Dual feedback symbol timing recovery method for non-los star-earth slmimo systems

Abstract

The application discloses a double feedback symbol timing recovery method for a non-identical star-ground line-of-sight MIMO system and belongs to the technical field of satellite communication and digital signal processing. In the application, each ground station receiving branch is controlled by a digital controlled oscillator to perform interpolation resampling, and a frequency domain conversion, a matched filtering and a bias-timing phase correction are completed by using an overlap-save method; a pair of spectrums with a half fast Fourier transform period interval is extracted in a roll-off area, a conjugate de-weighted accumulation is performed on the spectrums, and a phase angle is taken to obtain a timing phase deviation of each receiving branch; the timing phase deviation is divided into two feedback paths after being smoothed in a complex domain, one path generates a frequency domain rotation coefficient to perform a fine phase correction, and the other path extracts a timing frequency deviation through a loop filter to drive a digital controlled oscillator and a time domain resampler to continuously track. The application can independently calibrate each receiving clock under the conditions of a star-ground line-of-sight multiple input multiple output inherent channel interference and a non-identical multi-station clock without pilot assistance.

Description

Technical Field

[0001] This invention belongs to the field of satellite communication and digital signal processing technology, specifically relating to a dual-feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems. Background Technology

[0002] With the rapid development of applications such as remote sensing, meteorological monitoring, and Earth observation, the amount of data generated and transmitted back daily by low-Earth orbit (LEO) satellites continues to increase. Traditional Ka / Ku bands and existing satellite-to-ground link technologies are gradually becoming insufficient to meet the ultra-high-speed data transmission requirements of high-resolution payloads. After the International Telecommunication Union (ITU) allocated the 71–76 GHz and 81–86 GHz frequency bands for satellite communications, utilizing E-band ultra-wideband resources to increase satellite-to-ground transmission bandwidth has become a feasible approach. Satellite-to-ground propagation is primarily line-of-sight (LOS) propagation, with relatively weak multipath components. Line-of-sight multiple-input multiple-output (LOI) technology, by rationally configuring the spacing between satellite antennas and multiple ground stations, constructs spatial degrees of freedom in ultra-long LOS links, significantly improving spectral efficiency. However, to obtain approximately orthogonal transmission channels, the spatial spacing between ground receiving stations typically reaches tens of kilometers, making it difficult for multiple receiving stations to share the same sampling clock; the Doppler spread differences caused by the high-speed motion of LEO satellites further cause the sampling deviation of each station to vary over time.

[0003] Clock non-homogeneity and Doppler spread differences result in varying timing phase and timing frequency deviations at different receiving stations. If these deviations are not corrected separately, inter-symbol interference (ISI) will occur. In Los-of-Sight (Line-of-Sight Multiple-Input Multiple-Output) joint reception, signals from multiple satellite-borne transmit antennas will also generate inter-channel interference (ICI) related to transmission delay differences. Symbol timing recovery is located in the pre-stage of the digital baseband receiver, and its performance directly affects the throughput and reliability of subsequent channel estimation, equalization, and MIMO detection.

[0004] Existing non-data-aided timing detection algorithms include the Mueller-Müller algorithm, the Gardner algorithm, and the Oerder-Meyr algorithm. The time-domain-based Gardner algorithm is easy to implement, but its accuracy decreases in high-order modulation scenarios. Frequency-domain-based Oerder-Meyr-like algorithms have higher accuracy, but typically require higher sampling rates and greater computational complexity. Frequency-domain estimation based on Barton's ideas can obtain the timing phase using the conjugate relationship of the root-raised cosine roll-off region, but the traditional normalized weighting process involves modulus calculation and division, which is not conducive to ultra-high-speed parallel hardware implementation.

[0005] Existing feedback-based timing recovery structures offer good steady-state accuracy, but converge slowly under large timing frequency deviations. Feedforward structures converge quickly, but struggle to continuously track dynamic sampling deviations. Existing time-frequency dual-feedback schemes combine frequency domain estimation with time-domain resampling to improve the resilience of single-input single-output broadband communication systems to timing frequency deviations; however, these algorithms have not yet addressed the issues arising from multiple non-coordinated ground stations and inherent inter-channel interference in LoS-MIMO.

