High-frequency carrier synchronization signal modulation system

Abstract

The invention relates to the technical field of high-frequency signal modulation, and discloses a high-frequency carrier synchronization signal modulation system. A carrier distortion compensation module of the system constructs a model based on a historical transmission data set, captures a phase jitter parameter, a spectrum leakage component and a modulation pulse sequence of a transmitting end carrier in real time, and outputs a reconstructed baseband parameter; the multi-dimensional distortion analysis module compares the reconstructed baseband parameter with a receiving end demodulation baseband parameter through composite difference detection, and generates a channel-level distortion coefficient tensor; the topological positioning network module is used for positioning signal out-of-step physical nodes and generating a probability distribution thermodynamic diagram by combining impedance characteristics of transmission nodes and signal path delay information; and the adaptive modulation strategy module starts a multi-band carrier injection mode for the high-probability out-of-step nodes according to the abnormal probability gradient of the thermodynamic diagram, and applies a phase disturbance test to adjacent transmission links. The system can improve the adaptability and accuracy of high-frequency carrier signal synchronous modulation.

Description

Technical Field

[0001] The present invention relates to the technical field of high-frequency signal modulation, in particular to a high-frequency carrier synchronization signal modulation system. Background Art

[0002] In fields such as communications, radar, and industrial control, high-frequency carrier signals serve as the core carrier of information transmission. Their synchronization performance is directly related to the integrity and reliability of data transmission. As transmission rates and frequency bands increase, carrier signal distortion becomes increasingly prominent. Phenomena such as phase jitter, spectrum leakage, and pulse sequence offset often cause signal synchronization errors between the transmitter and receiver, leading to demodulation errors and data loss.

[0003] Traditional carrier synchronization systems often employ single-dimensional distortion compensation mechanisms, such as adjusting the carrier phase solely through a phase-locked loop or relying solely on filters to suppress spectral leakage. These approaches often overlook the interdependencies among distortion factors—phase jitter can overlap with spectral leakage, while pulse sequence time offsets can amplify the effects of these distortions, limiting compensation effectiveness. Furthermore, existing systems lack in-depth utilization of historical transmission data when addressing distortion, making it difficult to develop adaptive compensation models. This can lead to compensation lags or overcompensation in complex channel environments.

[0004] Traditional distortion analysis methods focus on single-dimensional detection in the time or frequency domains. For example, they only calculate the phase cumulative error in the time domain or analyze the energy distribution distortion in the frequency domain. This single-dimensional analysis method fails to fully reflect the distortion characteristics at the channel level, making the generated distortion parameters difficult to support subsequent precise positioning and modulation strategy adjustments.

[0005] Existing technologies for locating carrier desynchronization often rely on simple path delay measurements, ignoring the impact of transmission node impedance characteristics on signal propagation. Fluctuations in node impedance can cause signal reflections and attenuation, altering the distribution of desynchronization paths. Locating the desynchronized physical node based solely on delay information often results in ambiguity and large errors, complicating troubleshooting and system maintenance.

[0006] Existing adaptive modulation strategies often use fixed patterns, such as applying uniform modulation parameters to all transmission links. This lacks differentiated handling of desynchronization risks. When some nodes exhibit a high probability of desynchronization, these rigid modulation strategies cannot be adjusted in a timely manner, making it difficult to prevent further deterioration in signal synchronization performance, hindering the stable operation of the entire system in complex environments. Summary of the Invention

[0007] The object of the present invention is to provide a high-frequency carrier synchronization signal modulation system to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides a high-frequency carrier synchronization signal modulation system, the system comprising: Carrier distortion compensation module: Builds a carrier distortion compensation model based on historical transmission data sets, captures the phase jitter parameters, spectrum leakage components, and modulated pulse sequence of the transmitting end carrier signal in real time, and reconstructs the baseband parameters through the output of the carrier distortion compensation model; Multi-dimensional distortion analysis module: performs composite difference detection on the reconstructed baseband parameters and the demodulated baseband parameters at the receiving end. The composite difference detection includes time domain phase accumulation offset, frequency domain distortion energy distribution and pulse sequence time alignment to generate a channel-level distortion coefficient tensor; Topology positioning network module: inputs the distortion coefficient tensor into the synchronization path topology network, combines the transmission node impedance characteristics and signal path delay information, generates a probability distribution heat map of the carrier out-of-sync path, and locates the physical node where the signal is out of sync; Adaptive modulation strategy module: configures a synchronous modulation strategy according to the abnormal probability gradient of the probability distribution heat map, including: enabling a multi-band carrier injection mode for nodes with high probability of out-of-sync, and applying a phase perturbation test to adjacent transmission links.

[0009] Preferably, the carrier distortion compensation module specifically includes: Historical feature extraction unit: This unit performs multi-scale decomposition on historical transmission data sets, including using empirical mode decomposition to separate the energy proportions of the steady-state phase component and transient noise component of the carrier signal, establishing a correlation matrix between spectrum leakage components and channel signal-to-noise ratio through spectral correlation analysis, and using a dynamic time warping algorithm to align modulated pulse train patterns at different bandwidths. The distortion compensation model construction unit inputs the processed historical transmission data into a hybrid compensation network. The hybrid compensation network includes: a gated cyclic timing prediction unit based on the channel aging curve to generate reference carrier parameters; a fully convolutional network embedded with a distortion attention mechanism to correct phase deviation caused by spectrum leakage; and a pulse train compensator that dynamically adjusts compensation weights based on the modulated pulse train captured in real time. Baseband parameter generation unit: The multi-channel acquisition unit deployed at the RF front end synchronously acquires the phase jitter variance of the carrier signal, the 3rd to 40th harmonic components and phase offset angle of the spectrum leakage, the amplitude jump gradient and time interval entropy of the modulated pulse sequence, and inputs the real-time acquired data into the carrier distortion compensation model to output the reconstructed baseband parameters.

