distance r is derived and expressed as E(r)=E0×e (-αr) , where E0 is the incident field strength, α is the attenuation coefficient, and r is the radial position. Through this model, the laws of propagation, attenuation, and scattering of external electromagnetic interference signals in different shielding layers are described. Based on this mathematical model, a theoretical calculation of the attenuation coefficient is performed for the inner shielding layer. The magnitude of the attenuation coefficient α is related to the conductivity, permeability, and operating frequency of the material, and is specifically calculated by the formula α = 4.7×10 3 ×f 0.5×σ×μPerform quantitative analysis, where σ represents the conductivity of the silver-copper alloy of the inner shielding layer, μ is its relative permeability, and f is the target frequency range, usually covering 0.1 to 18 GHz. Using the same idea, perform an independent calculation of the attenuation coefficient for the outer shielding layer. The parameters are only different in terms of materials. The outer shielding layer mainly uses a tin-copper alloy, and the differences in its conductivity and permeability will directly affect the attenuation ability of the outer layer to electromagnetic fields. After obtaining the attenuation coefficients of the inner and outer shielding layers respectively, calculate the ratio of the attenuation coefficients of the inner and outer layers. This ratio reflects the shielding synergistic effect of the inner and outer shielding layers on electromagnetic interference of different frequencies. During the setting process of the ratio of the attenuation coefficients of the shielding layer, according to the target application scenario and the desired shielding effectiveness, combined with engineering experience and electromagnetic simulation results, set an optimal ratio to ensure that the inner layer has a stronger shielding effect in the high-frequency band, while the outer layer takes into account mechanical protection and interference suppression in a wide frequency band. After setting the ratio of the attenuation coefficients, trace back to the material ratio, and by adjusting the proportions of components such as silver, copper, and tin in the inner and outer shielding layers, optimize their conductivity and permeability to make the attenuation coefficient in the actual structure match the theoretically set value. Input the adjusted material ratios of the inner and outer shielding layers and all relevant geometric parameters (such as shielding layer thickness, braiding density, layer spacing, etc.) into a professional electromagnetic field analysis system, and use simulation software to perform multiple rounds of iterative calculations on the electromagnetic response of the entire shielding structure within the actual working frequency band. After detailed simulation and optimization in the frequency domain and time domain, select the best parameter combination that can continuously maintain a high shielding effectiveness in the 0.1-18 GHz broadband. Output the design parameters of the shielding structure.

[0018] 102. Install an electromagnetic induction probe on the triangular support frame structure to collect the differential interference signal between the inner shielding layer and the outer shielding layer; Specifically, multiple high-sensitivity electromagnetic induction probes are embedded in the top structure of the triangular support frame, with one probe installed on each of the three support columns of each triangular support frame structure, and the three probes are symmetrically distributed at 120 degrees in space to ensure uniform coverage in the circumferential direction. At the same time, each probe is deeply inserted into the gap area between the inner and outer shielding layers to capture the local electromagnetic disturbance changes existing in this space to the greatest extent. Since the differential interference signal is extremely weak and has a high frequency, an isolated micro coaxial lead is used as the signal transmission medium, and each probe is independently connected to an external signal acquisition module to avoid signal crosstalk and ensure the electromagnetic integrity of each channel. The signal acquisition module inputs the analog signals of each channel into a high-performance analog-to-digital converter system, uses an ADC module with a high dynamic range and high sampling accuracy, sets an independent sampling path for each channel, with a sampling rate of 40 MSPS and a bit width of 24 bits to completely retain the amplitude and time characteristics of the high-frequency interference signal, forming multiple parallel digital differential signal data streams. To improve data processing efficiency and suppress noise, digital filtering and decimation processing are performed on each signal channel. The filtering link is implemented through a FIR or IIR digital filter to initially suppress the interference in the non-target frequency band of the spectrum, and the decimation process compresses the signal to a more suitable processing rate, such as 10 MSPS, by down-converting the sampling rate, thereby reducing the processing load while retaining the main interference characteristics, forming multiple preliminarily processed differential signals. The multiple preliminarily processed differential signals are subjected to window slicing and phase correction processing. Each channel signal is sliced in segments in the form of a sliding window with a fixed duration. The common window width is 25 ms, and a 50% overlap rate is set between the windows to avoid loss of boundary information; then, a time-domain phase comparison and compensation method is used to perform phase correction on each signal path. By calculating the delay difference between each channel and adjusting its time axis position, the signals of all channels are synchronized in time. After completing the time alignment, a weighted average processing is performed on all the corrected channel signals, and the weight coefficient is dynamically calculated through the least mean square error adaptive algorithm to ensure that the weighted result maximally suppresses the inconsistent interference and fluctuations between channels, and finally outputs the synthesized differential interference signal.

[0019] 103. Perform a joint time-frequency domain analysis on the differential interference signal to obtain an interference source feature mapping spectrum; Specifically, the differential interference signal is split into two channels, and the same signal is respectively introduced into the frequency-domain analysis channel and the time-domain analysis channel to obtain information on the signal in two dimensions: the spectral structure and the time-domain transient change. In the frequency-domain analysis channel, the high-precision fast Fourier transform algorithm is used to transform the input signal, and the original time-domain signal is resolved into its spectral distribution within the entire target frequency band. At the same time, combined with non-linear frequency band division processing, the wide frequency band from 0.1 GHz to 18 GHz is divided into several characteristic sub-bands according to the distribution laws of typical interference sources (such as cellular communication, wireless local area network, radar, etc.). For each sub-band, the power spectral density, in-band peak value, energy distribution, frequency band edge characteristics are extracted, and complex characteristics such as spectral non-Gaussianity and third-order spectral moments are obtained using methods such as high-order spectral analysis to form a frequency-domain feature set. At the same time, the time-domain analysis channel performs multi-scale wavelet packet decomposition processing on the differential signal, selects wavelet basis functions with good time-frequency localization ability such as db4, and decomposes the signal into a series of time-domain sub-band coefficients at multiple scales and resolutions. For each sub-band obtained by wavelet packet decomposition, statistical indicators such as its energy, variance, kurtosis, skewness, and entropy are calculated to reflect the local mutation, pulse interference, and non-stationary dynamic characteristics of the signal. These time-domain characteristics help to identify instantaneous interference, burst interference, and reveal long-term structural disturbances in complex electromagnetic environments. Based on the frequency-domain feature set and the time-domain feature set, collaborative feature extraction is performed, and collaborative feature extraction methods such as principal component analysis or independent component analysis are used to obtain a mixed-domain feature matrix covering multi-dimensional attributes. A non-linear dimensionality reduction algorithm such as t-SNE or autoencoder is used to perform high-dimensional data mapping and compression on the mixed-domain feature matrix, and then through unsupervised learning methods such as density clustering and spectral clustering, different interference patterns are naturally aggregated in the reduced-dimensional space, and discriminative interference feature fingerprints are extracted. Each group of interference feature fingerprints represents the unique structure and time-frequency behavior of potential interference sources in the current differential signal. The interference feature fingerprints are calculated with the pre-constructed interference source library at multiple levels, combined with various measurement criteria such as cosine similarity and Euclidean distance, the most similar interference source type and its confidence level are selected, and the matching result is mapped into a visual interference source feature mapping atlas.

[0020] 104. Perform multi-modal interference recognition and suppression according to the interference source feature mapping atlas, generate an interference cancellation signal, and based on the interference cancellation signal, perform real-time elimination and feedback on the electromagnetic interference in the inner shielding layer and the outer shielding layer to obtain the interference cancellation feedback effect.