[0006] On the other hand, existing MIMO timing synchronization largely relies on pilots or training sequences, employing time-domain correlation, singular value decomposition, or least squares methods to jointly perform synchronization and channel equalization. The resilience of these data-assisted methods to timing frequency deviations is closely related to the length and period of the training sequence, making them unsuitable for scenarios with extremely large deviations and independent drift at each receiving station; increasing the pilot ratio also reduces transmission efficiency. Schemes based on beamforming or antenna delay pre-compensation are strongly coupled to specific channel and geometric conditions, resulting in insufficient adaptability.

[0007] Therefore, there is an urgent need in this field for a symbol timing recovery method that does not require multiple stations to share a clock, does not rely on pilots, can suppress inherent inter-channel interference of LoS-MIMO, and can continuously track large timing frequency deviations, while making its computational structure suitable for parallel implementation of broadband signals above 3Gsym / s. Summary of the Invention

[0008] To address the aforementioned problems in the existing technology, this invention provides a dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems. The technical problem to be solved by this invention is achieved through the following technical solution: In a first aspect, the present invention provides a dual-feedback symbol timing recovery method for non-homogeneous satellite-to-ground line-of-sight MIMO systems, applicable to systems with... N b A single-origin spaceborne launch channel M r A satellite-to-ground line-of-sight MIMO demodulation system with multiple non-homogeneous ground receiving channels and a centralized processing node, the dual-feedback symbol timing recovery method includes: S1, generate the control word for the numerically controlled oscillator based on the symbol rate, input sampling rate and timing frequency deviation of the loop feedback, and use the control word to interpolate and resample the baseband signal of the ground receiving channel to generate a resampled signal; S2, the resampled signal is divided into blocks according to the overlap-preservation method, and each overlap-preservation data block is executed sequentially. N The frequency domain signal of the ground receiving channel is obtained by performing point fast Fourier transform, frequency domain matched filtering, and offset-timing phase correction. S3, extract the phase gap from the positive and negative roll-off regions of the frequency domain signal. NThe pair of spectra at two frequency points are combined, and the pair of spectra are conjugate unweighted accumulation. The timing phase deviation estimate of the ground receiving channel is obtained based on the phase angle of the accumulation result. S4, the estimated timing phase deviation is smoothed by first-order infinite impulse response to obtain a smoothed timing phase deviation; S5, the smoothed timing phase deviation is divided into two feedback paths, which are fed back to the bias-timing phase correction and the numerically controlled oscillator respectively, to form a dual feedback closed loop; S6 performs an inverse fast Fourier transform on the frequency domain signal that has completed bias-timing phase correction to extract the effective part of the overlapping and retained data block, and obtains the output signal that has completed symbol timing recovery.

[0009] In one embodiment of the present invention, M r Each non-similar ground receiving channel is formed by separate ground stations that do not share a sampling clock. Each ground station transmits the baseband signal to the centralized processing node; m The ground receiving channel receives the first n Normalized timing phase deviation of signals from each satellite-borne transmission channel satisfy: ; in, For the first m The normalized timing phase deviation of the signal received by each ground receiving channel from the first satellite-borne transmitting channel. These are the discrete time points after resampling. For mold taking operation, For the first n From the first satellite launch channel to the second... m The transmission delay difference of each ground receiving channel This is the sampling rate corresponding to twice the symbol rate.

[0010] In one embodiment of the present invention, the interpolation resampling process is completed using a fourth-order Lagrange interpolator with a Farrow structure; m One ground receiving channel t Resampled signal at time 1 The expression is as follows: ; in, For the first m The input digital baseband signal of each receiving channel As the interpolation reference position, The first Lagrange interpolation coefficient matrix p Line 1 q Column elements, for t Time of the first The interpolation interval corresponding to each frequency domain data block.

[0011] In one embodiment of the present invention, S2 includes: The resampled signal is divided into blocks according to the overlap-preservation method to obtain overlap-preserved data blocks; Convert each overlapping reserved data block into N The point frequency domain signal is passed through a root-raised cosine matched filter to obtain the joint frequency response of the shaping filter and the matched filter. Offset-timing phase correction is then performed using the frequency domain rotation coefficient generated by the first feedback path to obtain the frequency domain signal for the ground receiving channel.

[0012] In one embodiment of the present invention, the first m The ground receiving channel corresponds to the first... κ The estimated timing phase deviation of each estimated segment The expression is as follows: ; in, L The number of frequency domain data blocks included in a single timing deviation estimate. , For the effective roll-off half-width, The root-raised cosine roll-off factor, For the first m The ground receiving channel is in the first κ In the estimated segment, the first... l Frequency domain signal at time 10:00 It is a complex conjugate.

