Satellite common view time synchronization method and system for distributed digital phased array

Abstract

This invention discloses a satellite common-view time synchronization method and system for distributed digital phased arrays, relating to the field of satellite navigation technology. It optimizes common-view satellites and synchronously acquires dual-frequency pseudorange and carrier phase observations, constructing a single-difference carrier phase equation to eliminate satellite clock bias and ephemeris errors. It combines baseline distance dynamic compensation for atmospheric delay and utilizes an integer ambiguity search algorithm to calculate the picosecond-level relative clock bias between master and slave nodes. Local pseudo-random pilot signals are injected into the RF front-end, and channel group delays are measured and subtracted in real time to obtain the end-to-end absolute synchronization deviation matrix. The deviation is decomposed, and the local temperature-controlled crystal is tamed using a digital phase-locked loop to eliminate nanosecond-level macroscopic deviations. Baseband complex multiplication is used for frequency domain phase rotation to offset picosecond-level microscopic deviations. This invention eliminates fiber optic dependence, removes hardware thermal drift black-box errors, achieves end-to-end coherent-level synchronization of distributed nodes, and improves array mobility.

Description

Technical Field

[0001] This invention belongs to the field of satellite navigation technology, and more specifically, relates to a satellite common-view time synchronization method and system for distributed digital phased arrays. Background Technology

[0002] As modern radar and wireless communication technologies evolve towards higher gain and greater mobility, distributed digital phased array technology has emerged. This technology breaks down traditional centralized large antenna arrays into multiple spatially separated independent radio frequency nodes, achieving high-gain beamforming through coherent power combining of spatial electromagnetic waves. Its core premise is that each independent node must maintain extremely high time and phase synchronization accuracy, typically requiring time errors to be strictly controlled at the picosecond level and phase errors to be less than a dozen degrees.

[0003] Existing high-precision time synchronization technologies primarily rely on fiber optic networks to transmit reference clock signals. For example, the White Rabbit protocol, widely used in high-energy physics, can achieve synchronization accuracy down to the sub-nanosecond or even picosecond level. In practical distributed radar deployments, nodes are typically connected in a star topology by extending multiple fiber optic cables several kilometers long from a central control vehicle. However, while traditional global navigation satellite system (GNSS) timing technologies offer the advantage of being cable-free, they rely heavily on pseudorange measurements. Limited by symbol width and residual atmospheric delay errors, their synchronization accuracy is typically only in the nanosecond to tens of nanosecond range. Even traditional satellite common-view technology used for time comparison between remote timekeeping laboratories suffers from poor real-time performance and does not account for delay drift within the receiver's internal radio frequency channels, making it difficult to achieve accuracy exceeding hundreds of picoseconds.

[0004] This reveals the following problems with existing technologies: fiber optic synchronization schemes limit the field mobility, damage resistance, and scalability of distributed arrays, and are prone to system failure due to cable breakage; existing cableless satellite timing and common-view technologies have orders of magnitude gaps in synchronization accuracy and have not eliminated the black box delay caused by temperature aging of the internal radio frequency channels, thus failing to meet the picosecond-level end-to-end synchronization requirements necessary for spatial coherent synthesis of microwave phased arrays. Summary of the Invention

[0005] (a) Technical problems to be solved To address the problems in related technologies, this invention provides a satellite common-view time synchronization method and system for distributed digital phased arrays, thereby overcoming the aforementioned technical problems existing in the prior art.

[0006] (II) Technical Solution To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: On the one hand, the present invention provides a satellite common-view time synchronization method for distributed digital phased arrays, comprising the following steps: S1. Obtain real-time elevation angle data, real-time signal-to-noise ratio data, and position accuracy factor of visible navigation satellites. Calculate scores and filter them using the satellite quality assessment function to obtain a set of common-view satellites. S2. Lock the dual-frequency navigation signal of the common-view satellite set, extract the dual-frequency pseudorange observation sequence and carrier phase observation sequence, and use a global counter to add a local timestamp to construct the master node observation dataset and the slave node observation dataset. S3. Perform single-difference calculation on the master node observation dataset and the slave node observation dataset, and call the corresponding atmospheric delay compensation model to process the single-difference data according to the three-dimensional spatial straight-line distance between the master node and the slave node, and calculate the relative clock difference coarse value and the single-difference carrier phase equation. S4. Based on the coarse value of relative clock difference, the integer ambiguity of the carrier phase is searched through the integer ambiguity algorithm model, and the precise integer value of the integer ambiguity is output. Substitute the exact integer value of the integer ambiguity back into the single-difference carrier phase equation to calculate the high-precision relative clock difference between the master and slave nodes. S5. The master node and slave node generate local calibration pilot signals and inject them into the input ports of their respective RF front-ends. The RF channel group delay of the master node and the RF channel group delay of the slave node are measured respectively. By combining the high-precision relative clock difference between master and slave nodes, the delay of the master node RF channel group, and the delay of the slave node RF channel group, a reference plane conversion is performed to calculate the end-to-end absolute synchronization deviation matrix. S6. Divide the absolute synchronization deviation matrix into a nanosecond-level integer deviation part and a picosecond-level fractional deviation part; The nanosecond-level integer deviation portion is input into a digital phase-locked loop to generate a voltage-controlled oscillator control voltage, so as to tame the node's local temperature-controlled crystal oscillator in real time. The picosecond-level fractional deviation is converted into a frequency domain phase rotation factor, and the frequency domain phase rotation factor is used in the digital baseband to perform complex multiplication phase compensation on the radar waveform data of the slave node. This invention eliminates satellite clock and ephemeris errors by constructing a single-difference carrier phase equation between master and slave nodes, and achieves picosecond-level spatial link relative time comparison by combining an integer ambiguity fixing algorithm. By introducing a self-calibration mechanism for local RF channel group delay, the reference plane for time synchronization is accurately moved down from the physical antenna aperture to the digital baseband, eliminating black box errors caused by hardware thermal drift. Through a dual closed-loop control mechanism combining macroscopic discipline of isothermal crystal oscillator and microscopic phase rotation in the baseband frequency domain, coherent time and phase synchronization across nodes and end-to-end is achieved, eliminating the system's dependence on fiber optic cables.

[0007] Preferably, the satellite quality assessment function in S1 calculates scores and filters them into a common-view satellite set, specifically including the following steps: S11, Read the... i Based on the real-time elevation angle data, real-time signal-to-noise ratio data, and position accuracy factor of the visible navigation satellites, an initial candidate satellite list is established; S12. Set the elevation angle threshold; determine if the real-time elevation angle data is less than the elevation angle threshold; if so, directly set the threshold value. i Each visible navigation satellite is assigned a score of 0 and removed from the initial candidate satellite list, resulting in an updated candidate satellite list. Otherwise, proceed to step S13 to calculate the score and retain it in the updated candidate satellite list; S13. Call the satellite quality assessment function to calculate the score; the satellite quality assessment function is as follows: divide the real-time elevation angle data and the real-time signal-to-noise ratio data by their respective physical extreme values ​​for positive normalization and multiply by the first positive weighting coefficient; divide the position accuracy factor by the physical extreme value for negative normalization and multiply by the second negative penalty coefficient; add the positive normalization result and the negative normalization result to obtain the score. S14. Sort all visible navigation satellites with scores greater than 0 in the updated candidate satellite list in descending order of scores, extract the numbers of the top N visible navigation satellites, and generate a common-view satellite set. This invention introduces a multi-dimensional satellite quality assessment function, comprehensively considering the penetration path length, received signal-to-noise ratio, and spatial geometric distribution of satellite signals. Through a quantitative scoring and ranking mechanism, it dynamically eliminates low-quality satellite signals severely affected by multipath effects, ensuring that the navigation signals ultimately selected for the common-view satellite set have optimal physical purity and geometric solution accuracy. This improves the anti-interference capability and convergence speed of subsequent carrier phase single-difference calculations from the source.

[0008] Preferably, step S3 involves calling the corresponding atmospheric delay compensation model to process the single-difference data based on the three-dimensional straight-line distance between the master node and the slave node, specifically including the following steps: S31. Read the three-dimensional geodetic coordinates of the master node and the slave node, and use the Euclidean distance algorithm to calculate the three-dimensional straight-line distance between the master node and the slave node; set the baseline cancellation threshold. S32. Determine whether the straight-line distance in three-dimensional space is less than or equal to the baseline cancellation threshold; if so, set the ionospheric delay single-difference term and the tropospheric delay single-difference term in the initial single-difference carrier phase equation to 0, and obtain the compensated single-difference carrier phase equation. Otherwise, extract real-time meteorological data from the locations of the master and slave nodes, input the real-time meteorological data into the Hopfield tropospheric model to calculate the absolute delay and subtract them to obtain the tropospheric delay single-difference compensation value; substitute the tropospheric delay single-difference compensation value into the initial single-difference carrier phase equation and perform a subtraction operation to obtain the compensated single-difference carrier phase equation. This invention dynamically switches atmospheric delay compensation strategies based on the actual physical distance between master and slave nodes; in short baseline scenarios, direct cancellation is used to avoid introducing secondary noise into the model, while in long baseline scenarios, real-time meteorological data is introduced for gradient compensation, thus balancing the computational efficiency of the algorithm with the robustness of the space environment.