[0021] Specifically, according to the candidate interference sources and their confidence levels in the interference source feature mapping spectrum, classify the types of interference existing in the current signal, and use clustering analysis and pattern recognition algorithms to divide all identified interference signals into three categories: narrowband interference, broadband interference, and burst interference. For narrowband interference, the system automatically matches a notch filter array. The center frequency and quality factor of each group of notch filters are set according to the interference spectrum characteristics to ensure deep suppression of the interference frequency points while maximizing the protection of the integrity of the target signal. After processing by the notch filter array, the continuous narrowband interference components in the corresponding frequency band are effectively filtered out to obtain a narrowband interference suppression signal. For broadband interference, due to its wide energy distribution and large affected frequency band span, threshold shrinkage processing based on wavelet transform is adopted for broadband interference. The energy distribution of the signal is analyzed in the multi-scale wavelet domain, and the main interference energy components are dynamically compressed to the background noise level through soft and hard thresholding strategies, effectively retaining the key information in the original signal to form a broadband interference suppression signal. At the same time, for burst interference, due to its strong time characteristics, short duration, and large amplitude fluctuations, an adaptive prediction and correction model is constructed based on the historical observation window and timing characteristics. Technologies such as ARMA and Kalman filtering are used to perform real-time modeling and prediction on the burst disturbance, and an in-phase compensation signal is generated in advance before the interference arrives to correct the impact of the burst interference in the first time and output a burst interference suppression signal. To maximize the interference cancellation effect, according to the confidence level data of each candidate interference source in the interference source feature mapping spectrum, the narrowband, broadband, and burst interference suppression signals are dynamically weighted and fused according to the confidence level weights, and the minimum mean square error criterion or Bayesian adaptive strategy is used to obtain a comprehensively optimized interference cancellation signal. The interference cancellation signal is input into a two-layer heterogeneous recurrent neural network for interference prediction processing. The inner layer focuses on the timing dynamic modeling of high-frequency complex interference with an LSTM structure, and the outer layer focuses on the fast response to external burst interference with a GRU structure. The two exchange information in the attention mechanism fusion layer to extract the multi-dimensional interference evolution law in real time. The neural network system combines time-frequency domain features, signal historical states, and external environment parameters to predict the interference trend in the short term in the future, and uses the prediction result as a feedback signal to guide the front-end filter, adaptive compensator, and overall parameter adjustment to achieve feedforward suppression and closed-loop dynamic optimization of the interference in the inner and outer shielding layers. Finally, the interference cancellation feedback effect is output.

[0022] Continuously sample the original interference data within S milliseconds before the occurrence of a sudden interference event, and segment it using a fixed-length time window, dividing the original signal into several overlapping or non-overlapping short-time segment samples. The samples within each window are normalized to eliminate the influence of amplitude and baseline drift, forming a structured and standardized interference sample set. Perform a stationarity test (such as the ADF test) and a white noise test (such as the Ljung-Box test) on the interference sample set, analyze the autocorrelation and randomness of the interference sequence, and then determine its time series characteristics, including whether it is suitable for using linear time series modeling methods. Based on the obtained time series characteristic analysis results, select the autoregressive moving average (ARMA) model to model the sudden interference, and establish an initial ARMA prediction model according to the order and lag parameters set by the model. In the model parameter estimation link, initially fit the coefficient matrix of the ARMA model by the maximum likelihood estimation method, and continuously adjust the model structure using the Box-Jenkins iterative calculation framework, and recursively optimize the parameters relying on the prediction residuals of the model to obtain a set of candidate model coefficients. Use the Akaike information criterion (AIC) to evaluate the parameter sets of different models, and select the optimal parameter set that comprehensively balances the model complexity and prediction ability by minimizing the AIC value. Substitute the optimal model parameters into the ARMA prediction model, perform back-substitution fitting on the existing interference sample data, calculate the fitting residual sequence to evaluate the model accuracy, and adjust the parameters in a timely manner according to the difference between the residual distribution and the model output to ensure that the model can truly reflect the evolution law of the interference waveform and obtain the interference waveform prediction function. Use the interference waveform prediction function to perform forward extrapolation calculation on the interference waveform within the next H milliseconds, generating a series of highly time-sensitive interference prediction values. To effectively suppress the sudden interference, these prediction values are processed by phase inversion and amplitude matching. The predicted waveform is phase-inverted by 180 degrees in time, and the energy level of the inverted signal is adaptively adjusted according to the amplitude of the original interference signal, so that the inverted signal can cancel the actual interference to the greatest extent during the superposition process. To avoid high-frequency ringing or other discontinuous noises brought by the newly generated inverted signal, finally, the inverted cancellation signal is input into a finite impulse response filter for smoothing processing. By optimizing the filter coefficients, the output signal not only has an ideal cancellation state with the interference in terms of amplitude and phase, but also has continuity and engineering feasibility, forming a high-quality signal for real-time suppression of sudden electromagnetic interference.

[0023] The frequency band of the interference cancellation signal is separated. Through digital filtering and band-pass / band-stop separation techniques, the signal is decomposed into a high-frequency interference component reflecting the characteristics of the inner shielding layer and a low-frequency interference component representing the dynamics of the outer shielding layer. The high-frequency component contains complex and rapidly changing interference source information from the core part of the cable, while the low-frequency component is more susceptible to factors such as the external environment, transient disturbances, and structural resonances. The extracted high-frequency interference component of the inner shielding layer is input into the long short-term memory network (LSTM) part of the double-layer heterogeneous recurrent neural network architecture. Utilizing its ability to model time-series data and capture the advantages of long-distance dependence information, a deep learning analysis is performed on the evolution pattern of high-frequency disturbances along the time axis. The LSTM network can utilize the historical state of the input sequence, dynamically adjust the parameters of the forget gate and input gate, and effectively predict and reconstruct the amplitude, phase, and periodic changes of high-frequency interference, outputting the interference prediction result of the inner shielding layer. The low-frequency interference component of the outer shielding layer is input into the gated recurrent unit network (GRU) branch of the same heterogeneous recurrent neural network to quickly identify and agilely predict external sudden interferences and slow evolution processes, obtaining the interference prediction result of the outer shielding layer. The prediction results output by the LSTM and GRU are input into a fusion layer designed with a multi-head attention mechanism. This layer dynamically weights the importance of different time series and spatial features, mining the internal correlation and mutual influence between the interference components of the inner and outer shielding layers. The multi-head attention mechanism can adaptively capture the activity intensity of key interference sources at different times and different spatial intervals, and achieve a high-precision aggregated expression of the distribution characteristics and influence paths of interference sources in space, obtaining the spatial location information of the interference sources. Based on the spatial location information of the interference sources, a hierarchical decoupling process is performed on the interference signals of the inner and outer shielding layers. Combining the mask masking strategy and feature independence analysis, the independent influence amounts of the inner-layer high-frequency interference and the outer-layer low-frequency interference are effectively separated, ensuring that the two types of interferences do not cross-couple or have energy leakage in subsequent suppression and feedback links. Through quantitative analysis of the separated interference effects of these two types, the system calculates their interference intensities at the target frequency band and critical moments respectively, forming a standardized and structured set of interference intensity vectors. The separated high-frequency and low-frequency interference intensities, spatial location data, and other multi-source information are comprehensively output to form a traceable and tunable interference cancellation feedback effect.

[0024] In the embodiments of the present invention, the function separation of the inner and outer shielding layers is achieved by using a radially progressive impedance gradient double-layer shielding structure design. The inner shielding layer focuses on protecting the signal line transmission stability, and the outer shielding layer focuses on protecting the charging wire from interference, significantly improving the anti-interference ability of the cable without increasing the overall diameter. The setting of the triangular support frame structure not only improves the compressive strength of the cable, but also creates an ideal geometric space between the inner and outer shielding layers, enabling the shielding efficiency to remain stable when the cable is compressed and deformed, achieving high-efficiency shielding in the entire frequency band. The shielding structure optimization algorithm based on the field strength decay coefficient precisely controls the material composition and geometric parameters of the inner and outer shielding layers, ensuring the maximization of shielding efficiency under the given thickness limit and effectively reducing the cable manufacturing cost. The real-time differential signal acquisition system for the inner and outer shielding layers enables the cable to have the ability to perceive the electromagnetic environment, be able to monitor the characteristics of the interference source in real time, and provide accurate data support for subsequent interference suppression. The interference feature extraction method based on time-frequency domain joint analysis realizes the precise identification and positioning of the interference source. Through the collaborative analysis of frequency domain and time domain features, different types and sources of interference signals can be distinguished. The multi-modal adaptive interference identification and elimination algorithm is optimized for different frequency bands and interference characteristics, greatly improving the signal transmission quality without changing the physical structure and effectively compensating for signal reflection and crosstalk between the internal layers of the cable. The double-layer heterogeneous recurrent neural network interference prediction system separates the influence of inner-layer high-frequency interference and outer-layer low-frequency interference through hierarchical decoupling technology, solves the problem of cross-influence of interference sources in traditional methods, and performs outstandingly in an environment with complex electromagnetic interference.