[0010] Preferably, the process of generating the reconstructed baseband parameters includes executing: Environmental interference filtering: Adaptive filters are designed based on the temperature and humidity parameters of the transmission environment to eliminate phase noise caused by multipath effects. Multi-domain feature fusion operation: Integrate the time-frequency correlation features of carrier phase, spectral components and pulse sequences through cross-domain correlation algorithms; Dynamic parameter output: Generates reconstructed baseband parameters containing normal fluctuation thresholds, and the reconstructed baseband parameters are dynamically updated with the channel aging coefficient.

[0011] Preferably, the multi-dimensional distortion analysis module specifically includes: Time domain phase offset calculation unit: performs sliding comparison between the reconstructed baseband parameters and the demodulated baseband parameters in a preset time window, uses a dynamic time warping algorithm to align the asynchronous sampling sequences, calculates the cumulative phase offset in each time window, and generates a time domain offset vector; Frequency domain distortion detection unit: performs frequency band energy decomposition on the spectrum leakage components of the reconstruction parameters and demodulation parameters, extracts the energy density ratio within the preset sensitive frequency band, quantifies the distortion index of each harmonic, and constructs the frequency domain distortion vector; Pulse sequence alignment evaluation unit: Based on the structural similarity algorithm, it matches the spatiotemporal distribution characteristics of the reconstructed and demodulated pulse sequences, calculates the time synchronization error of the pulse amplitude jump point, quantifies the Jensen-Shannon divergence of the pulse interval distribution, and generates a pulse alignment vector; Distortion coefficient tensor generation unit: The time domain offset vector, frequency domain distortion vector and pulse alignment vector are fused into a three-dimensional tensor, and the dimension difference is eliminated through normalization processing weighted by the characteristic entropy value, and the output dimension is a third-order distortion coefficient tensor of [channel number × timestamp × distortion type].

[0012] Preferably, the frequency domain distortion detection unit performs the following steps: In the frequency domain decomposition stage, the corresponding spectral leakage components in the reconstruction and demodulation baseband parameters are extracted, and multi-resolution frequency band segmentation is performed on each group of harmonic signals using wavelet packet transform to extract the energy distribution characteristics within a preset interference frequency band interval; the interference frequency band interval covers the harmonic resonance range of typical communication equipment; after the extraction is completed, the energy spectral density of the reconstructed and demodulated data within the interference frequency band interval is quantified respectively, and the distortion index of each order harmonic is generated based on the relative distortion degree, and the distortion results of all orders are summarized to construct a frequency domain distortion vector.

[0013] Preferably, in the pulse sequence alignment evaluation unit, the structural similarity algorithm adopts Hausdorff distance metric to perform position matching on the jump point sets of the two pulse sequences to identify time synchronization errors.

[0014] Preferably, the topology positioning network module specifically includes: Transmission topology modeling unit: Constructs a transmission node connection topology diagram based on signal path delay information, annotates the impedance characteristic parameters between nodes, superimposes the phase delay constraints of multipath reflection nodes on the topology diagram, and generates a transmission topology model including an impedance matrix and a node admittance matrix; The out-of-sync path deduction unit maps the distortion coefficient tensor to the corresponding nodes of the transmission topology model; performs out-of-sync path simulation based on a graph convolutional network, which includes: calculating the attenuation factor of the out-of-sync signal based on the node impedance characteristics; capturing cross-node out-of-sync correlation features through a multi-head graph attention mechanism; and using a random walk algorithm to simulate the propagation trajectory of the out-of-sync signal in the topological network; Heat map generation unit: Counts the frequency of out-of-step signals on each path in the simulated propagation, calculates the out-of-step residence probability value based on the impedance characteristic parameters, generates an out-of-step probability distribution heat map covering the entire network, and marks the set of suspicious paths whose probability values ​​exceed the preset out-of-step threshold; Physical node positioning unit: performs density cluster analysis on the probability distribution heat map to identify the out-of-sync probability cluster area; and delineates the out-of-sync physical node boundary based on the transmission node location and topological connection relationship.

[0015] Preferably, the topology positioning network module further includes outputting suspicious node identification and main out-of-sync path, wherein: Suspicious node identification: Based on the transmission nodes associated with the suspicious path set, the nodes are bound to the physical device addresses to generate a suspicious node identification list; Main out-of-sync path: During the random walk simulation, the node path sequence of each round of propagation is recorded, the frequency of occurrence of all paths is counted, the path with the highest frequency is selected as the main out-of-sync path, and an ordered node number sequence is output.

[0016] Preferably, the adaptive modulation strategy module specifically includes: Carrier injection strategy unit: When the out-of-sync probability of a node in the probability distribution heat map exceeds a preset threshold, it sends a modulation instruction to the RF unit associated with the node to execute the following: increase the carrier sampling rate to 8-12 times the base rate; synchronously enable the harmonic phase tracking mode to capture phase jump events of the 3rd to 40th harmonics; and deploy a transient response recorder on the terminal to store phase jump waveform fragments; Phase perturbation test unit: applies a multi-band phase perturbation sequence to the adjacent node links with the largest out-of-sync probability gradient in the probability distribution heat map.