[0013] In one embodiment of the present invention, the smoothed timing phase deviation is divided into two feedback paths, which are respectively fed back to the bias-timing phase correction and the numerically controlled oscillator to form a dual feedback closed loop, including: The smooth timing phase deviation is divided into two feedback paths. The first feedback path is converted into a frequency domain rotation coefficient and fed back to the bias-timing phase correction. The second feedback path is filtered through a loop to extract the timing frequency deviation and fed back to the numerically controlled oscillator, thus forming a dual feedback closed loop that combines fine frequency domain correction with continuous time domain tracking.

[0014] In one embodiment of the present invention, the loop filter adopts a proportional-integral structure and updates the first loop filter according to the following formula: m Timing frequency deviation of each ground receiving channel: ; in, For the first m The ground receiving channel corresponds to the first... κ The timing frequency deviation of each estimated segment, For the firstm The ground receiving channel corresponds to the first... The timing frequency deviation of each estimated segment, This is the proportional branch coefficient. For the first m The ground receiving channel corresponds to the first... κ The smoothed timing phase deviation of each estimated segment, For the first m The ground receiving channel corresponds to the first... Smooth timing phase deviation of each estimated segment, This is the integral branch coefficient.

[0015] In a second aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of the dual feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems provided in the first aspect of the present invention.

[0016] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the steps of the dual-feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems provided in the first aspect of the present invention.

[0017] The beneficial effects of this invention are: The solution provided in this invention addresses the LoS-MIMO scenario where multiple ground stations cannot share a clock. It establishes an independent closed loop for each receiving branch, eliminating the need for cross-site clock deployment to complete timing calibration before joint reception. The proposed conjugate unweighted estimation method, compared to traditional conjugate normalized weighted estimation, eliminates modulus and division operations. Furthermore, it improves estimation gain by preserving strong spectral components under low signal-to-noise ratio conditions, reducing the estimation variance by approximately 4 dB in noise-free scenarios during simulation. Employing a dual-feedback structure of frequency-domain fine correction and time-domain resampling, it can continuously track timing frequency deviations on the order of kiloppm under reasonable estimation segmentation and loop parameters, while reducing residual timing jitter. Moreover, this invention utilizes Fast Fourier Transform, parallel frequency point accumulation, Farrow interpolation, and independent receiving branches, making it suitable for parallel implementation on FPGAs or dedicated integrated circuits, and can serve E-band broadband satellite data transmission above 3Gsym / s. Attached Figure Description

[0018] Figure 1A schematic diagram of the steps of a dual-feedback symbol timing recovery method for a non-same-source satellite-to-ground line-of-sight MIMO system provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a dual-feedback symbol timing recovery method for a non-same-source satellite-to-ground line-of-sight MIMO system provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of a satellite-to-ground LosS-MIMO communication scenario provided in an embodiment of the present invention; Figure 4 This is a diagram of the multi-receiver branch dual-feedback symbol timing recovery architecture provided in an embodiment of the present invention; Figure 5 This is a simulation diagram of the estimated mean and error of the timing deviation detection algorithm under 4×4 MIMO conditions provided in this embodiment of the invention; Figure 6 This is an estimated variance plot of different timing deviation detection methods under 4×4 MIMO conditions provided in the embodiments of the present invention; Figure 7 This is a diagram illustrating the impact of transmission delay difference on the timing frequency deviation resistance and timing jitter variance of the symbol timing recovery (STR) loop, as provided in this embodiment of the invention. Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0020] To achieve the desired objectives, embodiments of the present invention provide a dual-feedback symbol timing recovery method, electronic device, and storage medium for non-same-source satellite-to-ground line-of-sight MIMO systems.

[0021] Below, we will first introduce a dual-feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems provided by an embodiment of the present invention.

[0022] like Figure 1 and Figure 2 As shown in the embodiment of the present invention, a dual-feedback symbol timing recovery method for non-similar satellite-to-ground line-of-sight MIMO systems is applied to systems with... N b A single-origin spaceborne launch channel M r A satellite-to-ground line-of-sight multiple-input multiple-output demodulation system with multiple non-same-source ground receiving channels and a centralized processing node, such as Figure 3 As shown, the dual-feedback symbol timing recovery method may include the following steps: S1 generates a control word for the numerically controlled oscillator based on the symbol rate, input sampling rate, and timing frequency deviation of the loop feedback. The control word is then used to interpolate and resample the baseband signal of the ground receiving channel to generate a resampled signal.