[0009] Preferably, the calculation of the relative clock bias coarse value and the single-difference carrier phase equation in step S3 specifically includes the following steps: S33. Construct an initial single-difference carrier phase equation that includes a single difference term for carrier phase observations, a single difference term for geometric distance, a single difference term for relative clock error, a single difference term for ionospheric delay, a single difference term for tropospheric delay, a single difference term for integer ambiguity, and a noise term. S34. Extract the dual-frequency pseudorange observation sequence obtained in step S2, replace the carrier phase observation sequence in the initial single-difference carrier phase equation with the dual-frequency pseudorange observation sequence, and remove the integer ambiguity single-difference term in the equation to generate an ambiguity-free pseudorange observation equation. S35. Substitute the geometric coordinates of the master node and slave node into the unambiguous pseudorange observation equation, and solve iteratively using the least squares method to output the coarse relative clock difference value at the nanosecond level. This invention constructs an unambiguous pseudorange observation equation and uses the least squares method for iterative solution, which can obtain a coarse relative clock difference value at the nanosecond level in a very short time. Although the accuracy of this coarse value is limited, it provides a highly reliable and well-defined preliminary convergence center for subsequent complex carrier phase integer ambiguity search, which greatly reduces the search blind zone in the multidimensional space and improves the real-time performance of the algorithm.

[0010] Preferably, step S4 involves searching for the integer ambiguity of the carrier phase using an integer ambiguity algorithm model and outputting precise integer values ​​for the integer ambiguity. This specifically includes the following steps: S41. Extract the covariance matrix of the compensated single-difference carrier phase equation output in step S3; construct the Z-transform integer matrix, and multiply the covariance matrix of the compensated single-difference carrier phase equation on the left and on the right by the Z-transform integer matrix to generate the orthogonal search space matrix after downcorrelation processing. S42. Input the coarse relative clock difference value output in step S3 into the orthogonal search space matrix for coordinate transformation, and calculate the three-dimensional coordinates of the center of the multidimensional search sphere. S43. Using the central three-dimensional coordinates as the origin, set a search radius threshold; within the sphere volume limited by the search radius threshold, use a sequential conditional search algorithm to find the first set of integer vectors that minimizes the sum of squared residuals and the second set of integer vectors that have the second smallest sum of squared residuals. S44. Calculate the ratio of the second smallest residual sum of squares to the smallest residual sum of squares; determine whether the ratio is greater than the preset confidence threshold; if so, determine that the first set of integer vectors is an integer value of integer ambiguity; otherwise, determine that the search has failed and discard the observation data of the current epoch. This invention utilizes the Z-transform integer matrix to reduce the correlation of highly correlated covariance matrices, transforming the narrow search ellipsoid into an orthogonal sphere. It also combines a coarse relative clock difference value to limit the search radius and employs a sequential conditional search algorithm to quickly lock the optimal solution. This breakthrough overcomes the bottleneck of integer ambiguity technology in carrier phase observation, significantly improving the relative time comparison accuracy from the nanosecond level of pseudorange to the picosecond level, laying a data foundation for the realization of coherent synthesis in the microwave band.

[0011] Preferably, the step S5, measuring the radio frequency channel group delay value transmitted from the local calibration pilot signal to the digital baseband output port, specifically includes the following steps: S51. Generate a Gold polynomial pseudo-random sequence inside the field-programmable gate array of the slave node as a local calibration pilot signal; S52. Convert the local calibration pilot signal into an analog radio frequency signal, and use a directional coupler to inject the analog radio frequency signal in reverse into the radio frequency front-end input port of the slave node; S53. The injected analog radio frequency signal passes through a low-noise amplifier and an intermediate frequency filter in sequence, and is then resampled by an analog-to-digital converter and input to the digital baseband output port to obtain a resampled signal. S54. Perform a sliding correlation operation between the resampled signal and the originally generated Gold polynomial pseudo-random sequence; find the number of sliding clock cycles when the correlation peak occurs; multiply the number of sliding clock cycles by the system clock cycle duration to calculate the RF channel group delay value; This invention utilizes a pseudo-random sequence generated by a digital baseband as a local calibration pilot signal and injects it inversely into the RF front end, allowing it to undergo the same physical transmission path as real satellite signals. By accurately extracting the channel group delay through sliding correlation operations, this mechanism transforms the closed analog RF link into an observable system, effectively quantifying the internal transmission delay caused by device aging and thermal expansion and contraction, and providing key state parameters for achieving high-precision end-to-end synchronization.

[0012] Preferably, calculating the end-to-end absolute synchronization deviation matrix in step S5 specifically includes the following steps: S55. Read the high-precision relative clock difference between the master and slave nodes calculated in step S4 as the reference time parameter, and read the RF channel group delay value calculated in step S54 as the internal subtraction parameter. S56. Combine the high-precision relative clock difference between the master and slave nodes with the delay of the master node's RF channel group and the delay of the slave node's RF channel group to perform a reference plane conversion. Then, shift the physical reference plane for time synchronization from the antenna phase center of the master and slave nodes to their respective digital baseband processing chips to generate an initial absolute synchronization deviation sequence. S57. Construct a sliding mean filter, input the initial absolute synchronization deviation sequence into the sliding mean filter to remove transient spikes and jumps, and output a smoothed end-to-end absolute synchronization deviation matrix. This invention combines the high-precision relative clock difference of the space link with the delay of the radio frequency channel group inside the hardware to achieve precise translation of the physical reference plane for time synchronization. At the same time, a sliding mean filter is introduced to effectively suppress transient spike jumps and output a smooth and stable absolute synchronization deviation matrix, avoiding control lockout caused by sudden strong electromagnetic interference from the outside, and ensuring the continuity and reliability of subsequent hardware discipline commands.

[0013] Preferably, step S6 involves inputting the nanosecond-level integer deviation portion into a digital phase-locked loop to generate a voltage-controlled oscillator control voltage, so as to tame the node's local temperature-controlled crystal oscillator in real time. This specifically includes the following steps: S61. Read the nanosecond-level integer deviation portion of the absolute synchronization deviation matrix; S62. Construct a proportional-integral controller model that includes proportional gain coefficient and integral gain coefficient; S63. Input the nanosecond-level integer deviation portion into the proportional-integral controller model, and calculate the control voltage increment command by summing the proportional term and the integral term; S64. The control voltage increment command is superimposed on the base voltage of the previous cycle, and the analog control voltage is output to the voltage control pin of the local temperature-controlled crystal oscillator through the digital-to-analog converter chip to change the crystal resonant frequency to eliminate the nanosecond-level integer deviation. This invention utilizes a proportional-integral controller model to perform closed-loop adjustment of nanosecond-level integer deviations in the absolute synchronization deviation matrix, generating precise control voltage increment commands to tame the local temperature-controlled crystal oscillator. This mechanism fully leverages the voltage-controlled characteristics of the crystal oscillator, eliminating macroscopic time drift between master and slave nodes from the source through physical means, ensuring that the clock frequency of the slave node always closely follows the master node, thus guaranteeing the global time uniformity and stability of the distributed system under long-term operation.

[0014] Preferably, step S6 involves converting the picosecond-level fractional deviation into a frequency-domain phase rotation factor, and then using this frequency-domain phase rotation factor to perform complex multiplication phase compensation on the radar waveform data of the slave node in the digital baseband. Specifically, this includes the following steps: S65. Read the picosecond-level fractional deviation portion of the absolute synchronization deviation matrix; S66. Extract the specific frequency value of the current working radio frequency carrier of the distributed digital phased array; S67. Multiply the specific frequency value by the picosecond-level decimal deviation part, and multiply by the pi constant factor to calculate the radian value of the frequency domain phase rotation factor. S68. Input the radian value of the frequency domain phase rotation factor into the coordinate rotation digital computer algorithm IP core to generate a complex rotation matrix containing in-phase cosine components and orthogonal sine components; perform point-by-point complex multiplication operation between the radar waveform complex data stream being processed and the complex rotation matrix to cancel out the picosecond-level decimal deviation in physical time. This invention addresses the picosecond-level residual error that cannot be eliminated by the voltage-controlled resolution of temperature-controlled crystal oscillators. It innovatively transforms this error into a frequency-domain phase rotation factor and generates a complex rotation matrix in the digital baseband for point-by-point compensation. Without changing the underlying hardware physical state, it utilizes the extremely high flexibility and precision of digital signal processing to smooth out microscopic time deviations, achieve coherent phase alignment, and maximize the spatial power synthesis efficiency of the phased array.

[0015] On the other hand, the present invention also provides a satellite common-view time synchronization system for distributed digital phased arrays, used to implement a satellite common-view time synchronization method for distributed digital phased arrays, the system comprising: The device consists of a master node and at least one slave node; both the master node and the slave node are equipped with a dual-band GNSS antenna, an RF front-end receiving module, and a baseband signal processing module. The baseband signal processing module includes a field-programmable gate array (FPGA) chip; the FPGA chip is internally configured with a satellite signal acquisition and tracking logic unit, a timestamp latch generation logic unit, a digital phase-locked loop control logic unit, and a complex multiplication phase compensation logic unit; The master node device and the slave node device exchange observation data required for single difference calculation via a wireless microwave data link.

[0016] (III) Beneficial Effects The present invention has the following beneficial effects: This invention breaks through the accuracy bottleneck of traditional satellite time synchronization and achieves picosecond-level ultra-high precision comparison. By introducing carrier phase single-difference solution combined with short baseline atmospheric error dynamic cancellation technology, and using the decorrelation matrix and conditional search algorithm to quickly fix integer ambiguity, this invention improves the satellite common-view time comparison accuracy from the nanosecond level of pseudorange to the picosecond level, thus meeting the physical requirements of microwave phased array for spatial coherent synthesis.