[0017] Preferably, the phase disturbance test unit execution process includes: Disturbance signal injection operation: inject 0.5-5kHz swept frequency phase disturbance signal through controllable phase modulator; Theoretical response calculation operation: Calculate the theoretical phase response spectrum based on the transmission topology model and impedance matrix; Measured response acquisition operation: collect the node phase response after the disturbance signal is injected and generate the measured phase response spectrum; Abnormal judgment operation: Calculate the Manhattan distance between the theoretical phase response spectrum and the measured phase response spectrum. When the distance value exceeds the preset offset threshold, it is marked as a phase abnormality-related node.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This high-frequency carrier synchronization signal modulation system forms a complete distortion processing and synchronization optimization mechanism through the synergistic effect of four modules. The carrier distortion compensation module builds a model based on historical transmission data sets, capable of capturing the phase jitter parameters, spectral leakage components, and modulated pulse sequences of the transmitting carrier signal in real time. The fusion of multi-dimensional parameters makes the reconstructed baseband parameters more consistent with actual transmission scenarios, avoiding the information loss caused by single-parameter processing. The introduction of historical data sets allows the model to learn distortion patterns from past transmission cases, making the reconstruction process adaptable to complex channel environments.

[0019] The multi-dimensional distortion analysis module performs a composite difference detection between the reconstructed baseband parameters and the demodulated baseband parameters at the receiving end, covering the cumulative phase offset in the time domain, the energy distribution of the distortion in the frequency domain, and the time alignment of the pulse sequence. The resulting channel-level distortion coefficient tensor breaks through the limitations of traditional single-dimensional analysis. This composite detection method comprehensively characterizes the interaction and superposition effects of various distortion types in the channel, making the description of distortion characteristics closer to the actual transmission state, providing a richer basis for subsequent desynchronization location and modulation strategy adjustment.

[0020] The topology positioning network module inputs the distortion coefficient tensor into the synchronization path topology network and combines the transmission node impedance characteristics with signal path delay information to generate a probability distribution heat map, thereby locating the physical nodes along the carrier desynchronization path. The transmission node impedance characteristics directly affect the signal propagation characteristics, while path delay reflects the time loss of signal transmission. The combination of these two makes the probability distribution of the desynchronization path more consistent with the actual physical link, making the location of the desynchronization node more accurate and reducing the need for blind troubleshooting.

[0021] The adaptive modulation strategy module configures its strategy based on the abnormal probability gradient of the probability distribution heat map. It enables multi-band carrier injection for nodes with a high probability of desynchronization. This can mitigate distortion accumulation in a single frequency band through the complementary effects of multi-band signals. Phase perturbation testing of adjacent transmission links further clarifies the mutual influence between links through perturbation feedback, making modulation strategy adjustments more targeted. This differentiated modulation approach avoids the inadequate adaptability of a unified strategy to nodes with different desynchronization risks, allowing the system to flexibly adjust based on the actual desynchronization situation and improving its ability to maintain synchronization in complex transmission environments.

[0022] The coordinated operation of various modules, from source compensation and comprehensive analysis of distortion to precise positioning of desynchronized nodes and flexible adaptation of modulation strategies, forms a closed-loop optimization system. This enables the high-frequency carrier synchronization signal modulation system to have more comprehensive processing capabilities when facing complex problems such as phase jitter, spectrum leakage, and node desynchronization, and can maintain the stability and reliability of signal synchronization in various channel environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a working principle diagram of the high-frequency carrier synchronization signal modulation system of the present invention; Figure 2 This is the design diagram of the multi-dimensional distortion analysis module; Figure 3 Flowchart of the frequency domain distortion detection unit; Figure 4 Flowchart for positioning network modules for topology. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] See also Figure 1 The present invention provides a high-frequency carrier synchronization signal modulation system, the system comprising: The carrier distortion compensation module constructs a carrier distortion compensation model based on the historical transmission data set, captures the phase jitter parameters, spectrum leakage components and modulated pulse sequence of the transmitting end carrier signal in real time, and reconstructs the baseband parameters through the output of the carrier distortion compensation model.

[0026] The multi-dimensional distortion analysis module performs composite difference detection on the reconstructed baseband parameters and the demodulated baseband parameters at the receiving end. The composite difference detection includes time domain phase accumulation offset, frequency domain distortion energy distribution and pulse sequence time alignment to generate a channel-level distortion coefficient tensor.

[0027] The topology positioning network module inputs the distortion coefficient tensor into the synchronous path topology network, combines the transmission node impedance characteristics and the signal path delay information, generates a probability distribution heat map of the carrier out-of-sync path, and locates the physical node where the signal is out of sync.

[0028] The adaptive modulation strategy module configures a synchronous modulation strategy according to the abnormal probability gradient of the probability distribution heat map, including: enabling a multi-band carrier injection mode for nodes with high probability of out-of-sync, and applying a phase disturbance test to adjacent transmission links.

[0029] Example 1: See Figure 2 The operation of the carrier distortion compensation module relies on the collaborative work of the historical feature extraction unit, the distortion compensation model construction unit and the baseband parameter generation unit. Each unit processes data and outputs results according to a specific process.

[0030] When the historical feature extraction unit performs multi-scale decomposition on the historical transmission data set, it first uses the empirical mode decomposition method to decompose the carrier signal into multiple intrinsic mode functions. By calculating the energy contribution of each function, the steady-state phase component and the transient noise component are separated. In this process, the steady-state phase component reflects the stable characteristics of the carrier signal under normal transmission conditions, while the transient noise component includes signal fluctuations caused by various types of sudden interference. Subsequently, through spectral correlation analysis, the correlation coefficient between the spectrum leakage component and the channel signal-to-noise ratio is calculated, and a correlation matrix is ​​established between the two. The element values ​​in the matrix reflect the corresponding relationship between different levels of spectrum leakage and signal-to-noise ratio levels. At the same time, the dynamic time warping algorithm is used to process the modulated pulse sequence patterns under different bandwidths. By stretching or compressing the time axis of the sequence, the pulse sequences under different bandwidths are morphologically aligned.