[0025] In a specific embodiment, the process of executing step 101 may specifically include the following steps: Perform cylindrical central positioning on the cable conductor to obtain a copper core conductor, and extrude and coat the copper core conductor with polytetrafluoroethylene material to obtain an insulating layer; Perform silver-copper alloy wire mesh braiding treatment on the insulating layer to obtain an inner shielding layer, and perform fluorinated ethylene propylene material coating treatment on the inner shielding layer to obtain an insulating isolation layer; Perform tin-copper alloy wire mesh braiding treatment on the insulating isolation layer to obtain an outer shielding layer; Manufacture a triangular support frame structure based on reinforced nylon material, and set the triangular support frame structure along the cable axis to obtain an electromagnetic isolation space between the inner shielding layer and the outer shielding layer.

[0026] Specifically, cylindrical central positioning is performed on the cable conductor. The geometric center of the copper core conductor is related to the overall symmetry of the cable. Any slight eccentricity will lead to problems such as uneven insulation layer thickness, impedance discontinuity, and signal distortion in the subsequent process. In the pre-processing stage of the conductor, a numerical control machine tool and a high-precision coaxial fixture are used to align the axis of the copper core conductor in the longitudinal and radial directions, ensuring that its cylindrical geometric dimensions meet the design standards. After the central positioning is completed, the online detection system monitors the axis position, diameter uniformity, and surface roughness of the conductor throughout the process, excluding errors caused by mechanical processing or raw material fluctuations. Polytetrafluoroethylene (PTFE) material with excellent high-frequency dielectric properties and extremely low loss factor is selected to perform continuous extrusion coating on the positioned copper core conductor. The extrusion coating process is carried out in an extruder unit with precise temperature control. It is required that the PTFE material forms a dense and bubble-free coating layer after heating and softening, and the thickness is controlled within the range of 0.8 ± 0.05 mm. Through the synchronous adjustment of process parameters such as temperature, pressure, and linear velocity, the insulation layer is uniformly formed on the surface of the copper core conductor. After coating, the insulation layer undergoes online spark detection and insulation withstand voltage testing to ensure that its continuity, integrity, and electrical safety performance meet the standards. The insulation layer is processed by braiding a silver-copper alloy wire mesh. An automated braiding device is used to evenly and alternately braid high-purity silver-copper alloy wires with a braiding density of 96 cores × 0.10 mm around the surface of the insulation layer to form a grid-shaped inner shielding layer. The high-density and wide-coverage silver-copper alloy wire mesh exhibits extremely low surface impedance in a high-frequency electromagnetic environment, and through the structural advantage of multi-point grounding, it effectively shields the signals transmitted inside the cable from being invaded by external high-frequency interference and maintains signal integrity. During the braiding process, the braiding tension and braiding angle are dynamically adjusted to avoid defects such as wire mesh relaxation, overlap, or breakage caused by uneven mechanical stress. The impedance and structural consistency are monitored in real time throughout the braiding process, and an inner shielding layer with ideal shielding ability is output. The inner shielding layer is subjected to a secondary coating treatment with fluorinated ethylene propylene (FEP) material. The FEP material has excellent insulation withstand voltage and thermal stability. Through the melt extrusion or casting coating process, an FEP insulation isolation layer with a thickness of 0.5 ± 0.03 mm is uniformly coated on the outside of the inner shielding layer to achieve a multi-layer insulation system, effectively isolating the potential leakage current and electric field coupling between the inner and outer shields, suppressing surface discharge and insulation aging problems, and improving the operation reliability of the entire cable under high voltage, high frequency, and complex climate conditions. The insulation isolation layer is processed by braiding a tin-copper alloy wire mesh. 64-core × 0.12 mm tin-copper alloy wires are used to perform precise grid braiding on the outside of the FEP insulation isolation layer with a high braiding coverage rate. The conductivity and cost of the tin-copper alloy are balanced, enabling it to not only ensure the effective blocking and energy dissipation of the outer shield against high-frequency interference but also improve the mechanical strength and corrosion resistance of the entire cable. The wire diameter uniformity, braiding tension, and coverage density are monitored synchronously during the braiding process to avoid shielding failure or damage to the shielding layer caused by external mechanical impact due to uneven tightness, and a high-performance outer shielding layer is formed.Through a multi-layer composite shielding structure and a gradient impedance layout between the inner and outer shielding layers, external electromagnetic waves are dissipated and attenuated layer by layer when penetrating the cable structure, enhancing the overall anti-interference ability. In order to form a stable and controllable electromagnetic isolation space between the inner and outer shielding layers and endow the cable with higher mechanical strength and structural stability, a triangular support frame structure is fabricated based on reinforced nylon material. Reinforced nylon, with its high mechanical strength, excellent wear resistance, and low dielectric loss, becomes an ideal material for the support structure. Through high-precision injection molding or 3D printing processes, support columns with an equilateral triangle cross-section are manufactured. Each group of triangular support frames consists of three support columns distributed at 120 degrees, and one group is precisely arranged every 50 ± 2 mm along the cable axis direction, realizing the full physical isolation and support of the shielding layer spacing. The side length of each support column is controlled within 2.0 ± 0.1 mm, and the spacing accuracy is maintained at 1.0 ± 0.05 mm. The support frame eliminates the distance fluctuations between the inner and outer layers caused by pressure, tension, or bending of the cable and provides an ideal space for the arrangement of micro-probes in the subsequent differential signal acquisition system.

[0027] In a specific embodiment, the method for anti-interference of high-frequency coaxial cables further includes the following steps: Conduct a mathematical modeling of the electromagnetic field distribution of the high-frequency coaxial cable to obtain a model of the relationship between the electromagnetic field strength and the radial distance; Based on the model of the relationship between the electromagnetic field strength and the radial distance, calculate the attenuation coefficient of the inner shielding layer to obtain the inner shielding layer attenuation coefficient; Based on the model of the relationship between the electromagnetic field strength and the radial distance, calculate the attenuation coefficient of the outer shielding layer to obtain the outer shielding layer attenuation coefficient; Based on the inner shielding layer attenuation coefficient and the outer shielding layer attenuation coefficient, set the ratio of the inner and outer layer attenuation coefficients to obtain the attenuation coefficient ratio; According to the attenuation coefficient ratio, adjust the material compositions of the inner and outer shielding layers to obtain the optimized material ratio of the inner and outer shielding layers; Input the optimized material ratio and geometric parameters of the inner and outer shielding layers into the electromagnetic field analysis system for iterative calculation to obtain the shielding structure design parameters.