[0017] This invention achieves end-to-end synchronization through a joint calibration mechanism for radio frequency channels. It uses a pseudo-random pilot signal generated by the digital baseband to be injected into the radio frequency front end in reverse, and measures and subtracts the group delay caused by temperature drift and device aging in the analog link in real time. It accurately moves the physical reference plane for time synchronization from the antenna port to the inside of the digital baseband, eliminating the black box error inside traditional equipment.

[0018] This invention constructs a macro- and micro-level dual closed-loop compensation architecture, endowing the distributed system with excellent mobility and robustness. By splitting the synchronization deviation into nanosecond-level integers and picosecond-level fractions, it performs joint compensation through macro-level voltage control discipline of a temperature-controlled crystal oscillator and micro-level frequency domain phase rotation of a digital baseband, respectively. It eliminates the dependence on fiber optic cables and significantly improves the mobile deployment capability, damage resistance and reconfiguration capability, and flexible scalability of the distributed radar array in complex field environments.

[0019] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the satellite common-view time synchronization method for distributed digital phased arrays according to the present invention. Figure 2 This is a graph showing the change in the ratio of integer ambiguity search residual convergence to confidence level in this invention. Figure 3 This is a convergence curve of the output voltage of the crystal oscillator disciplined PI controller of the present invention; Figure 4 This is a diagram of the satellite common-view time synchronization system architecture of the distributed digital phased array of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the invention, and not all embodiments. Based on the embodiments of the invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the invention.