[0031] The distortion compensation model construction unit inputs the historical transmission data that has undergone the above processing into the hybrid compensation network. The hybrid compensation network contains three key components. The gated cyclic timing prediction unit based on the channel aging curve generates reference carrier parameters according to the time-varying patterns of the channel parameters recorded in the historical data. These parameters serve as the ideal reference standard for the carrier signal. The fully convolutional network embedded with the distortion attention mechanism focuses on correcting the phase deviation caused by spectrum leakage. By learning the characteristic patterns of spectrum leakage, the network gives higher attention weight to the phase deviation areas related to spectrum leakage during the processing process, thereby adjusting the phase more accurately. The pulse sequence compensator can receive the modulated pulse sequence captured in real time and dynamically adjust the compensation weight according to the characteristic changes of the sequence. When the pulse sequence has an amplitude jump or an abnormal time interval, the compensation weight will change accordingly to adapt to these situations.

[0032] The baseband parameter generation unit operates through a multi-channel acquisition unit deployed in the RF front-end. This acquisition unit synchronously acquires multiple real-time parameters of the carrier signal, including phase jitter variance, which reflects the discreteness of phase jitter; the 3rd to 40th harmonic components of spectrum leakage and the corresponding phase offset angles of each harmonic; the amplitude transition gradient of the modulated pulse sequence, that is, the rate of change of the amplitude of adjacent pulses; and the time interval entropy, which describes the degree of disorder in the pulse sequence time interval. Before this real-time data is input into the carrier distortion compensation model, it undergoes environmental interference filtering, multi-domain feature fusion, and dynamic parameter output.

[0033] In the environmental interference filtering phase, an adaptive filter is designed based on the temperature and humidity parameters of the transmission environment. The filter parameters automatically adjust with changes in temperature and humidity. When temperature and humidity fluctuations increase multipath effects, the filter coefficients change accordingly, effectively eliminating phase noise caused by multipath. Multi-domain feature fusion is performed using a cross-domain correlation algorithm. This algorithm analyzes the time-domain variations of the carrier phase, the frequency-domain distribution of spectral components, and the time-series characteristics of the pulse sequence, exploiting the time-frequency correlation between the three and integrating these discrete features into a unified feature set. The reconstructed baseband parameters generated in the dynamic parameter output phase include normal fluctuation thresholds. These thresholds are not fixed but are dynamically updated based on the channel aging coefficient, which is calculated based on factors such as channel usage time and transmitted data volume. As the coefficient increases, indicating increased channel aging, the normal fluctuation thresholds are adjusted accordingly to adapt to the changing channel state.

[0034] Example 2: See Figure 3 The operation of the multi-dimensional distortion analysis module depends on the sequential collaboration of the time domain phase offset calculation unit, the frequency domain distortion detection unit, the pulse sequence alignment evaluation unit and the distortion coefficient tensor generation unit. Each unit processes the data according to the established logic and forms the final result.

[0035] When the time-domain phase offset calculation unit is in operation, it first sets a preset time window. The length of this time window is determined based on the baud rate of the signal transmission and the channel characteristics to ensure that it can completely cover one or more modulation cycles. Subsequently, a sliding comparison is performed on the reconstructed baseband parameters and the demodulated baseband parameters. During the sliding process, the time window moves along the time axis at a fixed step size, and the two sets of parameters within the window are compared once each movement. Because the sampling clocks of the transmitter and receiver may deviate, resulting in asynchronous sampling sequences, a dynamic time warping algorithm is used to align the asynchronous sampling sequences. This algorithm stretches or compresses the time axis of one set of sequences to make the two sets of sequences match as much as possible in terms of morphology, and then calculates the cumulative phase offset within each time window. The phase cumulative offset is calculated based on the aligned sequences. By comparing the phase values ​​at corresponding moments, the total offset within the entire time window is accumulated, and finally a time-domain offset vector is generated. Each element in the vector corresponds to the cumulative phase offset result of a time window.

[0036] The frequency domain distortion detection unit processes the spectral leakage components of the reconstruction and demodulation parameters. During the frequency domain decomposition stage, the corresponding spectral leakage components of the reconstructed and demodulated baseband parameters are first extracted. These components contain harmonic components generated during signal transmission due to channel non-idealities. Multi-resolution frequency band segmentation is then performed on each set of harmonic signals using the wavelet packet transform. The wavelet packet transform decomposes the signal into different frequency subbands, enabling detailed analysis of signals in different frequency bands. The energy distribution characteristics within a pre-set interference frequency band are extracted. This interference frequency band is determined based on the operating frequency range and harmonic resonance characteristics of typical communication equipment and covers the frequency range likely to cause strong interference. After extraction, the energy spectral density of the reconstructed and demodulated data within this interference frequency band is quantified. The energy spectral density is quantified by calculating the signal energy per unit frequency. Distortion indices for each harmonic order are generated based on the relative distortion level. The relative distortion level is determined by the ratio of the energy spectral density difference between the demodulated and reconstructed data to the energy spectral density of the reconstructed data. The distortion indices for all orders are aggregated to construct a frequency domain distortion vector.

[0037] The work of the pulse sequence alignment assessment unit revolves around the spatiotemporal distribution characteristics of the reconstructed and demodulated pulse sequences. Based on the structural similarity algorithm, the waveform morphology, amplitude changes, and occurrence time of the two pulse sequences are matched to identify the corresponding pulse signals. The time synchronization error of the pulse amplitude jump point is calculated, that is, the difference in the occurrence time of the same pulse jump event in the reconstructed sequence and the demodulated sequence. This difference reflects the deviation in the time alignment of the two sequences. At the same time, the Jensen-Shannon divergence of the pulse interval distribution is quantified. The Jensen-Shannon divergence is used to measure the difference between two probability distributions. By calculating the divergence value of the pulse time interval distribution in the two sequences, the inconsistency in the interval regularity of the pulse sequence can be reflected. The time synchronization error is integrated with parameters such as the Jensen-Shannon divergence to generate a pulse alignment vector. The elements in this vector reflect the degree of alignment of the two pulse sequences from different angles.