[0028] Specifically, based on Maxwell's equations, the coaxial cable is simplified into a multi-layer cylindrical symmetric structure, corresponding to the conductor, inner insulation, inner shielding layer, isolation layer, outer shielding layer, and external environment respectively. Since the coaxial cable is used for high-frequency signal transmission, both the conduction characteristics of the cable itself and the incident, reflection, and absorption mechanisms of the external electromagnetic field need to be considered during analysis. For this purpose, the cylindrical coordinate system (r, φ, z) is adopted to describe the variation of the electric field strength at different radial positions. The external incident electromagnetic field is regarded as a plane wave, which undergoes reflection, absorption, and transmission under the action of the shielding layer, and its radial distribution law is represented by an exponential decay model, that is, the electromagnetic field strength E(r) = E0·e (-αr), where \(E_0\) is the incident point field strength, \(\alpha\) is the attenuation coefficient, and \(r\) is the radial distance. This model can describe the attenuation process of electric field energy in a multi-layer shielding structure and is also convenient for subsequent quantitative analysis of the attenuation coefficient. Based on the above mathematical model, the attenuation coefficients are calculated separately for the inner shielding layer and the outer shielding layer. The attenuation coefficient \(\alpha\) is the core parameter for measuring the shielding ability of materials against high-frequency electromagnetic waves, and its value is closely related to the conductivity \(\sigma\), permeability \(\mu\), and operating frequency \(f\) of the material. For the inner shielding layer, a high-purity silver-copper alloy is selected, and its attenuation coefficient is given by the formula \(\alpha_1 = 4.7\times10\) 3 \(\times f\) 0.5×σ1×μ1 , where \(\sigma_1\) is the conductivity of the silver-copper alloy and \(\mu_1\) is its relative permeability. The outer shielding layer is mainly made of a tin-copper alloy, and its attenuation coefficient \(\alpha_2 = 4.7\times10\) 3 \(\times f\) 0.5×σ2×μ2. Through the above formula, in the working frequency band of 0.1-18GHz, the α1 and α2 values ​​at different frequency points are calculated for specific design parameters and material properties to achieve a quantitative description of the electromagnetic energy attenuation path. In actual engineering, the multi-layer structural parameters and material physical properties of the cable are input into the simulation software, and the full-band distribution curve of the reduction coefficient is obtained by frequency domain scanning. Combined with the typical interference sources that the cable may encounter at different frequencies, the shielding effectiveness is fully evaluated. After obtaining the reduction coefficients of the inner and outer shielding layers, in order to achieve the coordinated optimization of the shielding structure, the reduction coefficient ratio of the two is set. This ratio determines the distribution and reduction efficiency of energy in the multi-layer structure. Engineering experience and simulation analysis show that when the reduction coefficient ratio K=α1 / α2 of the inner and outer shielding layers is set to a certain optimal constant, it can achieve the maximum shielding of high-frequency signals and the balance of signal integrity under the constraints of limited thickness and cost. If the K value is too large, the inner layer shielding effectiveness is redundant and the outer layer is relatively weak, which is susceptible to external interference; if the K value is too small, the outer layer consumes too many resources and the inner layer is insufficiently protected, making it difficult to fully resist high-frequency crosstalk. Through theoretical modeling and simulation iteration, the appropriate K value is selected for structural optimization by combining different interference scenarios, working conditions and application requirements. Based on the determination of the K value, the material ratio of the inner and outer shielding layers is adjusted in reverse. Taking silver-copper alloy and tin-copper alloy as examples, the conductivity and magnetic permeability of the material are directly affected by adjusting the content of silver or tin. In order to meet the constraints of the decreasing coefficient ratio, the silver content of the inner layer is controlled at about 30%, so that it has low impedance and high shielding capability at high frequencies, and the tin content of the outer layer is set to 15%, taking into account cost, mechanical strength and broadband anti-interference. While adjusting the material ratio, the geometric parameters such as weaving density, wire thickness and coverage are optimized in a linked manner. For example, the inner shielding layer is designed with a high weaving density and a small wire diameter to enhance the initial suppression of high-frequency weak interference, and the outer shielding layer is appropriately thickened and the weaving density is reduced to improve impact resistance and disperse external energy. The optimized material ratio of the inner and outer shielding layers and all key geometric parameters (such as the thickness of each layer, the spacing between the shielding layers, the size of the support frame, etc.) are input into the electromagnetic field analysis system, and a complete three-dimensional model is established using electromagnetic simulation software. Full-frequency simulation and multiple rounds of iterations are performed on typical interference scenarios in the 0.1-18GHz frequency band. In the simulation, the reflection loss, absorption loss and multiple reflection correction of the incident electromagnetic wave by the shielding structure are examined, and the overall shielding effectiveness SE (including SEi, SEr, and SEa) is parameter scanned. The material composition and structural parameters are automatically adjusted through numerical optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.), combined with multi-objective trade-offs such as shielding effectiveness, mechanical strength, thermal stability and process feasibility, to screen out parameter combinations that can ensure shielding efficiency higher than 85dB in the full frequency band, and finally output the shielding structure design parameters.

[0029] In a specific embodiment, the process of executing step 102 may specifically include the following steps: Install electromagnetic induction probes on the triangular support frame structure to obtain multiple electromagnetic induction probes evenly distributed in the gap between the inner shielding layer and the outer shielding layer; Conduct isolated coaxial lead connections for the multiple electromagnetic induction probes to obtain multiple probe signal channels connected to the signal acquisition module, and perform digital conversion processing on the signals collected by the multiple probe signal channels to obtain multiple digitized differential signals; Perform digital filtering and decimation processing on the multiple digitized differential signals to obtain multiple differentially processed signals after preliminary processing; Perform window slicing and phase correction processing on the multiple differentially processed signals after preliminary processing to obtain time-aligned multiplexed signals, and perform weighted average processing on the time-aligned multiplexed signals to obtain differential interference signals.

[0030] Specifically, an electromagnetic induction probe is installed on the triangular support frame structure. The support frame with an equilateral triangular cross-section is made of high-strength and low-dielectric-loss reinforced nylon material. Each group of support frames consists of three support columns distributed along the circumference of the cable at an angle of 120°. At the top or side of the support frame, a preset installation hole diameter or card slot is provided. According to the designed spacing and angle, a high-sensitivity platinum-rhodium alloy micro induction probe with a diameter of 0.5±0.02 mm is embedded therein, so that each group of support frames can carry three probes, and the probe detection head penetrates into the gap area of about 1 mm between the inner and outer shielding layers, realizing the real-time perception of small changes in the electromagnetic field. With the uniform arrangement of the support frames along the entire length of the cable, multiple groups of evenly distributed electromagnetic induction arrays are formed in the space between the inner and outer shielding layers. Each probe is independently connected to the signal acquisition module outside the cable by an isolated micro coaxial lead. The shielding layer of the coaxial lead is grounded and connected to the shielding structure of the cable body, the central conductor is connected to the signal end of the probe, and the outer shielding is grounded to the system, forming a high-frequency signal transmission path with low loss and anti-crosstalk. Inside the signal acquisition module, a high-performance multi-channel analog-to-digital conversion system (ADC) is set. Each probe signal channel is configured with a dedicated amplifier, buffer, anti-aliasing filter, and 24-bit high-dynamic-range ADC. The sampling rate is as high as 40 MSPS, which can capture nanosecond-level interference mutations and high-frequency broadband interference information. After all channel signals enter the digital domain, bandwidth control and high- and low-frequency noise suppression are performed through a digital filter. The passband, stopband, and transition band widths are set according to the target frequency band through the FIR (finite impulse response) filter structure to suppress spurious interference and system background noise in non-target frequency bands. To reduce the redundant data volume and improve the efficiency of backend processing, a digital decimation operation is performed on each signal channel, and the high-frequency sampling data is compressed to 10 MSPS or lower at a controllable downsampling rate, while ensuring that the target interference components are not distorted or folded, forming multiple preliminarily processed digital differential signals. Window slicing and phase correction are performed on all preliminarily processed signals. Window slicing is to slice the signal according to a fixed time length, with 50% overlap for each slice to avoid cross-boundary interference loss, and the segmentation of multiple channels of windows is processed synchronously. Phase correction is based on signal correlation analysis and synchronization alignment technology. For each channel signal, its main peak arrival time, phase shift, and clock drift are calculated respectively. Through methods such as digital delay compensation and shift alignment, multiple channel signals are strictly aligned within the same window period, eliminating the out-of-step phenomenon caused by inconsistencies in physical or electronic links. After the multiple channel signals are time-aligned, weighted average fusion processing is performed on the multi-channel signals within each window. By using the complementarity between the signals collected at different spatial positions, accidental interference, random noise, and local anomalies are suppressed to the greatest extent, while highlighting the dominant components of common interference signals, and finally a differential interference signal is obtained.