[0023] Please see Figure 1 This invention discloses a satellite common-view time synchronization method for distributed digital phased arrays, comprising the following steps: S1. Obtain real-time elevation angle data, real-time signal-to-noise ratio data, and position accuracy factor of the visible navigation satellites. Calculate the score using a preset satellite quality evaluation function and select the top N visible navigation satellites as a common-view satellite set. In step S1, a score is calculated using a preset satellite quality evaluation function, and the top N visible navigation satellites are selected as a common-view satellite set. This specifically includes the following steps: S11. Read the real-time elevation angle data, real-time signal-to-noise ratio data, and position accuracy factor of the i-th visible navigation satellite, and establish an initial candidate satellite list; Specifically, in embodiment S11, the application scenario of the present invention is a distributed digital phased array radar system deployed in complex terrain in the field; the system consists of a core radar vehicle as the master node and multiple unmanned radar sub-array vehicles distributed within a radius of several kilometers as slave nodes; since it is necessary to perform 10GHz X-band spatial electromagnetic wave coherent power synthesis, the signal transmission time error between each node must be controlled within 5 picoseconds, corresponding to a phase error of less than 18 degrees; After the system powers on, the GNSS receiver baseband signal processing modules of the master and slave nodes begin parallel satellite search; the system calculates the geometric parameters of all visible navigation satellites in the current sky by parsing the ephemeris data in the navigation message and combining it with local coarse coordinates; the system reads the first... i The system stores the real-time elevation angle data of the visible navigation satellites, such as the current elevation angle of the BeiDou PRN07 satellite being 22 degrees; real-time signal-to-noise ratio data, such as a carrier-to-noise ratio of 45 dB-Hz; and position accuracy factors, such as a spatial geometric distribution factor of 2.1. The system then stores the satellite numbers containing these three parameters into the system memory to establish an initial candidate satellite list. S12. Set the elevation angle threshold to 15 degrees; determine if the real-time elevation angle data is less than 15 degrees; if the real-time elevation angle data is less than 15 degrees, then directly set the threshold to 15 degrees. i The score of each visible navigation satellite is assigned to 0 and removed from the initial candidate satellite list to obtain an updated candidate satellite list; if the real-time elevation angle data is greater than or equal to 15 degrees, then step S13 is executed to calculate the score and retain it in the updated candidate satellite list. Specifically, in embodiment S12, in order to ensure the accuracy of the common-view solution, low-elevation satellites that are severely affected by ground multipath reflection must be eliminated; because the signals of low-elevation satellites penetrate a very long layer of atmosphere and are easily reflected and refracted by ground buildings or trees, introducing serious non-line-of-sight errors. The system determines whether the real-time elevation angle data is less than 15 degrees. If the elevation angle of a satellite is only 8 degrees, the determination result is less than 15 degrees. The system directly assigns the score of the satellite to 0, performs a memory release operation, forcibly removes it from the initial candidate satellite list, and generates an updated candidate satellite list after removing inferior satellites. If the determination result is greater than or equal to 15 degrees, it is retained in the list and prepared to enter the next scoring stage. S13. Call the satellite quality assessment function to calculate the score; the satellite quality assessment function is as follows: divide the real-time elevation angle data and the real-time signal-to-noise ratio data by their respective physical extreme values ​​for positive normalization and multiply by the first positive weighting coefficient; divide the position accuracy factor by the physical extreme value for negative normalization and multiply by the second negative penalty coefficient; add the positive normalization result and the negative normalization result to obtain the score. Specifically, in embodiment S13, the system calculates a score for each satellite in the updated candidate satellite list by calling the satellite quality assessment function. The specific mathematical formula is as follows: ; in, High school i For real-time elevation angle data, SNR i This is real-time signal-to-noise ratio data. PDOP i This is the position accuracy factor; in the formula, W The positive weighting coefficient for the elevation angle is set to 0.4. The larger the elevation angle, the smaller the tropospheric delay and the more accurate the model. W 2 is the positive weighting coefficient for the signal-to-noise ratio, which is set to 0.4. The signal-to-noise ratio directly determines the tracking thermal noise of the carrier phase-locked loop. W 3 is the negative penalty coefficient for geometric accuracy, set to 0.2. The larger the PDOP value, the denser the satellite distribution and the greater the geometric solution error, so it is used as a deduction penalty; the denominators 90, 50, and 10 are the physical extreme value normalization constants of each parameter, respectively. The W 1. W 2 and W3. A three-factor, three-level orthogonal experimental design was used for calibration. The success rate of integer ambiguity fixation and the convergence time of clock error resolution were used as comprehensive evaluation indicators. Elevation angle weight, signal-to-noise ratio weight, and PDOP penalty weight were selected as experimental factors, with three gradient levels set for each factor. The experiment covered six typical field deployment scenarios, including low-elevation-angle multipath propagation and strong electromagnetic interference, and a total of 200 observation epochs were collected for statistical analysis. Verification was performed using range analysis. W 1 = 0.4 W 2 = 0.4 W With a weight combination of 3=0.2, the success rate of fixed ambiguity is improved by 12.7%, the convergence time is shortened by 18.3%, and the overall performance is optimal. Furthermore, it should be noted that PDOP is an indicator where a smaller value represents a better satellite spatial geometry distribution. In order to linearly add it to indicators such as elevation angle and signal-to-noise ratio, which represent better quality, within the same scoring system, this indicator must be reversed. The subtraction operation in the formula is the mathematical implementation of this reverse process. That is, when the PDOP value is larger, the subtraction term has a larger score, and the total score is reduced accordingly, which constitutes a negative penalty term for poor spatial geometry distribution. Dividing by the physical extreme value of 10 and multiplying by the negative penalty coefficient constitutes the complete reverse normalization process. S14. Sort all visible navigation satellites with scores greater than 0 in the updated candidate satellite list in descending order of scores, extract the numbers of the top N visible navigation satellites, and generate a common-view satellite set. Specifically, in embodiment S14, the system uses a quick sorting algorithm to sort all visible navigation satellites with scores greater than 0 in the updated candidate satellite list in descending order of scores from high to low. In this embodiment, N=6 is set, and the system extracts the numbers of the top 6 satellites, such as PRN03, PRN07, PRN12, PRN15, PRN22, and PRN28, to generate a common-view satellite set. The common-view satellite set output in this step is directly transmitted to the subsequent radio frequency receiving link to ensure that the satellite signals participating in the high-precision calculation are the carriers with the optimal current spatial geometric distribution and the purest signal quality. This embodiment details the optimization process of the common-view satellite ensemble; by integrating elevation angle, signal-to-noise ratio and geometric accuracy factor to construct a quantitative evaluation system, the system can adaptively select the best satellite signals in complex field environments; it not only eliminates interference from non-line-of-sight errors, but also ensures the purity of observation data from the physical source, providing reliable data support for high-precision common-view calculation of distributed phased arrays in harsh electromagnetic environments; S2. Lock the dual-frequency navigation signal of the common-view satellite set, and extract the dual-frequency pseudorange observation sequence and carrier phase observation sequence; A global counter is used to add a local timestamp to the dual-frequency pseudorange observation sequence and the carrier phase observation sequence to construct the master node observation dataset and the slave node observation dataset. Specifically, in embodiment S2, both the master node and the slave node are equipped with choke coil anti-multipath antennas supporting the L1 1575.42MHz and L2 1227.60MHz frequency bands; the RF front-end receiving module downconverts the high-frequency microwave signal to the intermediate frequency and performs high-frequency analog-to-digital conversion; inside the baseband processing chip, the system allocates independent hardware tracking channels to these 6 satellites according to the common-view satellite set generated in step S14; Each hardware channel contains a delay phase-locked loop (DLL) and a phase phase-locked loop (PLL). The DLL continuously adjusts the phase of the locally copied pseudo-random code to align it with the received satellite ranging code, thereby measuring the signal propagation time. Multiplying this by the speed of light yields a dual-frequency pseudorange observation sequence. The pseudorange observation is characterized by its lack of ambiguity, but due to the large symbol width, its measurement noise is typically in the decimeter to meter range, corresponding to a time error in the nanosecond range. Phase-locked loops (PLLs) track the carrier phase of satellite signals and output a sequence of carrier phase observations with a measurement resolution down to the millimeter level. The wavelength of the carrier phase is extremely short, such as about 19 centimeters in the L1 band, and the measurement resolution can reach one-hundredth of the wavelength, i.e., the millimeter level, corresponding to a time error of the picosecond level. However, the carrier phase is a periodic signal that is constantly rotating. The receiver can only measure a fractional part of the phase that is less than a complete wavelength and the integer number of cycles accumulated from the start of locking. In the space between the signal being transmitted from the satellite and the receiver locking, there are an unknown number of complete integer wavelengths. This is the integer ambiguity problem that must be solved in subsequent steps. To enable comparison of these observation data between the master and slave nodes, the system utilizes a global counter running at 100MHz within a field-programmable gate array (FPGA). At the rising edge of the local second pulse signal, it synchronously latches the dual-frequency pseudorange and carrier phase observation sequences of all current channels and adds a local timestamp. The master and slave nodes package these time-stamped observations to construct master-node and slave-node observation datasets respectively. The slave-node observation datasets are then sent to the master node via a wireless microwave data link for differential processing. The observations include pseudorange, phase, and Doppler shift. Furthermore, the dual-frequency pseudorange observation sequence and carrier phase observation sequence in this invention can utilize, for example, L1 1575.42MHz and L2 The 1227.60MHz frequency point is combined to eliminate ionospheric delay. Specifically, using an ionospheric-free combination method, the pseudorange or carrier phase observations of frequencies L1 and L2 are linearly combined. Through the phase or pseudorange difference between the two frequencies, the first-order term of ionospheric delay can be calculated and eliminated. This first-order term accounts for more than 99% of the total ionospheric delay, which allows for significant suppression of residual ionospheric errors at the single-node solution level, even in scenarios where complete atmospheric cancellation of short baselines is not applicable, thus improving the convergence speed of the subsequent single-difference model. The wide-lane combination of L1 and L2 is used to construct a wide-lane carrier phase observation with a wavelength of approximately 86.2 cm. This ultra-long wavelength characteristic can greatly alleviate the convergence pressure of integer ambiguity search and can be used as a preprocessing step for accurate integer ambiguity search in step S4. First, the integer ambiguity is fixed using the wide-lane combination, and then the result is mapped back to the basic ambiguity of L1 and L2, improving the search success rate. This embodiment uses the global counter inside the FPGA to achieve synchronous latching of multi-channel pseudorange and carrier phase observations under the unified triggering of the local second pulse; it ensures that the observation data acquired by the master and slave nodes in different locations have strict time consistency, eliminates the systematic deviation introduced by the misalignment of data sampling time, and builds an accurate timestamp mapping benchmark for subsequent single difference calculation; S3. Perform single-difference calculation on the master node observation dataset and the slave node observation dataset to eliminate satellite clock error and satellite ephemeris error. Then, call the corresponding atmospheric delay compensation model to process the single-difference data according to the three-dimensional spatial straight-line distance between the master node and the slave node, and calculate the relative clock error coarse value and the single-difference carrier phase equation. Specifically, in embodiment S3, the error sources of the satellite navigation system are extremely complex; the satellite's own atomic clock has satellite clock errors, and the orbital positions broadcast by the satellite have ephemeris errors; due to the master node m With slave node u Simultaneous observation of the same common-view satellite i The signals emitted by the satellite propagate to these two nodes along almost parallel paths; In step S3, the corresponding atmospheric delay compensation model is invoked to process the single difference data based on the three-dimensional straight-line distance between the master node and the slave node. This specifically includes the following steps: S31. Read the three-dimensional geodetic coordinates of the master node and the slave node, and use the Euclidean distance algorithm to calculate the three-dimensional straight-line distance between the master node and the slave node; set the baseline cancellation threshold to 5000 meters. Specifically, in embodiment S31, the baseline cancellation threshold is established by extracting past meteorological sounding data to create a fitting curve of the spatial correlation coefficient of the total tropospheric zenith delay decreasing with distance; when the spatial correlation coefficient drops to 0.95, the corresponding physical distance is the atmospheric environment consistency boundary. After receiving the observation dataset from the slave node, the master node subtracts the carrier phase equation from the carrier phase equation in the slave node's observation dataset to obtain the initial single-difference carrier phase equation. Through subtraction, the huge satellite clock error and satellite ephemeris error in the equation are mathematically canceled out. Subsequently, the system reads the three-dimensional geodetic coordinates of the master node and the slave node, and uses the Euclidean distance algorithm to calculate the three-dimensional straight-line distance between the master node and the slave node. S32. Determine whether the straight-line distance in three-dimensional space is less than or equal to 5000 meters; if the straight-line distance in three-dimensional space is less than or equal to 5000 meters, then directly assign the ionospheric delay single-difference term and the tropospheric delay single-difference term in the initial single-difference carrier phase equation to 0, and obtain the compensated single-difference carrier phase equation. Specifically, in embodiment S32, if the straight-line distance in three-dimensional space, such as the distance between the master and slave radar vehicles, is less than 5000 meters (1200 meters), the system determines that the total ionospheric electron content and tropospheric water vapor distribution above the master and slave nodes are completely consistent. At this time, the delay experienced by the master node is equal to the delay experienced by the slave node, and the difference is 0. The system directly forces the ionospheric delay single-difference term and the tropospheric delay single-difference term in the initial single-difference carrier phase equation to be 0, performs complete cancellation, and outputs the compensated single-difference carrier phase equation, which simplifies the equation and avoids the secondary noise introduced by the model. Furthermore, regarding the ionospheric delay single difference term, in this embodiment of the invention, for scenarios where the baseline cancellation threshold of 5000 meters is less than or equal to the baseline cancellation threshold, based on the spatial correlation assumption, the ionospheric delay single difference term and the tropospheric delay single difference term are forcibly assigned to 0 for direct cancellation; for scenarios where the straight-line distance in three-dimensional space is greater than 5000 meters, as an optional processing strategy, if only single-frequency observation data is used, the parameters of the international reference ionospheric model and the Klobuchar model can be further combined to calculate the zenith ionospheric delay and obtain the absolute amount of the slant delay through the mapping function for bi-station differential compensation; If the straight-line distance in three-dimensional space is greater than 5000 meters, then extract the real-time meteorological data of the location of the master node and the slave node, input the real-time meteorological data into the Hopfield tropospheric model to calculate the absolute delay respectively and subtract them to obtain the tropospheric delay single difference compensation value. Substitute the tropospheric delay single-difference compensation value into the initial single-difference carrier phase equation and perform a subtraction operation to obtain the compensated single-difference carrier phase equation. If the straight-line distance in three-dimensional space, such as in a super-large cross-regional distributed array, is greater than 5000 meters (15000 meters apart), there will be spatial gradient differences in the atmospheric environment. The system extracts real-time meteorological data transmitted from meteorological sensors at the locations of the master node and slave node, such as temperature 25℃, air pressure 1013hPa, and relative humidity 60%, and inputs them into the Hopfield tropospheric model to calculate the absolute delay of the master node and slave node respectively. The difference is then used to obtain the tropospheric delay single-difference compensation value. The system substitutes the tropospheric delay single-difference compensation value into the initial single-difference carrier phase equation for subtraction and outputs the gradient-corrected compensated single-difference carrier phase equation. Furthermore, the sources of the real-time meteorological data include, but are not limited to, the node's built-in meteorological sensors, wireless data link synchronization, and access to external meteorological grid data; The nodes are equipped with weather sensors, and the master and slave node devices can integrate a micro weather station module. They can directly transmit locally measured temperature, air pressure, and humidity data back to the baseband processing module via a serial interface. The wireless data link synchronization allows the master and slave nodes to receive current meteorological environmental parameters broadcast by the master node when some simplified configuration slave nodes are not equipped with weather sensors. The external meteorological grid data access allows the system to call real-time meteorological grid forecast products issued by the weather forecast center or local meteorological department based on the three-dimensional geodetic coordinates of the master and slave nodes to interpolate the meteorological data of the target point, provided that there is network access or a pre-stored digital map. This embodiment dynamically and adaptively switches atmospheric delay compensation strategies based on atmospheric spatial correlation boundaries; it directly cancels out the delay within a 5000-meter short baseline to improve computational efficiency, and introduces meteorological data for precise compensation in ultra-long baseline scenarios; this design enables the system to flexibly adapt to different scales of distributed phased array deployments, effectively eliminating the impact of spatial environmental gradient differences on time synchronization accuracy. Furthermore, the calculation of the relative clock bias coarse value and the single-difference carrier phase equation in S3 specifically includes the following steps: S33. Construct an initial single-difference carrier phase equation that includes a single difference term for carrier phase observations, a single difference term for geometric distance, a single difference term for relative clock error, a single difference term for ionospheric delay, a single difference term for tropospheric delay, a single difference term for integer ambiguity, and a noise term. Specifically, in embodiment S33, the master node m Common-view satellite i The carrier phase equation is ;in, f i m Represents the carrier phase observation value. r i m Represents the true geometric distance. cδtm Indicates receiver clock bias. cδt i Indicates satellite clock bias, I i m and T i m Indicates the delay between the ionosphere and the troposphere; N i m Indicates the ambiguity of the integer period. e i m Indicates noise; Subtracting the slave node observation equations from the master node observation equations yields the master node observation equation. m With slave node u For the same shared satellite i The specific mathematical formula of the single-difference carrier phase equation ; where Δ f i mu This represents the single difference term of the carrier phase observation; Δ r i mu Represents the single difference term of geometric distance; c Δ represents the speed constant of light in a vacuum; δt mu The relative clock error term to be solved; Δ I i mu Indicates the single difference term of ionospheric delay; Δ T i mu The term represents the tropospheric delay single difference; λ represents the carrier wavelength; Δ N i mu Δ represents the integer ambiguity single difference term to be solved; e i mu This represents the multipath effect and the single difference term of thermal noise; S34. Extract the dual-frequency pseudorange observation sequence obtained in step S2, replace the carrier phase observation sequence in the initial single-difference carrier phase equation with the dual-frequency pseudorange observation sequence, and remove the integer ambiguity single-difference term in the equation to generate an ambiguity-free pseudorange observation equation. Specifically, in embodiment S34, since pseudorange observations inherently do not have ambiguity issues, the system uses the dual-frequency pseudorange observation sequence obtained in step S2 to replace the carrier phase observations on the left side of the above equation, and directly removes them on the right side of the equation. This step generates an unambiguous pseudorange observation equation; S35. Substitute the geometric coordinates of the master node and slave node into the unambiguous pseudorange observation equation, and solve iteratively using the least squares method to output the coarse relative clock difference value at the nanosecond level. Specifically, in embodiment S35, the system uses the least squares method to iteratively solve the ambiguity-free pseudorange observation equation to obtain the variable to be solved in the equation, namely the coarse value of the relative clock difference. Since the pseudorange noise is large, the accuracy of the coarse value of the relative clock difference calculated at this time is usually between 1 nanosecond and 3 nanoseconds. The system saves it to provide a convergence boundary for the next step of the solution. This embodiment constructs a rigorous single-difference carrier phase equation and an unambiguous pseudorange observation equation, and cleverly utilizes the physical characteristic that pseudorange data has no multiple solutions. It then uses the least squares method to quickly calculate the coarse relative clock difference value at the nanosecond level. Although this process cannot directly meet the synchronization requirements of the phased array, it defines an accurate initial convergence interval for the subsequent complex carrier phase ambiguity search, which greatly reduces the computational load of the system. S4. Input the coarse value of the relative clock difference into the integer ambiguity fixing algorithm model, search for the integer ambiguity of the carrier phase in the single-difference carrier phase equation, and output the precise integer value of the integer ambiguity. Substitute the exact integer value of the integer ambiguity back into the single-difference carrier phase equation to calculate the high-precision relative clock difference between the master and slave nodes. Specifically, in embodiment S4, the single-difference carrier phase equation includes an unknown integer ambiguity single-difference term Δ. N i mu Since it is an integer, it can theoretically be determined through exhaustive search; however, when multiple satellites are combined, the computational cost of the multidimensional search space explodes exponentially; this invention uses the LAMBDA algorithm for fast fixing. In step S4, the carrier phase integer ambiguity in the single-difference carrier phase equation is searched, and the precise integer value of the integer ambiguity is output. This specifically includes the following steps: S41. Extract the covariance matrix of the compensated single-difference carrier phase equation output in step S3; construct the Z-transform integer matrix, and multiply the covariance matrix of the compensated single-difference carrier phase equation on the left and on the right by the Z-transform integer matrix to generate the orthogonal search space matrix after downcorrelation processing. Specifically, in embodiment S41, the compensated single-difference carrier phase equation contains an unknown integer ambiguity single-difference term; due to the covariance between the observation equations of different satellites, the search space is an extremely narrow and tilted multidimensional ellipsoid. The system extracts the covariance matrix of the compensated single-difference carrier phase equation output in step S3, and constructs a multidimensional Z-transform integer matrix using the LAMBDA algorithm. The covariance matrix is ​​then multiplied on the left and right by the Z-transform integer matrix. Through mathematical space mapping, the elongated ellipsoid is transformed into a nearly isotropic orthogonal spherical space, generating an orthogonal search space matrix after decorrelation processing. S42. Input