[0038] The distortion coefficient tensor generation unit receives the time offset vector from the time-domain phase offset calculation unit, the frequency distortion vector from the frequency distortion detection unit, and the pulse alignment vector from the pulse sequence alignment assessment unit, and performs a three-dimensional tensor fusion on these three vectors. During this fusion process, the time offset, frequency distortion, and pulse alignment are treated as three distinct dimensions, with the values ​​in each dimension corresponding to corresponding feature parameters. Dimensional differences are eliminated through a normalization process weighted by feature entropy. The feature entropy value is determined based on the information content of each feature parameter in the historical data; feature parameters with greater information content receive higher weights. Normalization maps the values ​​of each feature parameter to the same range, preventing dimensional differences from affecting subsequent analysis. The final output is a third-order distortion coefficient tensor of the dimensions [channel number × timestamp × distortion type]. The channel number distinguishes different transmission channels, the timestamp corresponds to the specific sampling time, and the distortion type encompasses various distortion features related to the time domain, frequency domain, and pulse sequence. This tensor fully presents the various distortion conditions across different channels at different times.

[0039] Example 3: See Figure 4 When the pulse sequence alignment evaluation unit is running, the Hausdorff distance metric is used as the core of the structural similarity algorithm to match the positions of the jump point sets of the two pulse sequences. Specifically, the time coordinates of all amplitude jump points are first extracted from the reconstructed pulse sequence and the demodulated pulse sequence to form two independent point sets. By calculating the Hausdorff distance between the two point sets, the matching degree of the two sets of jump points on the time axis is quantified, and then the time synchronization error is identified. When the Hausdorff distance value is small, it indicates that the position distribution of the two sets of jump points is relatively close and the time synchronization is good; when the distance value increases, it means that there is a more obvious time offset.

[0040] The topology positioning network module begins with the transmission topology modeling unit. This unit collects information about the path delays of signals transmitted between nodes. This information is obtained by inserting timestamps during transmission and recording the time of reception. Based on this path delay information, a topology diagram of transmission node connections is constructed. Each node in the diagram represents a signal transmission or forwarding device, and the lines between nodes represent the actual signal transmission links. The impedance characteristic parameters of each link are annotated. These parameters, including resistance, inductance, and capacitance, are pre-acquired during node deployment using specialized impedance measurement equipment. Phase delay constraints for multipath reflection nodes are superimposed on the topology diagram. Multipath reflection nodes are equivalent nodes formed by obstacles or reflective surfaces encountered during signal transmission. Their phase delay is calculated based on the ratio of the reflection path length to the signal wavelength. Combining this information, a transmission topology model is generated, consisting of an impedance matrix and a node admittance matrix. Elements in the impedance matrix represent the equivalent impedance between two nodes, while the node admittance matrix reflects the signal conduction capability of each node.

[0041] The out-of-sync path deduction unit maps the distortion coefficient tensor to the corresponding nodes in the transmission topology model. This mapping process is based on the correspondence between the channel numbers in the tensor and the node numbers in the topology model, so that each node is associated with various distortion data for that channel at different timestamps. Out-of-sync path simulation is performed based on a graph convolutional network. During the simulation, the attenuation factor of the out-of-sync signal is first calculated based on the node impedance characteristics. The attenuation factor is proportional to the modulus of the node impedance; the larger the impedance, the greater the attenuation of the signal after passing through the node. A multi-head graph attention mechanism is used to capture cross-node out-of-sync correlation characteristics. This mechanism uses multiple attention heads to calculate the correlation weights between different nodes and then integrates these weights to reflect the complex interactions between nodes. A random walk algorithm is used to simulate the propagation trajectory of the out-of-sync signal in the topological network. During the random walk, starting from any node, the probability of the next jump is determined based on the out-of-sync probability of the adjacent nodes and the attenuation factor of the connecting link. This process is repeated until a preset number of steps are reached or until the starting node is returned, thus simulating the possible propagation path of the out-of-sync signal.

[0042] The heat map generation unit counts the frequency of out-of-step signals on each path in the simulated propagation. The frequency statistics are based on the results of multiple random walk simulations. The number of times each path is recorded is its frequency of occurrence. The out-of-step residence probability value is calculated in combination with the impedance characteristic parameters. The calculation formula is:

[0043] in, represents the out-of-step dwell probability value, Indicates the frequency of occurrence of a path. represents the product of the impedance moduli of all nodes on the path, represents the total number of possible paths, Indicates the The frequency of occurrence of the path, Indicates the The product of the impedance modulus values ​​of all nodes on the path. Based on the calculated out-of-step residence probability value, a heat map of the out-of-step probability distribution covering the entire network is generated. Different colors in the heat map represent different probability values. At the same time, a set of suspicious paths with probability values ​​exceeding the preset out-of-step threshold is marked. The preset out-of-step threshold is determined based on the probability distribution characteristics of historical out-of-step events.

[0044] The physical node positioning unit performs density cluster analysis on the probability distribution heat map. By calculating the probability density of each region in the heat map, it identifies areas with higher density as areas of clustered synchronization probability. Based on the locations of transmission nodes and topological connectivity, it delineates the boundaries of physical nodes with synchronization loss. This boundary delineation focuses on the nodes within the clustered area and, combined with the actual physical distance between nodes and the relevance of the connecting links, determines the range of physical nodes affected by the synchronization loss.