[0031] In a specific embodiment, the process of performing step 103 may specifically include the following steps: Perform two-channel splitting on the differential interference signal to obtain a frequency-domain analysis channel signal and a time-domain analysis channel signal; Perform fast Fourier transform and non-linear frequency band division processing on the frequency-domain analysis channel signal to obtain a frequency-domain feature set, and perform multi-scale wavelet packet decomposition processing on the time-domain analysis channel signal to obtain a time-domain feature set; Perform collaborative feature extraction based on the frequency-domain feature set and the time-domain feature set to obtain a mixed-domain feature matrix, and perform non-linear dimensionality reduction and clustering processing on the mixed-domain feature matrix to obtain an interference feature fingerprint; Perform multi-level similarity calculation on the interference feature fingerprint and a pre-constructed interference source library to obtain an interference source feature mapping map.

[0032] Specifically, the differential interference signal is split into two channels, and the same set of signals is simultaneously introduced into the frequency-domain analysis channel and the time-domain analysis channel at the hardware or digital processing level to synchronously obtain the structured descriptions of the signals in two complementary spaces of the frequency domain and the time domain. After receiving the differential interference signal, the frequency-domain analysis channel uses a high-performance DSP or FPGA platform to implement the fast Fourier transform, converting the signal from a time-domain sequence into a spectrum representation containing amplitude and phase information. Due to the highly complex distribution of interference sources in the high-frequency coaxial cable application environment, the system not only performs a global FFT on the entire frequency band but also adopts a non-linear frequency band division method. According to the distribution laws of typical interference sources such as communication base stations, radars, Wi-Fi, etc., the range of 0.1 - 18 GHz is dynamically divided into several refined characteristic sub-bands, and high-order statistics such as power spectral density, main peak energy, in-band mean, peak-to-peak distance, third-order spectral moment, amplitude modulation and frequency modulation characteristics, and frequency band edge steepness are calculated for different sub-bands respectively. Each frequency-domain feature reflects the energy distribution and dominant frequency of a certain type of interference, and reveals the modulation mode, non-Gaussianity, and aliasing risk of the interference signal, thus forming a frequency-domain feature set. At the same time, the time-domain analysis channel performs multi-scale wavelet packet decomposition on the differential interference signal, selects wavelet bases with excellent time-frequency localization performance such as Daubechies (db4), and decomposes the signal at multiple scales and different time resolutions to obtain 128 groups of time-domain sub-band coefficients. For these wavelet sub-bands, multiple time-domain features such as their energy distribution, standard deviation, skewness, kurtosis, mutual information between sub-bands, entropy value, transient amplitude fluctuation, and local extreme value statistics are calculated to capture the mutations, short-time pulses, overlapping disturbances, and persistent changes of the interference signal on the time axis. The wavelet packet decomposition reveals the energy aggregation and diffusion of the signal within a local window and is suitable for identifying complex interferences such as sudden, non-stationary, or multi-source superposition. These time-domain features are constructed into a time-domain feature set. Collaborative feature extraction is performed based on the frequency-domain feature set and the time-domain feature set. Specific methods include feature normalization, principal component analysis, independent component analysis, and feature mutual information measurement, etc. By reasonably setting feature weights, the high-dimensional, multi-source, and heterogeneous original feature vectors are reduced to a set of discriminative and highly complementary mixed-domain feature matrices. Non-linear dimensionality reduction is performed on the mixed-domain feature matrix, such as through t-SNE (t-distributed stochastic neighbor embedding), autoencoders, multi-dimensional scaling analysis, etc. Through the non-linear mapping from high-dimensional features to a low-dimensional embedding space, the local similarity and global distribution structure of the data are retained, and the structured compression of interference features is achieved. The reduced-dimensional feature space helps to improve the convergence speed and recognition accuracy of clustering algorithms. Through clustering techniques such as K-means, the characteristic clusters of different types of interference sources are automatically separated within the embedding space. The center of each feature cluster or the clustering core represents the "feature fingerprint" of a typical interference signal, and these interference feature fingerprints have robust discriminative capabilities across time and frequency domains. Perform multi-level similarity calculations between the above-extracted interference feature fingerprints and the pre-constructed interference source library.The interference source library includes the time-frequency templates, statistical features, and historical evolution trajectories of various typical interference sources commonly encountered in practical applications, covering multi-source scenarios such as wireless communication, industrial equipment, environmental noise, and radar. The multi-level similarity matching combines complex metrics such as Mahalanobis distance, Gaussian kernel correlation, and dynamic time warping to comprehensively compare each set of feature fingerprints with the templates in the library and output the top N candidate interference sources with the highest confidence. At the same time, by integrating the matching scores under different feature dimensions and using voting or weighting strategies, the optimal interference type determination and related confidence distribution are obtained, and an interference source feature mapping graph is generated.

[0033] In a specific embodiment, the process of executing step 104 may specifically include the following steps: According to the candidate interference sources and their confidence levels in the interference source feature mapping graph, the interference types are classified to obtain narrowband interference, broadband interference, and burst interference after classification; The narrowband interference is processed by a notch filter array to obtain a narrowband interference suppression signal; The broadband interference is processed by wavelet domain threshold shrinking to obtain a broadband interference suppression signal; The burst interference is processed by prediction and correction modeling to obtain a burst interference suppression signal; The narrowband interference suppression signal, broadband interference suppression signal, and burst interference suppression signal are weighted and fused according to the confidence levels in the interference source feature mapping graph to obtain an interference cancellation signal; The interference cancellation signal is input into a double-layer heterogeneous recurrent neural network for interference prediction processing to obtain an interference cancellation feedback effect.

[0034] Specifically, based on the interference source feature mapping atlas, for the specific types and confidence levels of candidate interference sources, a clustering label, confidence ranking, and multi-level feature voting mechanism are used to classify all interference signals encountered by the current cable. According to the time-frequency structure, energy distribution, transient change characteristics of the interference source, and the comparison results with historical templates, each interference event is determined to be one of three categories: narrowband interference, broadband interference, or burst interference. And a weight related to its confidence level is assigned to each type of interference signal. Narrowband interference is manifested as single or a few frequency points with narrow distribution and concentrated energy in the signal spectrum, such as radio frequency spurs, carrier leakage, harmonics, or some wireless communication interferences. Broadband interference is manifested as interference with energy covering multiple continuous or discontinuous sub-bands and broad frequency distribution, such as pulse trains, broadband emission sources, power electronic device clutter, etc. Burst interference often has sudden changes in the time domain, short duration, obvious transient peaks, and unpredictability, such as switching impulses, surge spikes, etc. For narrowband interference, a notch filter array is used for precise suppression. The center frequency of each group of notch filters is automatically locked to the interference frequency point, and the bandwidth and quality factor (Q value) are dynamically adjusted according to the spectrum details. The notch filters are implemented using IIR or FIR, and a very deep band suppression effect is formed through multi-stage cascading. The entire array works in parallel to deeply suppress multiple different frequency points or narrowbands simultaneously, and at the same time, the notch center is continuously adjusted through a dynamic tracking algorithm to ensure that the suppression effect is adaptively real-time with the appearance of frequency drift or multi-source combined interference. The interference energy near the target frequency point of the processed signal is effectively suppressed, and other frequency bands of the main signal maintain their original characteristics, and the output is a narrowband interference suppression signal. For broadband interference, because the frequency range it involves is wide and the energy distribution is complex, directly using notch filtering will cause a large amount of useful signal loss. The system uses multi-scale wavelet packet transform to decompose the signal, projects the broadband signal into different time-frequency components, and combines soft threshold and hard threshold strategies to dynamically detect and suppress the high-energy interference components in each wavelet sub-band. The selection of the threshold is adaptively adjusted based on noise statistical estimation, sub-band energy distribution, or Bayesian learning. After threshold shrinkage in the wavelet domain, the main noise or abnormal components in the broadband interference signal are reduced to near the noise floor, and the instantaneous details and key information of the signal are retained to the greatest extent. The output is a broadband interference suppression signal with smoothness, low noise, and high fidelity characteristics. For burst interference, due to its extremely strong temporal unpredictability and mutability, the system is based on temporal statistical modeling and intelligent prediction mechanisms, and autoregressive moving average, Kalman filter, or recurrent neural networks (such as LSTM, GRU) are selected to dynamically model the interference data. After performing stationarity tests and white noise analysis on the interference historical data within the previous S milliseconds, and determining the sequence characteristics, the model parameters are trained using iterative methods such as maximum likelihood estimation and Box-Jenkins, and the optimal model order and parameter set are selected through the Akaike information criterion.After obtaining the accurate interference prediction function, the signal waveform within the next H milliseconds is extrapolated forward to generate interference prediction values. Then, through phase inversion and amplitude matching techniques, an anti-phase cancellation signal is synthesized. Finally, after smoothing by a finite impulse response filter, a high-quality signal that can directly cancel the impact of burst interference, namely the burst interference suppression signal, is obtained. According to the confidence level in the feature mapping atlas, the narrowband, broadband, and burst interference suppression signals are weighted and fused. Strategies such as weighted average, least mean square error criterion, and Bayesian posterior fusion are adopted, and the fusion coefficients are allocated based on the probability contribution and harm intensity of each type of interference in the current environment to achieve intelligent resource allocation of "strong interference with strong suppression and weak interference with weak processing". The interference cancellation signal generated after multi-channel signal fusion has both broad-spectrum suppression ability and main signal protection ability in a multi-source interference environment. The interference cancellation signal is input into a two-layer heterogeneous recurrent neural network for interference prediction processing. One layer of the network is a long short-term memory (LSTM) unit, which is specifically used to model high-frequency, continuous, or long-term correlation interference and capture the implicit patterns and slow-varying trends of the signal over time. The other layer is a gated recurrent unit (GRU), which mainly focuses on low-frequency, burst, and environment-related interference, and has the advantages of fast response speed, simple parameters, and high sensitivity to short-term dynamic changes. After parallel processing, the two types of networks are fused through an attention mechanism to extract the most discriminative temporal and spatial features. The system automatically learns the propagation patterns and spatial evolution laws of different types of interference in the current cable shielding system to achieve accurate temporal prediction and spatial positioning of the interference source. At the output stage of the neural network, the system gives multi-dimensional feedback on the future interference trend, including the interference intensity at key frequency points in the short term, the dominant direction of spatial distribution, and the affected degree of each layer of the cable structure, and accordingly dynamically adjusts various front-end filtering, compensation, and adjustment parameters to form a full-process closed-loop optimization from acquisition, recognition, suppression, fusion to prediction and feedback.