the coarse relative clock difference value output in step S3 into the orthogonal search space matrix for coordinate transformation, and calculate the three-dimensional coordinates of the center of the multidimensional search sphere. Specifically, in embodiment S42, the system inputs the relative clock difference coarse value with nanosecond precision output in step S3 into the orthogonal search space matrix for linear coordinate transformation; this coarse value directly determines the specific position of the multidimensional search sphere in the orthogonal space, and the system outputs the center three-dimensional coordinates of the multidimensional search sphere, which greatly limits the range of blind search; S43. Taking the central three-dimensional coordinates as the origin, set the search radius threshold to 3 times the equivalent distance error corresponding to the relative clock difference coarse value; within the sphere volume limited by the search radius threshold, use a sequential conditional search algorithm to find the first set of integer vectors that minimizes the sum of squared residuals and the second set of integer vectors that have the second smallest sum of squared residuals. Please see Figure 2 Specifically, in embodiment S43, the system uses the center three-dimensional coordinates output in step S42 as the origin and sets the search radius threshold to three times the equivalent distance error corresponding to the coarse relative clock error. For example, if the coarse error is 0.6 meters, the radius threshold is 1.8 meters. The search radius threshold is obtained by extracting the pseudorange error covariance matrix output when solving the coarse relative clock error using the least squares method in step S35, and calculating the square root of its main diagonal elements to obtain the standard deviation. Based on the Laida rule, i.e., the 3σ criterion, of the three-dimensional normal distribution, three times this standard deviation is objectively set as the search radius threshold to ensure that the true optimal solution is covered with a probability of 99.73%. Within this limited spherical volume, the system employs a depth-first sequential conditional search algorithm to quickly traverse all possible combinations of integer nodes, finding the first set of integer vectors that minimizes the sum of squared residuals of the observation equation, which is the most likely correct solution; and the second set of integer vectors that have the second smallest sum of squared residuals, which is the most likely perturbation solution. S44. Calculate the ratio of the second smallest residual sum of squares to the smallest residual sum of squares; determine whether the ratio is greater than the preset confidence threshold of 3.0; if the ratio is greater than 3.0, determine that the first set of integer vectors is an integer value of integer ambiguity; if the ratio is less than or equal to 3.0, determine that the search has failed and discard the observation data of the current epoch. Specifically, in embodiment S44, the confidence threshold is based on the chi-square distribution test model, which calculates the critical ratio that makes the suboptimal solution and the optimal solution have a significant statistical difference under the current number of observed satellites, i.e., degrees of freedom; in order to meet the system's requirement of a fixed error rate of less than one in 100,000, the confidence threshold is obtained by looking up a table and back interpolation. The system calculates the ratio of the second smallest residual sum of squares to the smallest residual sum of squares; if the ratio is greater than 3.0, it indicates that the optimal solution is significantly better than the second best solution, and the system determines that the first set of integer vectors is the exact integer value of the integer ambiguity; if the ratio is less than or equal to 3.0, the system determines that the search has failed and forcibly discards the observation data of the current epoch. Once the exact integer value of the integer ambiguity is successfully fixed, such as 105423 weeks for satellite 1 and 95321 weeks for satellite 2, the integer ambiguity single-difference term in the single-difference carrier phase equation after system substitution becomes a known constant. The system then re-solves the equation using the least squares method, and the resulting variable is the high-precision relative clock difference between the master and slave nodes. At this point, the clock difference accuracy is improved to the 5-10 picosecond level of the carrier phase. This embodiment completely eliminates the strong correlation of the multidimensional search space through the Z-transform matrix of the LAMBDA algorithm, and achieves efficient convergence of the sequential conditional search by using the search radius and confidence threshold based on statistical calculations; it successfully removes the integer ambiguity variable in the carrier phase observation, enabling the system to calculate the picosecond-level relative clock difference using phase data with millimeter-level precision, breaking through the accuracy bottleneck of traditional satellite time synchronization. S5. The master node and slave node generate local calibration pilot signals and inject them into the input ports of their respective RF front-ends. The RF channel group delay of the master node and the RF channel group delay of the slave node are measured respectively. By combining the high-precision relative clock difference between master and slave nodes, the delay of the master node RF channel group, and the delay of the slave node RF channel group, a reference plane conversion is performed to calculate the end-to-end absolute synchronization deviation matrix. Specifically, in embodiment S5, the picosecond-level high-precision relative clock difference calculated in step S4 has its physical reference plane located at the phase center of the GNSS antenna on the top of the radar vehicle; however, the coherent synthesis of the radar signal is triggered in the digital baseband processing chip; from the antenna to the baseband, there are radio frequency coaxial cables, low noise amplifiers (LNAs), mixers, surface acoustic wave filters (SAWs), and analog-to-digital converters (ADCs). These analog RF devices have an inherent group delay, which is subject to drastic changes in ambient temperature, such as the temperature difference between noon and midnight in the desert, and slow nanosecond-level drift due to device aging. Without eliminating this internal delay, picosecond-level synchronization at the antenna is meaningless for the final phased array beamforming. The master node employs the same calibration mechanism as the slave node: a local calibration pilot signal is generated within the master node's FPGA, injected backwards into the master node's RF front-end input port via a directional coupler, and after passing through the complete receive link, the master node's RF channel group delay is measured using a sliding correlation combined with interpolation peak measurement. t gd,m The group delay measurement of the master and slave nodes is completed synchronously in each calibration cycle, and the measurement results are stored in the deviation calculation module of the corresponding node. The step S5, which measures the radio frequency channel group delay value transmitted from the local calibration pilot signal to the digital baseband output port, specifically includes the following steps: S51. Generate Gold polynomial pseudo-random sequences inside the field-programmable gate arrays of the master node and slave node respectively, as their respective local calibration pilot signals; Specifically, in embodiment S51, the picosecond-level high-precision relative clock difference calculated in step S4 has a physical reference plane that remains at the antenna phase center; however, the radar signal is generated by the digital baseband, and the analog radio frequency devices between the antenna and the baseband have a group delay that drifts with temperature; in this embodiment, a Gold polynomial pseudo-random sequence with a length of 2047 chips is generated by using a linear feedback shift register inside the field programmable gate array of the slave node as a local calibration pilot signal; S52. Convert the local calibration pilot signal into an analog radio frequency signal, and use a directional coupler to inject the analog radio frequency signal in reverse into the radio frequency front-end input port of the slave node; Specifically, in embodiment S52, the system converts the digital local calibration pilot signal into an analog radio frequency signal through a digital-to-analog converter; a 20dB directional coupler is connected in series in the physical path; the analog radio frequency signal is injected in reverse into the radio frequency front-end input port of the slave node through the coupling end of the directional coupler; S53. The injected analog radio frequency signal passes through a low-noise amplifier and an intermediate frequency filter in sequence, and is then resampled by an analog-to-digital converter and input to the digital baseband output port to obtain a resampled signal. Specifically, in embodiment S53, the injected analog radio frequency signal is superimposed on the real weak satellite signal coming from the antenna, passes through the low noise amplifier, mixer, and intermediate frequency filter in sequence, undergoes the same physical delay process as the real satellite signal, and is finally resampled by the high-speed analog-to-digital converter at a sampling rate of 1 GSPS and input to the digital baseband output port to obtain the resampled signal. S54. Perform a sliding correlation operation between the resampled signal and the originally generated Gold polynomial pseudo-random sequence; find the number of sliding clock cycles when the correlation peak occurs; multiply the number of sliding clock cycles by the system clock cycle duration to calculate the RF channel group delay value; Specifically, in embodiment S54, in the digital baseband, the system performs a sliding correlation operation on the resampled signal and the originally generated Gold polynomial pseudo-random sequence; finds the number of sliding clock cycles when the correlation peak occurs, such as a delay of 452 clock cycles; the system multiplies the number of sliding clock cycles by the system clock cycle duration, such as a 100MHz clock with a period of 10 nanoseconds, to calculate the RF channel group delay value, such as 4520 nanoseconds; The master node performs the exact same measurement procedure as the slave node to obtain the master node RF channel group delay; To further improve the measurement accuracy of RF channel group delay, after obtaining the coarse peak position through sliding correlation, a subsampling-level interpolation peak measurement step is added. The sliding correlation operation clock is switched from the 100MHz system clock to the 1GSPS ADC sampling clock, and the coarse measurement step size is compressed from 10ns to 1ns, initially improving the time resolution. The correlation peak point and the correlation amplitude of the two adjacent sampling points are extracted. The envelope curve of the correlation peak is fitted by a three-point parabolic interpolation algorithm to obtain the offset of the non-integer sampling point corresponding to the peak, further improving the delay measurement resolution to 1 / 10 to 1 / 20 of the sampling period, i.e., 50 to 100ps. At the same time, the Gold sequence chip rate can be increased to over 100Mcps, and the interpolation error can be further reduced by using a narrower correlation main peak. Finally, the group delay measurement stability can be better than 20ps, meeting the calibration requirements of picosecond-level synchronization. This embodiment uses a directional coupler to inversely inject the pseudo-random pilot signal generated by the digital baseband into the analog RF front end; by using sliding correlation calculation, the actual physical time of the signal traveling through the internal devices is accurately measured, breaking the black box state of the traditional RF link, realizing real-time dynamic monitoring of temperature drift and device aging, and providing extremely critical quantitative compensation parameters for subsequent elimination of internal hardware delay and end-to-end synchronization. Further, in step S5, the end-to-end absolute synchronization deviation matrix is ​​calculated, specifically including the following steps: S55. Read the high-precision relative clock difference between the master and slave nodes calculated in step S4 as the reference time parameter, and read the RF channel group delay value calculated in step S54 as the internal subtraction parameter. Specifically, in embodiment S55, the system synchronously reads the high-precision relative clock difference between the master and slave nodes calculated in step S4 and the radio frequency channel group delay value calculated in real time in step S54 in the digital baseband processing unit. S56. Combine the high-precision relative clock difference between the master and slave nodes with the delay of the master node's RF channel group and the delay of the slave node's RF channel group to perform a reference plane conversion. Then, shift the physical reference plane for time synchronization from the antenna phase center of the master and slave nodes to their respective digital baseband processing chips to generate an initial absolute synchronization deviation sequence. Specifically, in embodiment S56, the system calls a specific mathematical formula to calculate. ;in E abs Δ represents the scalar numerical components in the absolute synchronization deviation matrix. δt mup This indicates the high-precision relative clock difference between the master and slave nodes; t gd,m Indicates the delay of the master node radio frequency channel group; t gd,u To reduce the delay of the radio frequency channel group from the node; through this subtraction operation, the system shifts the physical reference plane for time synchronization from the external antenna to the digital baseband, generating an initial absolute synchronization deviation sequence; S57. Construct a sliding mean filter with a length of 10 epochs, input the initial absolute synchronization deviation sequence into the sliding mean filter to remove transient spikes and jumps, and output a smoothed end-to-end absolute synchronization deviation matrix. Specifically, in embodiment S57, the end-to-end absolute synchronization deviation matrix is ​​denoted as... E abs Let be a 1×M dimensional row vector, where M is the total number of slave nodes in the system; the k-th element in the matrix corresponds to the end-to-end time deviation between the master node and the k-th slave node, expressed as: ; where Δ δt mu ( k ) indicates the relationship between the primary node and the second node. k The high-precision relative clock difference is obtained from the satellite common-view solution of each node, with the physical reference plane being the phase center of the two GNSS antennas at both ends; t gd,m This represents the delay of the master node's radio frequency receiving channel group, i.e., the transmission delay from the antenna phase center to the digital baseband sampling port; t gd,u ( k ) represents the group delay of the radio frequency receive channel of the k-th slave node; Furthermore, the synchronization deviation matrix achieves precise translation of the time synchronization reference plane, moving the antenna aperture reference plane of the traditional scheme down to the inside of the digital baseband processing chip, eliminating the group delay black box error of the RF analog link, and providing a true end-to-end deviation representation covering the entire signal path; the matrix elements are signed scalar time deviations, which serve as the common input basis for subsequent nanosecond-level crystal oscillator discipline and picosecond-level phase rotation two-stage compensation, determining the control output of the dual closed loop; the matrix fully represents the absolute time difference between the master node and all slave nodes at the digital baseband level, and is the state parameter for the distributed phased array to achieve spatial coherent power synthesis; This embodiment deeply integrates the solution results of the spatial link with the calibration results of the local hardware, completing the translation of the time reference plane from the outside to the inside; at the same time, a sliding mean filter is introduced to smooth the difference sequence, effectively filtering out transient jumps caused by occasional noise, and the output absolute synchronization deviation matrix more realistically reflects the steady-state synchronization error of the system, improving the robustness of subsequent hardware closed-loop control. S6. Divide the absolute synchronization deviation matrix into a nanosecond-level integer deviation part and a picosecond-level fractional deviation part; input the nanosecond-level integer deviation part into a digital phase-locked loop to generate a voltage-controlled oscillator control voltage, so as to tame the local temperature-controlled crystal oscillator of the node in real time; The picosecond-level fractional deviation is converted into a frequency domain phase rotation factor, and the frequency domain phase rotation factor is used in the digital baseband to perform complex multiplication phase compensation on the radar waveform data of the slave node. Specifically, in embodiment S6, after the system obtains the end-to-end absolute synchronization deviation, such as a total deviation of 125.432 nanoseconds, it needs to feed it back to the hardware actuator for physical correction. Considering the voltage regulation resolution limit of the temperature-controlled crystal oscillator OCXO and the flexibility of digital baseband compensation, the present invention adopts a dual closed-loop compensation mechanism that combines coarse adjustment and fine adjustment. Furthermore, it should be noted that the application background of this invention is a distributed digital phased array radar system. In phased array radar engineering practice, each radio frequency channel is usually designed as a time-division duplex or uses a transmit-receive isolated duplexer. In this invention, the radio frequency front-end of each node adopts a zero-IF or superheterodyne architecture with integrated transmit and receive and a shared local oscillator. Its up-conversion and down-conversion links share the same physical path and filter bank. The group delay of the analog links for receiving and transmitting has extremely high reciprocity in physical terms. Therefore, the group delay value obtained by measuring the receiving link can be used to characterize the physical delay of the transmitting link at the same port. The synchronization target of this invention is the triggering and data generation time of the digital baseband; the receiving channel group delay measured in step S5 is essentially to accurately shift the time reference plane from the antenna phase center to the inside of the baseband chip; once the reference plane is established inside the baseband, when performing phase rotation compensation on the radar waveform data in step S6, the equivalent time of the baseband signal is shifted as a whole, which already includes the expected fixed delay of the transmission link. Please see Figure 3 In step S6, the nanosecond-level integer deviation is input into a digital phase-locked loop to generate a voltage-controlled oscillator control voltage, so as to tame the node's local temperature-controlled crystal oscillator in real time. This specifically includes the following steps: S61. Read the nanosecond-level integer deviation portion of the absolute synchronization deviation matrix; Specifically, in embodiment S61, the system obtains the absolute synchronization deviation matrix. For example, if the total deviation is 125.432 nanoseconds, the integer part is extracted as the nanosecond-level integer deviation part, i.e., 125 nanoseconds. S62. Construct a proportional-integral controller model that includes proportional gain coefficient and integral gain coefficient; Specifically, in embodiment S62, the system constructs a proportional-integral controller model containing a proportional gain coefficient and an integral gain coefficient inside the FPGA; the proportional gain coefficient is set to 0.05 to provide fast error response correction; the integral gain coefficient is set to 0.001 to eliminate long-term static errors. To further explain, the proportional gain coefficient and integral gain coefficient in this embodiment are tuned using a step response identification combined with a phase margin constraint method. First, a standard voltage step signal is injected into the voltage-controlled pin of the local temperature-controlled crystal oscillator, and the dynamic response curve of the local clock frequency is collected to identify the first-order inertial equivalent model of the voltage-controlled channel. Then, with a phase margin of not less than 45° and no steady-state error as constraints, the PI parameter tuning is completed in the discrete digital domain. Finally, the proportional gain coefficient is determined to be 0.05 and the integral gain coefficient is 0.001, which takes into account both the disciplined response speed and the stability without overshoot. S63. Input the nanosecond-level integer deviation portion into the proportional-integral controller model, and calculate the control voltage increment command by summing the proportional term and the integral term; Specifically, in embodiment S63, the system inputs the nanosecond-level integer deviation portion to the proportional-integral controller model and utilizes the difference equation. Calculate the control voltage increment command; where Δ V ( k ) indicates the current number k The control voltage increment command output in each control cycle; K p Indicates the proportional gain coefficient; E int ( k ) indicates the current number k The nanosecond-level integer deviation portion of the input for each control cycle; K i Indicates the integral gain coefficient; This indicates the period from the first cycle to the current cycle. k The cumulative sum of nanosecond-level integer deviations over each cycle; S64. The control voltage increment command is superimposed on the base voltage of the previous cycle, and the analog control voltage is output to the voltage control pin of the local temperature-controlled crystal oscillator through the digital-to-analog converter chip to change the crystal resonant frequency to eliminate the nanosecond-level integer deviation. Specifically, in embodiment S64, the system superimposes the above-mentioned control voltage increment command onto the base voltage of the previous cycle, and outputs the actual analog control voltage to the voltage control pin of the local temperature-controlled crystal oscillator through a 16-bit high-precision digital-to-analog converter chip; by changing the control voltage, the piezoelectric resonant frequency inside the crystal is finely adjusted, so that the local clock frequency of the slave node is accelerated or decelerated, gradually leveling out and eliminating the 125 nanosecond macroscopic time deviation. This embodiment utilizes a classic proportional-integral controller model to convert the nanosecond-level macroscopic part of the absolute synchronization deviation into an analog voltage command, which acts on the local temperature-controlled crystal oscillator. This control loop changes the oscillation frequency of the slave node clock from the physical level, smoothly and continuously eliminating the time accumulation drift between the master and slave nodes, and ensuring the high consistency and operational stability of the distributed system on the macroscopic time scale. Further, in step S6, the picosecond-level fractional deviation is converted into a frequency-domain phase rotation factor, and the frequency-domain phase rotation factor is used in the digital baseband to perform complex multiplication phase compensation on the radar waveform data of the slave node. Specifically, this includes the following steps: S65. Read the picosecond-level fractional deviation portion of the absolute synchronization deviation matrix; Specifically, in embodiment S65, the system extracts the fractional part of the absolute synchronization deviation matrix as the picosecond-level fractional deviation part, i.e., 0.432 nanoseconds, which is equal to 432 picoseconds; this is the microscopic residual error that the temperature-controlled crystal oscillator cannot smoothly respond in a short time. S66. Extract the specific frequency value of the current working radio frequency carrier of the distributed digital phased array; Specifically, in embodiment S66, the system extracts the specific frequency value of the current working radio frequency carrier of the distributed digital phased array, such as the X-band radar carrier frequency being 10 GHz and the period being 100 picoseconds. S67. Multiply the specific frequency value by the picosecond-level decimal deviation part, and multiply by the pi constant factor to calculate the radian value of the frequency domain phase rotation factor. Specifically, in embodiment S633, the system calculates the radian value of the frequency domain phase rotation factor using a specific physical formula, the formula being: ;in i The frequency domain phase rotation factor is... F c The operating carrier frequency for the distributed phased array is 10 GHz in S66; E frac This represents the decimal part at the picosecond level, in seconds; substitute the value. i =2π×10×10 9 ×(432×10 -12= 2π × 4.32 radians. Based on the 2π periodicity of the phase, the 8π integer phase corresponding to the 4 complete periods is removed, and the final effective compensation phase is 0.64π radians. S68. Input the radian value of the frequency domain phase rotation factor into the coordinate rotation digital computer algorithm IP core to generate a complex rotation matrix containing in-phase cosine components and orthogonal sine components; perform point-by-point complex multiplication operation between the radar waveform complex data stream being processed and the complex rotation matrix to cancel out the picosecond-level decimal deviation in physical time. Specifically, in embodiment S644, the system inputs the radian value of the frequency domain phase rotation factor 0.32π into the internal coordinate rotation digital computer algorithm IP core to iteratively generate a complex rotation matrix containing in-phase cosine components and orthogonal sine components. When the node baseband is generating a complex data stream of radar waveforms ready for transmission, this batch of high-speed data streams is multiplied point by point with the complex rotation matrix. Through mathematical rotation on the complex plane, a phase lead or lag is artificially introduced into the digital signal. This digital phase shift cancels out the 432 picosecond delay that is difficult to eliminate in physical time after passing through the transmission channel. Through the closed-loop control combining macroscopic physical crystal oscillator discipline and microscopic digital baseband phase rotation, this invention successfully squeezes the synchronization error of each independent node to within 5 picoseconds in a field deployment environment that completely abandons optical fiber cables, thus meeting the stringent conditions for phased array spatial power coherent synthesis. To address the picosecond-level micro-errors that cannot be eliminated by hardware crystal oscillators, this embodiment fully leverages the advantages of digital signal processing, converting them into frequency domain phase rotation factors, and performing point-by-point complex multiplication operations in the baseband data stream using the CORDIC algorithm. This compensation mechanism achieves phase alignment, ultimately meeting the physical requirements for spatial power coherent synthesis of distributed phased arrays in the microwave band.