[0045] When outputting results, the topology positioning network module generates a list of suspicious node identifiers and the main out-of-sync path. The list of suspicious node identifiers is generated based on the transmission nodes associated with the suspicious path set. The logical number of each node is bound to the corresponding physical device address, which includes the device's MAC address or IP address, so that each identifier in the list can correspond to a specific hardware device. The main out-of-sync path is determined during the random walk simulation process. The node path sequence of each round of propagation is recorded, the frequency of occurrence of all paths is counted, and the path with the highest frequency of occurrence is selected as the main out-of-sync path. The ordered node number sequence of the path is output. The numbers in the sequence are arranged according to the order of the nodes in the path to clearly show the most likely propagation path of the out-of-sync signal.

[0046] Example 4: The operation of the adaptive modulation strategy module relies on the coordinated operation of the carrier injection strategy unit and the phase disturbance test unit. The two units respectively perform corresponding modulation actions according to the analysis results of the probability distribution heat map.

[0047] The carrier injection strategy unit continuously monitors the out-of-sync probability value of each node in the probability distribution heat map. When the out-of-sync probability of a node exceeds the preset threshold, it immediately sends a modulation instruction to the RF unit associated with the node. After receiving the instruction, the RF unit first increases the carrier sampling rate to 8-12 times the base rate. The specific value of the sampling rate is determined by the transmission bandwidth and signal complexity of the node. The larger the bandwidth and the more complex the signal modulation method, the higher the sampling rate increase multiple to ensure that more signal details can be captured. The harmonic phase tracking mode is synchronously enabled. This mode uses a dedicated phase detection circuit to track the phase changes of the 3rd to 40th harmonics in real time. When the harmonic phase suddenly changes, the time point of the mutation and the amount of phase change are immediately recorded. These mutation events often reflect the instability of the node's internal circuit or the intrusion of external interference. A transient response recorder is deployed at the terminal. This recorder uses a high-speed storage chip and can store waveform fragments at the moment of phase jump with microsecond time accuracy. The storage length of the waveform fragment is automatically adjusted according to the duration of the jump, usually covering the signal from 10 cycles before the jump to 20 cycles after the jump, providing raw data for subsequent analysis of the cause of the jump.

[0048] The phase perturbation test unit targets adjacent node links with the largest desynchronization probability gradient in the probability distribution heat map. The desynchronization probability gradient is calculated by calculating the ratio of the difference in desynchronization probability between adjacent nodes to the physical distance between the nodes. A larger gradient indicates a more dramatic change in the desynchronization risk between the two nodes. A multi-band phase perturbation sequence is applied to these links. This perturbation sequence is designed based on the transmission characteristics of the link and consists of multiple sinusoidal phase-modulated signals of different frequencies. The frequency spacing of each signal is determined by the link's resonant frequency to prevent mutual interference between the perturbation signals.

[0049] The following is a table showing the parameter configuration and response records when a phase perturbation test is applied to a link between adjacent nodes:

[0050] In the table, the perturbation signal frequency ranges from 0.5 to 5 kHz, with typical values ​​selected at even intervals. The perturbation amplitude is set based on the impedance matching of the link; the better the impedance matching, the greater the perturbation amplitude that can be applied. The perturbation duration decreases with increasing frequency because high-frequency signals decay more quickly, requiring shorter test times to minimize the impact on normal communications. Link response delay refers to the time interval from the injection of the perturbation signal to the detection of the response signal at the receiver. This delay includes the signal's transmission time in the link and the node's processing time. The response phase deviation is the difference between the actual phase change detected by the receiver and the theoretically calculated value. The magnitude of this difference reflects the link's sensitivity to phase perturbations.

[0051] The carrier injection strategy unit and the phase perturbation test unit operate in conjunction with each other. Carrier injection provides high-resolution signal sampling data, while the phase perturbation test actively stimulates potential link problems. Together, they can more accurately pinpoint the fault type of the out-of-sync node and provide a basis for subsequent modulation strategy adjustments. During the test, all parameter configurations and response data are uploaded to the system control center in real time. The control center processes the data in real time using dedicated data analysis software. If a parameter is found to be out of range, the corresponding alarm mechanism is automatically triggered, prompting the operator to conduct further inspection.

[0052] Example 5: The operation of the phase disturbance test unit is gradually advanced according to the process of disturbance signal injection operation, theoretical response calculation operation, measured response acquisition operation and abnormality judgment operation. Each operation link is closely connected to form a complete test closed loop.

[0053] The disturbance signal injection operation is achieved through a controllable phase modulator, which is deployed at the signal transmitting end of the adjacent node link and can accurately control the phase change of the output signal. The frequency range of the injected swept phase disturbance signal is set at 0.5-5kHz, and the sweep method adopts linear sweep, that is, the signal frequency starts from 0.5kHz and gradually increases to 5kHz at a fixed frequency change rate. The sweep period is determined according to the transmission delay characteristics of the link to ensure that the disturbance signal can completely cover all possible resonant frequency points of the link. The phase change amplitude of the disturbance signal is adjusted according to the impedance characteristics of the link. For links with higher impedance, the phase change amplitude is appropriately reduced to avoid excessive signal distortion; for links with lower impedance, the amplitude can be increased to enhance the disturbance effect. During the injection process, a synchronous trigger signal is used to ensure that the disturbance signal is synchronized with the normal carrier signal transmitted in the link in time to avoid test errors caused by time misalignment.