[0035] In a specific embodiment, the process of performing steps to predict and correct the modeling of burst interference to obtain the burst interference suppression signal may specifically include the following steps: The data of the first S milliseconds of the burst interference is segmented into samples and normalized to obtain an interference sample set for multiple time windows, and the interference sample set is subjected to a stationarity test and a white noise test to obtain the analysis results of the interference time series characteristics; Based on the analysis results of the interference time series characteristics, an autoregressive moving average model is established for the burst interference to obtain an ARMA prediction model; The maximum likelihood estimation and Box-Jenkins iterative calculation are performed on the ARMA prediction model to obtain a model coefficient matrix, and the Akaike information criterion is used to evaluate the model coefficient matrix to obtain an optimal model parameter set; Substitute the optimal model parameter set into the ARMA prediction model, perform back substitution fitting and residual analysis on the interference data to obtain the model fitting accuracy evaluation result, and fine-tune the model parameters according to the model fitting accuracy evaluation result to obtain the interference waveform prediction function; Perform forward extrapolation calculation on the interference waveform prediction function to obtain the interference waveform prediction values within the next H milliseconds, and perform phase inversion and amplitude matching processing on the interference waveform prediction values to obtain the anti-phase cancellation signal; Input the anti-phase cancellation signal into a finite impulse response filter for smoothing processing to obtain the burst interference suppression signal.

[0036] Specifically, the historical data of the first S milliseconds of the sudden interference signal is processed by sliding window segmentation, splitting the long-time series signal into a series of short-time segments with fixed lengths and overlapping each other. Normalization processing is performed on each window segment to make the sample mean zero and the standard deviation one, eliminating the influence of the original amplitude change, trend drift, or measurement drift on the modeling result, and forming an interference sample set under a unified scale. Stationarity test and white noise test are performed on the interference sample set. By performing stationarity test (such as ADF test) and white noise test (such as Ljung-Box test) on each time window segment, it is judged whether the sequence conforms to a stationary random process or has long memory characteristics. The analysis results directly determine the selection of subsequent ARMA model parameters and the setting of model complexity. If the interference time series characteristics show stationarity, the autoregressive moving average (ARMA) model can efficiently perform signal fitting and future prediction; if significant non-stationarity is found, differencing is introduced or extended to higher-order models such as ARIMA. In the ARMA modeling stage, a suitable model order is set based on the interference sample set and its statistical characteristics, and then the maximum likelihood estimation method is used to initially fit the coefficient parameters of the ARMA model. On this basis, with the help of the Box-Jenkins method, through the recursive analysis of the residual sequence and the iterative correction of model parameters, the optimal approximation of the coefficient matrix is achieved. During this process, the system uses the Akaike information criterion (AIC) to comprehensively evaluate the model complexity and goodness of fit at different orders, and the parameter set corresponding to the model with the minimum AIC is the optimal structure of the current sequence. The optimal parameter set is brought into the ARMA prediction model, and the entire S-millisecond data set is subjected to back substitution fitting, and statistical analysis is performed on the fitting residuals. By calculating the residual variance, distribution skewness, kurtosis, and autocorrelation, the fitting accuracy of the model is quantified. If systematic deviation or structural abnormality is found in the residuals, the model order and parameters are fine-tuned again until the residuals exhibit an approximate white noise distribution, and a prediction function that can accurately describe the interference evolution is obtained. Based on the high-precision interference waveform prediction function, forward extrapolation calculation is performed to continuously predict the interference waveform within the future H-millisecond time window, so as to predict the upcoming interference trend and its amplitude evolution. In order to achieve active suppression of sudden interference, the predicted interference waveform signal is subjected to phase inversion processing, that is, the waveform is flipped 180 degrees on the time axis, and at the same time, the amplitude of the inverted signal is adaptively matched so that its energy is as equal as possible to the actual interference signal but in the opposite direction. This operation ensures that when the inverted signal is superimposed on the actual interference, the cancellation effect can be maximally achieved. Considering that prediction and phase inversion processing introduce new high-frequency noise or local glitches, the inverted cancellation signal is input into a finite impulse response (FIR) digital filter for smoothing filtering of the signal. The FIR filter coefficients are optimized by the Parks-McClellan algorithm to ensure that the filtered signal has ideal passband and stopband characteristics, extremely low phase distortion, and small passband ripple in the frequency domain, and finally a sudden interference suppression signal is obtained.

[0037] In a specific embodiment, the process of inputting the interference cancellation signal into the double-layer heterogeneous recurrent neural network for interference prediction processing to obtain the interference cancellation feedback effect may specifically include the following steps: Perform frequency band separation on the interference cancellation signal to obtain the high-frequency interference component of the inner shielding layer and the low-frequency interference component of the outer shielding layer; Input the high-frequency interference component of the inner shielding layer into the long short-term memory network in the double-layer heterogeneous recurrent neural network to capture the temporal characteristics and obtain the interference prediction result of the inner shielding layer; Input the low-frequency interference component of the outer shielding layer into the gated recurrent unit network in the double-layer heterogeneous recurrent neural network to identify the burst interference and obtain the interference prediction result of the outer shielding layer; Perform attention mechanism fusion on the interference prediction result of the inner shielding layer and the interference prediction result of the outer shielding layer to obtain the interference source spatial positioning information; Based on the interference source spatial positioning information, perform hierarchical decoupling processing on the interference signals of the inner and outer shielding layers to obtain the separated inner-layer high-frequency interference influence and outer-layer low-frequency interference influence; Calculate the interference intensity based on the separated inner-layer high-frequency interference influence and outer-layer low-frequency interference influence to obtain the interference cancellation feedback effect.