[0024] For further details, please refer to Figure 4 A satellite common-view time synchronization system for distributed digital phased arrays, used to implement a satellite common-view time synchronization method for distributed digital phased arrays; the system includes: The device consists of a master node and at least one slave node; both the master node and the slave node are equipped with a dual-band GNSS antenna, an RF front-end receiving module, and a baseband signal processing module. The baseband signal processing module includes a field-programmable gate array (FPGA) chip; the FPGA chip is internally configured with a satellite signal acquisition and tracking logic unit, a timestamp latch generation logic unit, a digital phase-locked loop control logic unit, and a complex multiplication phase compensation logic unit; The master node device and the slave node device exchange the observation data required for single difference calculation via a wireless microwave data link; This invention breaks away from the inherent dependence of traditional distributed phased arrays on physical fiber optic connections and improves the accuracy of wireless satellite timing. This invention observes that to achieve spatial coherent synthesis in the microwave band, it is necessary not only to solve the picosecond-level comparison problem of the external space link but also to eliminate the delay drift of the internal analog radio frequency devices due to environmental changes. Therefore, this invention integrates carrier phase single-difference calculation and integer ambiguity fast fixing technology in the external space link, improving the accuracy of satellite navigation signals. In the internal hardware link, a self-calibration mechanism through the reverse injection of pseudo-random pilot signals makes the originally closed radio frequency link measurable. Through the deep integration of external comparison and internal calibration, coupled with the dual actuators of crystal oscillator macroscopic discipline and baseband microscopic phase rotation, this invention constructs a fully closed-loop, cableless, end-to-end high-precision time synchronization system. This liberates the physical form of the distributed array, giving it extremely high mobility and damage resistance.