[0054] The theoretical response calculation is based on the transmission topology model and the impedance matrix. The transmission topology model includes parameters such as the connection relationship between each node in the link, the transmission path length, and the signal attenuation coefficient. The impedance matrix records the equivalent impedance values ​​between nodes at different frequencies. Based on the frequency variation curve of the injected swept-frequency phase disturbance signal and the path delay parameters in the topology model, the phase delay of the signal as it propagates in the link is calculated. Using the data in the impedance matrix, the reflection coefficient and transmission coefficient of the signal at different frequencies are calculated. The reflection coefficient reflects the degree of reflection of the signal at the node, while the transmission coefficient indicates the attenuation of the signal after passing through the node. Taking the phase delay, reflection coefficient, and transmission coefficient into consideration, and deriving them through the transmission line equation in circuit theory, a theoretical phase response spectrum is generated. This response spectrum uses frequency as the horizontal axis and phase change as the vertical axis, showing the phase response law of the link to disturbance signals of different frequencies under ideal conditions.

[0055] The measured response acquisition operation is performed at the receiving end of the link, and a high-precision phase meter is used to collect the node phase response after the disturbance signal is injected in real time. The sampling frequency of the phase meter is set to more than 20 times the highest frequency of the disturbance signal to meet the requirements of the Nyquist sampling theorem and ensure that the phase change curve can be accurately restored. During the acquisition process, the timestamp information is recorded synchronously to keep the phase response data and the frequency change of the disturbance signal in time. The collected raw phase data is filtered to remove the high-frequency noise generated by the environmental electromagnetic interference. The filtering adopts the sliding average filtering method. The window size is determined according to the frequency characteristics of the noise. Usually 5-10 sampling points are selected as a window. The processed phase data is arranged in frequency order to generate the measured phase response spectrum. The spectrum also uses frequency as the horizontal axis and phase change as the vertical axis to intuitively display the actual phase response of the link to the disturbance signal.

[0056] The abnormality determination operation is achieved by comparing the theoretical phase response spectrum with the measured phase response spectrum. The Manhattan distance between the two is calculated. This distance is obtained by taking the absolute difference between the theoretical phase value and the measured phase value at each frequency point, and then accumulating the absolute difference of all frequency points. A preset offset threshold is set. The threshold is determined based on the historical test data and stability requirements of the link. For links with higher communication quality requirements, the threshold is set to a smaller value; for links that allow a certain degree of phase deviation, the threshold can be appropriately increased. When the calculated Manhattan distance value exceeds the preset offset threshold, the two nodes connected by the link are marked as phase abnormality associated nodes, indicating that there are factors in the link that cause phase response abnormality, which may be node hardware failure, poor contact of the connection line, or strong external interference. After the marking is completed, the number of the abnormal associated node and the corresponding frequency range information are recorded in the system log to provide a direction for subsequent troubleshooting.

[0057] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0058] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A high-frequency carrier synchronization signal modulation system, characterized in that: include: Carrier distortion compensation module: Builds a carrier distortion compensation model based on historical transmission data sets, captures the phase jitter parameters, spectrum leakage components, and modulated pulse sequence of the transmitting end carrier signal in real time, and reconstructs the baseband parameters through the output of the carrier distortion compensation model; Multi-dimensional distortion analysis module: performs composite difference detection on the reconstructed baseband parameters and the demodulated baseband parameters at the receiving end. The composite difference detection includes time domain phase accumulation offset, frequency domain distortion energy distribution and pulse sequence time alignment to generate a channel-level distortion coefficient tensor; Topology positioning network module: inputs the distortion coefficient tensor into the synchronization path topology network, combines the transmission node impedance characteristics and signal path delay information, generates a probability distribution heat map of the carrier out-of-sync path, and locates the physical node where the signal is out of sync; Adaptive modulation strategy module: configures a synchronous modulation strategy according to the abnormal probability gradient of the probability distribution heat map, including: enabling a multi-band carrier injection mode for nodes with high probability of out-of-sync, and applying a phase perturbation test to adjacent transmission links.

2. The high-frequency carrier synchronization signal modulation system according to claim 1, characterized in that: The carrier distortion compensation module specifically includes: Historical feature extraction unit: This unit performs multi-scale decomposition on historical transmission data sets, including using empirical mode decomposition to separate the energy proportions of the steady-state phase component and transient noise component of the carrier signal, establishing a correlation matrix between spectrum leakage components and channel signal-to-noise ratio through spectral correlation analysis, and using a dynamic time warping algorithm to align modulated pulse train patterns at different bandwidths. The distortion compensation model construction unit inputs the processed historical transmission data into a hybrid compensation network. The hybrid compensation network includes: a gated cyclic timing prediction unit based on the channel aging curve to generate reference carrier parameters; a fully convolutional network embedded with a distortion attention mechanism to correct phase deviation caused by spectrum leakage; and a pulse train compensator that dynamically adjusts compensation weights based on the modulated pulse train captured in real time. Baseband parameter generation unit: The multi-channel acquisition unit deployed at the RF front end synchronously acquires the phase jitter variance of the carrier signal, the 3rd to 40th harmonic components and phase offset angle of the spectrum leakage, the amplitude jump gradient and time interval entropy of the modulated pulse sequence, and inputs the real-time acquired data into the carrier distortion compensation model to output the reconstructed baseband parameters.

3. The high-frequency carrier synchronization signal modulation system according to claim 2, characterized in that: The process of generating the reconstructed baseband parameters includes executing: Environmental interference filtering: Adaptive filters are designed based on the temperature and humidity parameters of the transmission environment to eliminate phase noise caused by multipath effects. Multi-domain feature fusion operation: Integrate the time-frequency correlation features of carrier phase, spectral components and pulse sequences through cross-domain correlation algorithms; Dynamic parameter output: Generates reconstructed baseband parameters containing normal fluctuation thresholds, and the reconstructed baseband parameters are dynamically updated with the channel aging coefficient.