[0038] Specifically, frequency band separation processing is performed on the interference cancellation signal. Based on the physical structure of the inner and outer shielding layers and the electromagnetic field propagation mechanism, through band-pass, band-stop or adaptive filtering techniques, the original interference cancellation signal is divided into two complementary components in the frequency domain: one part is the high-frequency interference component reflecting the microscopic disturbances of the inner shielding layer, and the other part is the low-frequency interference component representing the macroscopic disturbances of the outer shielding layer. The high-frequency component of the inner shielding layer contains signal components with rapid changes, small amplitudes, and rich time-domain details, which are easily affected by high-frequency disturbances such as internal device switching and electromagnetic leakage. The low-frequency component of the outer shielding layer, on the other hand, shows large-scale, slow-changing, and energy-concentrated disturbance components, which originate from external power systems, environmental resonances, or low-frequency electromagnetic interference. The high-frequency component is input into the long short-term memory network (LSTM) branch of the double-layer heterogeneous recurrent neural network architecture. With its multi-gated structure and persistent memory ability, LSTM can fully capture the dynamic dependencies of the input sequence in both the long term and the short term. By performing temporal modeling on the high-frequency interference component of the inner shielding layer, LSTM can identify complex features such as the periodicity, transient pulses, and crosstalk evolution of high-frequency interference, and predict the disturbance intensity and dominant mode changes at future moments. The system performs normalization, batch training, and dynamic weight updating on the sequence input in the LSTM network, enabling the model to adaptively optimize the thresholds of the memory gate, input gate, and forget gate under different interference scenarios and parameter fluctuations, improving the perception sensitivity and prediction stability for high-frequency complex interference, and outputting high-resolution temporal prediction results of the inner shielding layer interference. At the same time, the low-frequency interference component is input into the gated recurrent unit (GRU) branch of the heterogeneous neural network architecture. GRU is suitable for capturing external disturbances with low frequency, suddenness, and strong non-stationarity. The system performs temporal analysis on the low-frequency component of the outer shielding layer within the GRU branch. With the dynamic adjustment of the reset gate and update gate, the network can flexibly switch between different historical states, accurately identify and temporally predict typical phenomena such as the instantaneous amplitude increase of sudden interference signals, non-stationary mode changes, and short-period periodic interference, and output the prediction results of the low-frequency disturbances of the outer shielding layer. The temporal prediction vectors of LSTM and GRU are sent to the multi-head attention mechanism fusion layer. In this layer, multiple groups of parallel attention heads are used to dynamically weight the importance of different feature channels. Based on the correlation of the input signal in the three dimensions of time, frequency, and space, it automatically focuses on the feature information that is most discriminative for the overall interference environment. Through the combination of global weighting and local weighting, the system real-time integrates the prediction data from the high-frequency and low-frequency components, improving the perception accuracy of core spatial attributes such as spatial interference distribution, signal source direction, and dominant propagation path, and outputting the spatial location information of the interference source. Based on the spatial location information output by the multi-head attention mechanism, hierarchical decoupling processing is performed on the interference signals of the inner and outer shielding layers.The hierarchical decoupling relies on signal processing means such as the mask mechanism, subspace separation algorithm, and independent component analysis to effectively strip the independent influences of high-frequency and low-frequency interferences at their respective levels, enabling the interference components of each shielding layer to maximally restore their true effects both mathematically and physically, and avoiding energy coupling or misjudgment of different-level interferences during the superposition and feedback processes. Through hierarchical decoupling, the separated inner-layer high-frequency interference influence and outer-layer low-frequency interference influence are obtained. Based on the separated inner-layer high-frequency interference influence and outer-layer low-frequency interference influence, the interference intensity is calculated. The quantization of the interference intensity is comprehensively given based on multi-dimensional indicators such as energy spectrum analysis, root mean square of amplitude, maximum peak statistics, and instantaneous energy integration. At the same time, combined with the predicted spatial distribution information, a hierarchical, frequency-divided, and space-divided interference intensity matrix is output to obtain the interference cancellation feedback effect.

[0039] The high-frequency coaxial cable anti-interference method in the embodiments of the present invention has been described above. Next, the high-frequency coaxial cable anti-interference device in the embodiments of the present invention will be described. Please refer to Figure 2 , an embodiment of the high-frequency coaxial cable anti-interference device in the embodiments of the present invention includes: A construction module 201, configured to construct a radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and a triangular support frame structure is arranged between the inner shielding layer and the outer shielding layer; An acquisition module 202, configured to install electromagnetic induction probes on the triangular support frame structure to acquire the differential interference signal between the inner shielding layer and the outer shielding layer; A joint analysis module 203, configured to perform joint time-frequency domain analysis on the differential interference signal to obtain an interference source feature mapping spectrum; A feedback module 204, configured to perform multi-modal interference recognition and suppression according to the interference source feature mapping spectrum, generate an interference cancellation signal, and perform real-time elimination and feedback on the electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference cancellation signal to obtain the interference cancellation feedback effect.

[0040] Through the collaborative cooperation of the above-mentioned various components, the functional separation of the inner and outer shielding layers is achieved by adopting the design of a radially progressive impedance gradient double-layer shielding structure. The inner shielding layer focuses on protecting the transmission stability of the signal line, and the outer shielding layer focuses on protecting the charging wire from interference, enabling the cable to significantly improve its anti-interference ability without increasing the overall diameter. The setting of the triangular support frame structure not only improves the compressive strength of the cable but also creates an ideal geometric space between the inner and outer shielding layers, enabling the shielding efficiency to remain stable even when the cable is compressed and deformed, achieving high-efficiency shielding across the entire frequency band. The shielding structure optimization algorithm based on the field strength decay coefficient precisely controls the material composition and geometric parameters of the inner and outer shielding layers, ensuring the maximization of shielding efficiency under the given thickness limit and effectively reducing the cable manufacturing cost. The real-time differential signal acquisition system for the inner and outer shielding layers enables the cable to have the ability to sense the electromagnetic environment, be able to monitor the characteristics of the interference source in real time, and provide accurate data support for subsequent interference suppression. The interference feature extraction method of joint time-frequency domain analysis realizes the precise identification and positioning of the interference source. Through the collaborative analysis of frequency domain and time domain characteristics, it can distinguish different types and sources of interference signals. The multi-modal adaptive interference identification and elimination algorithm is optimized for different frequency bands and interference characteristics, significantly improving the signal transmission quality without changing the physical structure and effectively compensating for signal reflection and crosstalk between the internal layers of the cable. The double-layer heterogeneous recurrent neural network interference prediction system separates the influence of inner-layer high-frequency interference and outer-layer low-frequency interference through hierarchical decoupling technology, solves the problem of cross-influence of interference sources in traditional methods, and performs outstandingly in an environment with complex electromagnetic interference.

[0041] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, systems, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0042] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0043] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for anti-interference of high-frequency coaxial cables, characterized in that, Including: Construct a radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and set a triangular support structure between the inner shielding layer and the outer shielding layer; Install electromagnetic induction probes on the triangular support structure to collect differential interference signals between the inner shielding layer and the outer shielding layer; Perform joint time-frequency domain analysis on the differential interference signals to obtain an interference source feature mapping spectrum; Execute multi-modal interference recognition and suppression according to the interference source feature mapping spectrum, generate an interference cancellation signal, and perform real-time elimination and feedback of electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference cancellation signal to obtain an interference cancellation feedback effect.

2. The high-frequency coaxial cable anti-interference method according to claim 1, wherein The construction of the radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and setting a triangular support structure between the inner shielding layer and the outer shielding layer includes: Perform cylindrical central positioning on the cable conductor to obtain a copper core conductor, and extrude and coat the copper core conductor with polytetrafluoroethylene material to obtain an insulating layer; Perform silver-copper alloy wire mesh weaving treatment on the insulating layer to obtain an inner shielding layer, and perform fluorinated ethylene propylene material coating treatment on the inner shielding layer to obtain an insulating isolation layer; Perform tin-copper alloy wire mesh weaving treatment on the insulating isolation layer to obtain an outer shielding layer; Manufacture a triangular support structure based on reinforced nylon material, and set the triangular support structure along the cable axis to obtain an electromagnetic isolation space between the inner shielding layer and the outer shielding layer.