[0025] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0026] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A satellite common-view time synchronization method for distributed digital phased arrays, characterized in that, Includes the following steps: S1. Obtain the real-time elevation angle, signal-to-noise ratio, and position accuracy factor of the visible navigation satellites, calculate the scores using the satellite quality assessment function, and filter them to obtain the common-view satellite set; S2. Lock the dual-frequency navigation signal of the common-view satellite set, extract pseudorange and carrier phase observation values, and use a global counter to mark the local time to construct the observation dataset of the master node and the slave node. S3. Perform single-difference calculation on the master node observation dataset and the slave node observation dataset, and call the atmospheric delay compensation model to compensate according to the spatial distance between the master node and the slave node, so as to obtain the relative clock difference coarse value and the single-difference carrier phase equation. S4. Based on the coarse value of relative clock difference, the integer ambiguity of the carrier phase is searched through the integer ambiguity algorithm to determine the precise integer value of the integer ambiguity, and then substituted back into the single-difference carrier phase equation to obtain the high-precision relative clock difference between the master and slave nodes. S5. Measure the radio frequency channel group delay of the master node and slave node respectively, and use the high-precision relative clock difference to perform reference plane conversion to obtain the end-to-end absolute synchronization deviation matrix. S6. Divide the absolute synchronization deviation matrix into a nanosecond-level integer deviation part and a picosecond-level fractional deviation part; The nanosecond-level integer deviation is input into a digital phase-locked loop to generate a voltage-controlled voltage to dissipate the node’s local temperature-controlled crystal oscillator in real time. The picosecond-level fractional deviation is converted into a frequency domain phase rotation factor, and phase compensation is performed on the slave node radar waveform data in the digital baseband through complex multiplication.

2. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 1, wherein the satellite quality evaluation function in step S1 calculates scores and selects satellites for the common-view set, specifically includes the following steps: S11, Read the... i Based on the real-time elevation angle data, real-time signal-to-noise ratio data, and position accuracy factor of the visible navigation satellites, an initial candidate satellite list is established; S12, Set the elevation angle threshold; Determine if the real-time elevation angle data is less than the elevation angle threshold; if so, directly set the threshold value to the next highest value. i Each visible navigation satellite is assigned a score of 0 and removed from the initial candidate satellite list, resulting in an updated candidate satellite list. Otherwise, proceed to step S13 to calculate the score and retain it in the updated candidate satellite list; S13. Call the satellite quality assessment function to calculate the score; the satellite quality assessment function is as follows: divide the real-time elevation angle data and the real-time signal-to-noise ratio data by their respective physical extreme values ​​for positive normalization and multiply by the first positive weighting coefficient; divide the position accuracy factor by the physical extreme value for negative normalization and multiply by the second negative penalty coefficient; add the positive normalization result and the negative normalization result to obtain the score. S14. Sort all visible navigation satellites with scores greater than 0 in the updated candidate satellite list in descending order of scores, extract the numbers of the top N visible navigation satellites, and generate a common-view satellite set.

3. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 1, wherein step S3 involves performing single-difference calculation on the master node observation dataset and the slave node observation dataset, and invoking an atmospheric delay compensation model for compensation based on the spatial distance between the master node and the slave node, specifically including the following steps: S31. Read the three-dimensional geodetic coordinates of the master node and the slave node, and use the Euclidean distance algorithm to calculate the three-dimensional straight-line distance between the master node and the slave node; set the baseline cancellation threshold. S32. Determine whether the straight-line distance in three-dimensional space is less than or equal to the baseline cancellation threshold. If so, the ionospheric delay single-difference term and the tropospheric delay single-difference term in the initial single-difference carrier phase equation are set to 0 to obtain the compensated single-difference carrier phase equation. Otherwise, extract real-time meteorological data from the locations of the master and slave nodes, input the real-time meteorological data into the Hopfield tropospheric model to calculate the absolute delay and subtract them to obtain the tropospheric delay single-difference compensation value; substitute the tropospheric delay single-difference compensation value into the initial single-difference carrier phase equation for subtraction to obtain the compensated single-difference carrier phase equation.

4. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 3, wherein the calculation of the relative clock difference coarse value and the single-difference carrier phase equation in step S3 specifically includes the following steps: S33. Construct an initial single-difference carrier phase equation that includes a single difference term for carrier phase observations, a single difference term for geometric distance, a single difference term for relative clock error, a single difference term for ionospheric delay, a single difference term for tropospheric delay, a single difference term for integer ambiguity, and a noise term. S34. Extract the dual-frequency pseudorange observation sequence obtained in step S2, replace the carrier phase observation sequence in the initial single-difference carrier phase equation with the dual-frequency pseudorange observation sequence, and remove the integer ambiguity single-difference term in the equation to generate an ambiguity-free pseudorange observation equation. S35. Substitute the geometric coordinates of the master node and slave node into the ambiguity-free pseudorange observation equation, and solve iteratively using the least squares method to output the coarse relative clock error value at the nanosecond level.

5. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 1, wherein step S4 involves searching for carrier phase integer ambiguity using an integer ambiguity algorithm model and outputting precise integer values ​​for integer ambiguity, specifically including the following steps: S41. Extract the covariance matrix of the compensated single-difference carrier phase equation output in step S3; construct the Z-transform integer matrix, and multiply the covariance matrix of the compensated single-difference carrier phase equation on the left and on the right by the Z-transform integer matrix to generate the orthogonal search space matrix after downcorrelation processing. S42. Input the coarse relative clock difference value output in step S3 into the orthogonal search space matrix for coordinate transformation, and calculate the three-dimensional coordinates of the center of the multidimensional search sphere. S43. Using the central three-dimensional coordinates as the origin, set a search radius threshold; within the sphere volume limited by the search radius threshold, use a sequential conditional search algorithm to find the first set of integer vectors that minimizes the sum of squared residuals and the second set of integer vectors that have the second smallest sum of squared residuals. S44. Calculate the ratio of the second smallest residual sum of squares to the smallest residual sum of squares; determine whether the ratio is greater than the preset confidence threshold; if so, determine that the first set of integer vectors is an integer value of integer ambiguity; otherwise, determine that the search has failed and discard the observation data of the current epoch.

6. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 1, wherein step S5, measuring the radio frequency channel group delay of the master node and the slave node respectively, specifically includes the following steps: S51. Generate a Gold polynomial pseudo-random sequence inside the field-programmable gate array of the slave node as a local calibration pilot signal; S52. Convert the local calibration pilot signal into an analog radio frequency signal, and use a directional coupler to inject the analog radio frequency signal in reverse into the radio frequency front-end input port of the slave node; S53. The injected analog radio frequency signal passes through a low-noise amplifier and an intermediate frequency filter in sequence, and is then resampled by an analog-to-digital converter and input to the digital baseband output port to obtain a resampled signal. S54. Perform a sliding correlation operation between the resampled signal and the originally generated Gold polynomial pseudo-random sequence. Find the number of sliding clock cycles when the relevant peak occurs; multiply the number of sliding clock cycles by the system clock cycle duration to calculate the RF channel group delay value.

7. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 6, wherein calculating the end-to-end absolute synchronization deviation matrix in step S5 specifically includes the following steps: S55. Read the high-precision relative clock difference between the master and slave nodes calculated in step S4 as the reference time parameter, and read the RF channel group delay value calculated in step S54 as the internal subtraction parameter. S56. Combine the high-precision relative clock difference between the master and slave nodes with the delay of the master node's RF channel group and the delay of the slave node's RF channel group to perform a reference plane conversion. Then, shift the physical reference plane for time synchronization from the antenna phase center of the master and slave nodes to their respective digital baseband processing chips to generate an initial absolute synchronization deviation sequence. S57. Construct a sliding mean filter, input the initial absolute synchronization deviation sequence into the sliding mean filter to remove transient spikes and jumps, and output a smoothed end-to-end absolute synchronization deviation matrix.

8. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 1, wherein step S6, inputting the nanosecond-level integer deviation portion into a digital phase-locked loop to generate a voltage-controlled oscillator control voltage to tame the node's local temperature-controlled crystal oscillator in real time, specifically includes the following steps: S61. Read the nanosecond-level integer deviation portion of the absolute synchronization deviation matrix; S62. Construct a proportional-integral controller model that includes proportional gain coefficient and integral gain coefficient; S63. Input the nanosecond-level integer deviation portion into the proportional-integral controller model, and calculate the control voltage increment command by summing the proportional term and the integral term; S64. The control voltage increment command is superimposed on the base voltage of the previous cycle, and the analog control voltage is output to the voltage control pin of the local temperature-controlled crystal oscillator through the digital-to-analog converter chip to change the crystal resonant frequency to eliminate the nanosecond-level integer deviation.

9. The satellite common-view time synchronization method for distributed digital phased arrays according to claim 1, wherein step S6 converts the picosecond-level fractional deviation into a frequency-domain phase rotation factor, and uses the frequency-domain phase rotation factor to perform complex multiplication phase compensation on the radar waveform data of the slave node in the digital baseband, specifically including the following steps: S611. Read the picosecond-level fractional deviation portion of the absolute synchronization deviation matrix; S622. Extract the specific frequency value of the current working radio frequency carrier of the distributed digital phased array; S633. Multiply the specific frequency value by the picosecond-level decimal deviation part, and multiply by the pi constant factor to calculate the radian value of the frequency domain phase rotation factor. S644. Input the radian value of the frequency domain phase rotation factor into the coordinate rotation digital computer algorithm IP core to generate a complex rotation matrix containing in-phase cosine components and orthogonal sine components; perform point-by-point complex multiplication operation between the radar waveform complex data stream being processed and the complex rotation matrix to cancel out the picosecond-level decimal deviation in physical time.

10. A satellite common-view time synchronization system for distributed digital phased arrays, used to implement the method of any one of claims 1 to 9; the system comprising: The device consists of a master node and at least one slave node; both the master node and the slave node are equipped with a dual-band GNSS antenna, an RF front-end receiving module, and a baseband signal processing module. The baseband signal processing module includes a field-programmable gate array (FPGA) chip; the FPGA chip is internally configured with a satellite signal acquisition and tracking logic unit, a timestamp latch generation logic unit, a digital phase-locked loop control logic unit, and a complex multiplication phase compensation logic unit; The master node device and the slave node device exchange observation data required for single difference calculation via a wireless microwave data link.

Images