4. The high-frequency carrier synchronization signal modulation system according to claim 1, characterized in that: The multidimensional distortion analysis module specifically includes: Time domain phase offset calculation unit: performs sliding comparison between the reconstructed baseband parameters and the demodulated baseband parameters in a preset time window, uses a dynamic time warping algorithm to align the asynchronous sampling sequences, calculates the cumulative phase offset in each time window, and generates a time domain offset vector; Frequency domain distortion detection unit: performs frequency band energy decomposition on the spectrum leakage components of the reconstruction parameters and demodulation parameters, extracts the energy density ratio within the preset sensitive frequency band, quantifies the distortion index of each harmonic, and constructs the frequency domain distortion vector; Pulse sequence alignment evaluation unit: Based on the structural similarity algorithm, it matches the spatiotemporal distribution characteristics of the reconstructed and demodulated pulse sequences, calculates the time synchronization error of the pulse amplitude jump point, quantifies the Jensen-Shannon divergence of the pulse interval distribution, and generates a pulse alignment vector; Distortion coefficient tensor generation unit: The time domain offset vector, frequency domain distortion vector and pulse alignment vector are fused into a three-dimensional tensor, and the dimension difference is eliminated through normalization processing weighted by the characteristic entropy value, and the output dimension is a third-order distortion coefficient tensor of [channel number × timestamp × distortion type].

5. The high-frequency carrier synchronization signal modulation system according to claim 4, characterized in that: The frequency domain distortion detection unit performs the following steps: In the frequency domain decomposition stage, the corresponding spectral leakage components in the reconstruction and demodulation baseband parameters are extracted, and multi-resolution frequency band segmentation is performed on each group of harmonic signals using wavelet packet transform to extract the energy distribution characteristics within a preset interference frequency band interval; the interference frequency band interval covers the harmonic resonance range of typical communication equipment; after the extraction is completed, the energy spectral density of the reconstructed and demodulated data within the interference frequency band interval is quantified respectively, and the distortion index of each order harmonic is generated based on the relative distortion degree, and the distortion results of all orders are summarized to construct a frequency domain distortion vector.

6. The high-frequency carrier synchronization signal modulation system according to claim 4, characterized in that: In the pulse sequence alignment evaluation unit, the structural similarity algorithm adopts the Hausdorff distance metric to perform position matching on the jump point sets of the two pulse sequences and identify the time synchronization error.

7. The high-frequency carrier synchronization signal modulation system according to claim 1, characterized in that: The topology positioning network module specifically includes: Transmission topology modeling unit: Constructs a transmission node connection topology diagram based on signal path delay information, annotates the impedance characteristic parameters between nodes, superimposes the phase delay constraints of multipath reflection nodes on the topology diagram, and generates a transmission topology model including an impedance matrix and a node admittance matrix; The out-of-sync path deduction unit maps the distortion coefficient tensor to the corresponding nodes of the transmission topology model; performs out-of-sync path simulation based on a graph convolutional network, which includes: calculating the attenuation factor of the out-of-sync signal based on the node impedance characteristics; capturing cross-node out-of-sync correlation features through a multi-head graph attention mechanism; and using a random walk algorithm to simulate the propagation trajectory of the out-of-sync signal in the topological network; Heat map generation unit: Counts the frequency of out-of-step signals on each path in the simulated propagation, calculates the out-of-step residence probability value based on the impedance characteristic parameters, generates an out-of-step probability distribution heat map covering the entire network, and marks the set of suspicious paths whose probability values ​​exceed the preset out-of-step threshold; Physical node positioning unit: performs density cluster analysis on the probability distribution heat map to identify the out-of-sync probability cluster area; and delineates the out-of-sync physical node boundary based on the transmission node location and topological connection relationship.

8. The high-frequency carrier synchronization signal modulation system according to claim 7, characterized in that: The topology positioning network module further includes outputting suspicious node identification and main out-of-sync path, wherein: Suspicious node identification: Based on the transmission nodes associated with the suspicious path set, the nodes are bound to physical device addresses to generate a suspicious node identification list; Main out-of-sync path: During the random walk simulation, the node path sequence of each round of propagation is recorded, the frequency of occurrence of all paths is counted, the path with the highest frequency is selected as the main out-of-sync path, and an ordered node number sequence is output.

9. The high-frequency carrier synchronization signal modulation system according to claim 1, characterized in that: The adaptive modulation strategy module specifically includes: Carrier injection strategy unit: When the out-of-sync probability of a node in the probability distribution heat map exceeds a preset threshold, it sends a modulation instruction to the RF unit associated with the node to execute the following: increase the carrier sampling rate to 8-12 times the base rate; synchronously enable the harmonic phase tracking mode to capture phase jump events of the 3rd to 40th harmonics; and deploy a transient response recorder on the terminal to store phase jump waveform fragments; Phase perturbation test unit: applies a multi-band phase perturbation sequence to the adjacent node links with the largest out-of-sync probability gradient in the probability distribution heat map.

10. The high-frequency carrier synchronization signal modulation system according to claim 9, characterized in that: The phase disturbance test unit execution process includes: Disturbance signal injection operation: inject 0.5-5kHz swept frequency phase disturbance signal through controllable phase modulator; Theoretical response calculation operation: Calculate the theoretical phase response spectrum based on the transmission topology model and impedance matrix; Measured response acquisition operation: collect the node phase response after the disturbance signal is injected and generate the measured phase response spectrum; Abnormal judgment operation: Calculate the Manhattan distance between the theoretical phase response spectrum and the measured phase response spectrum. When the distance value exceeds the preset offset threshold, it is marked as a phase abnormality-related node.

Images