3. The high-frequency coaxial cable anti-interference method according to claim 2, characterized in that The anti-interference method for the high-frequency coaxial cable further includes: Perform mathematical modeling on the electromagnetic field distribution of the high-frequency coaxial cable to obtain a model of the relationship between the electromagnetic field strength and the radial distance; Based on the model of the relationship between the electromagnetic field strength and the radial distance, calculate the attenuation coefficient of the inner shielding layer to obtain the inner shielding layer attenuation coefficient; Based on the model of the relationship between the electromagnetic field strength and the radial distance, calculate the attenuation coefficient of the outer shielding layer to obtain the outer shielding layer attenuation coefficient; Based on the inner shielding layer attenuation coefficient and the outer shielding layer attenuation coefficient, set the ratio of the inner and outer layer attenuation coefficients to obtain the attenuation coefficient ratio; According to the attenuation coefficient ratio, adjust the material composition of the inner shielding layer and the outer shielding layer to obtain an optimized material ratio of the inner and outer shielding layers; Input the optimized material ratio and geometric parameters of the inner and outer shielding layers into the electromagnetic field analysis system for iterative calculation to obtain the shielding structure design parameters.

4. The high-frequency coaxial cable anti-interference method according to claim 1, characterized in that The installation of electromagnetic induction probes on the triangular support structure to collect differential interference signals between the inner shielding layer and the outer shielding layer includes: Install electromagnetic induction probes on the triangular support structure to obtain a plurality of electromagnetic induction probes evenly distributed in the gap between the inner shielding layer and the outer shielding layer; Perform isolated coaxial lead connection on the plurality of electromagnetic induction probes to obtain a plurality of probe signal channels connected to the signal acquisition module, and perform digital conversion processing on the signals collected by the plurality of probe signal channels to obtain a plurality of digitized differential signals; Perform digital filtering and decimation processing on the multiple digitized differential signals to obtain multiple preliminarily processed differential signals; Perform window slicing and phase correction processing on the multiple preliminarily processed differential signals to obtain time-aligned multiplexed signals, and perform weighted average processing on the time-aligned multiplexed signals to obtain a differential interference signal.

5. The high-frequency coaxial cable anti-interference method according to claim 1, characterized in that, Perform joint time-frequency domain analysis on the differential interference signal to obtain an interference source feature mapping spectrum, including: Perform two-channel splitting on the differential interference signal to obtain a frequency domain analysis channel signal and a time domain analysis channel signal; Perform fast Fourier transform and non-linear frequency band division processing on the frequency domain analysis channel signal to obtain a frequency domain feature set, and perform multi-scale wavelet packet decomposition processing on the time domain analysis channel signal to obtain a time domain feature set; Perform collaborative feature extraction based on the frequency domain feature set and the time domain feature set to obtain a mixed domain feature matrix, and perform non-linear dimensionality reduction and clustering processing on the mixed domain feature matrix to obtain an interference feature fingerprint; Perform multi-level similarity calculation on the interference feature fingerprint and a pre-constructed interference source library to obtain an interference source feature mapping spectrum.

6. The high-frequency coaxial cable anti-interference method according to claim 1, characterized in that Perform multi-modal interference recognition and suppression according to the interference source feature mapping spectrum to generate an interference cancellation signal, and perform real-time elimination and feedback on the electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference cancellation signal to obtain an interference cancellation feedback effect, including: Classify the interference types according to the candidate interference sources and their confidence levels in the interference source feature mapping spectrum to obtain classified narrowband interference, broadband interference, and burst interference; Perform notch filter array processing on the narrowband interference to obtain a narrowband interference suppression signal; Perform wavelet domain threshold shrinkage processing on the broadband interference to obtain a broadband interference suppression signal; Perform prediction and correction modeling processing on the burst interference to obtain a burst interference suppression signal; Perform weighted fusion on the narrowband interference suppression signal, the broadband interference suppression signal, and the burst interference suppression signal according to the confidence levels in the interference source feature mapping spectrum to obtain an interference cancellation signal; Input the interference cancellation signal into a two-layer heterogeneous recurrent neural network for interference prediction processing to obtain an interference cancellation feedback effect.

7. The high-frequency coaxial cable anti-interference method according to claim 6, characterized in that, Perform prediction and correction modeling processing on the burst interference to obtain a burst interference suppression signal, including: Perform sample segmentation and normalization processing on the data of the first S milliseconds of the burst interference to obtain an interference sample set of multiple time windows, and perform stationarity test and white noise test on the interference sample set to obtain an interference time series characteristic analysis result; Perform autoregressive moving average modeling on the burst interference based on the interference time series characteristic analysis result to obtain an ARMA prediction model; Perform maximum likelihood estimation and Box-Jenkins iterative calculation on the ARMA prediction model to obtain a model coefficient matrix, and perform Akaike information criterion evaluation on the model coefficient matrix to obtain an optimal model parameter set; Substitute the optimal model parameter set into the ARMA prediction model, perform back substitution fitting and residual analysis on the interference data to obtain the model fitting accuracy evaluation result, and fine-tune the model parameters according to the model fitting accuracy evaluation result to obtain the interference waveform prediction function; Perform forward extrapolation calculation on the interference waveform prediction function to obtain the interference waveform prediction values within the next H milliseconds, and perform phase inversion and amplitude matching processing on the interference waveform prediction values to obtain the anti-phase cancellation signal; Input the anti-phase cancellation signal into a finite impulse response filter for smoothing processing to obtain the burst interference suppression signal.

8. The high-frequency coaxial cable anti-interference method according to claim 6, characterized in that, The inputting the interference cancellation signal into a double-layer heterogeneous recurrent neural network for interference prediction processing to obtain the interference cancellation feedback effect includes: Perform frequency band separation on the interference cancellation signal to obtain the high-frequency interference component of the inner shielding layer and the low-frequency interference component of the outer shielding layer; Input the high-frequency interference component of the inner shielding layer into the long short-term memory network in the double-layer heterogeneous recurrent neural network to capture the time series characteristics and obtain the interference prediction result of the inner shielding layer; Input the low-frequency interference component of the outer shielding layer into the gated recurrent unit network in the double-layer heterogeneous recurrent neural network to identify burst interference and obtain the interference prediction result of the outer shielding layer; Perform attention mechanism fusion on the interference prediction result of the inner shielding layer and the interference prediction result of the outer shielding layer to obtain the interference source spatial positioning information; Based on the interference source spatial positioning information, perform hierarchical decoupling processing on the interference signals of the inner and outer shielding layers to obtain the separated inner layer high-frequency interference influence and outer layer low-frequency interference influence; Perform interference intensity calculation based on the separated inner layer high-frequency interference influence and outer layer low-frequency interference influence to obtain the interference cancellation feedback effect.

9. A high-frequency coaxial cable anti-interference device, characterized in that, For implementing the high-frequency coaxial cable anti-interference method according to any one of claims 1-8, the high-frequency coaxial cable anti-interference device includes: A construction module for constructing a radially progressive impedance gradient double-layer shielding structure with an inner shielding layer and an outer shielding layer, and arranging a triangular support frame structure between the inner shielding layer and the outer shielding layer; An acquisition module for installing electromagnetic induction probes on the triangular support frame structure to acquire the differential interference signals between the inner shielding layer and the outer shielding layer; A joint analysis module for performing time-frequency domain joint analysis on the differential interference signals to obtain the interference source feature mapping atlas; A feedback module for performing multi-modal interference identification and suppression according to the interference source feature mapping atlas, generating an interference cancellation signal, and performing real-time elimination and feedback on the electromagnetic interference in the inner shielding layer and the outer shielding layer based on the interference cancellation signal to obtain the interference cancellation feedback effect.

Images