[0023] exist Figure 3 The single beam configuration of a low-Earth orbit satellite in the satellite-to-ground line-of-sight MIMO demodulation system shown. N b A single transmitting antenna forms a network of multiple ground-feed receiving stations spaced tens of kilometers apart. M r Each receiving channel has its own independent local oscillator and sampling clock. The spaceborne transmitters share the same time-frequency reference, while the ground stations corresponding to each receiving channel use independent local oscillator and sampling clocks. After completing down-conversion and analog-to-digital conversion, each ground station sends the digital baseband signal via optical fiber to a centralized processing node. This node performs symbol timing recovery for each receiving channel in parallel, followed by channel estimation, equalization, and MIMO detection.

[0024] In an embodiment of the present invention, N b The corresponding number of satellite-borne launch channels from the same source. M r The corresponding number of non-same-source ground receiving channels; n Indicates the satellite launch channel index, 1≤ n ≤ N b ; m Indicates the ground receiving channel index, 1≤m≤ M r .

[0025] Let the first m Each ground station at the sampling time t Receive the n The signal propagation delay of each satellite-borne transmission channel is The transmission delay difference relative to the first satellite-borne transmission channel is... Defined as: ; in, For the first m Each ground station at the sampling time t The signal propagation delay of receiving the first satellite-borne transmission channel.

[0026] M r Each non-similar ground receiving channel is formed by ground stations that are separate from each other and do not share a sampling clock. Each ground station transmits the baseband signal to the centralized processing node.

[0027] Affected by the combined effects of receiving crystal oscillator deviation, sampling clock deviation, and changes in satellite-to-ground propagation distance, the normalized timing phase deviation satisfies: ; in, For the first m Normalized oscillator frequency deviation of each ground station This is the sampling rate corresponding to twice the symbol rate.

[0028] No. m The ground receiving channel receives the first n Normalized timing phase deviation of signals from each satellite-borne transmission channel satisfy: ; in, For the first m The normalized timing phase deviation of the signal received by each ground receiving channel from the first satellite-borne transmitting channel. These are the discrete time points after resampling. For mold taking operation, For the first n From the first satellite launch channel to the second... m The transmission delay difference of each ground receiving channel This is the sampling rate corresponding to twice the symbol rate. express right Take the mold.

[0029] from As can be seen from the expression, in addition to the common receiving clock deviation, each space stream in the same ground receiving channel also includes an additional phase formed by the transmission delay difference. This additional phase manifests as inherent inter-channel interference (ICI) in frequency domain timing deviation detection (TOD). Based on the satellite-ground geometry, the transmission delay difference (TDD) of each receiver when receiving two adjacent transmitted signals is approximately equal; therefore, let be... Furthermore, the average phase difference from all transmitting antennas to the m-th receiving end is defined. for: ; in, For any two adjacent transmitted signals to the first m Timing deviation of each receiver end To obtain a complex phase angle, n For the satellite launch channel index, It is the imaginary unit.

[0030] The embodiments of the present invention do not require the separation of each transmit spatial stream beforehand, but instead utilize the statistical characteristics of the joint spectrum of the same receive branch to complete timing recovery.

[0031] Specifically, the reference control word of the numerically controlled oscillator and actual control word They are respectively: ; in, Symbol rate, Input the sampling rate to the receiving end. For the first m The ground receiving channel corresponds to the first... κ The timing frequency deviation of each estimated segment.

[0032] The phase decrement register is updated according to the following formula: ; in, for t The register result of the phase decrement register corresponding to the given time. for The register result of the phase decrement register corresponding to the given time.

[0033] The reference control word of the numerically controlled oscillator is initialized. Before the loop is locked, interpolation resampling is performed according to the reference control word. After the loop obtains the timing frequency deviation, it is superimposed on the reference control word to gradually make the resampled signal... The average sampling rate approaches twice the symbol rate.

[0034] The interpolation resampling process can be accomplished using a fourth-order Lagrange interpolator with a Farrow structure; m One ground receiving channel t Resampled signal at time 1 The expression is as follows: ; in, For the first m The input digital baseband signal of each receiving channel As the interpolation reference position, The first Lagrange interpolation coefficient matrix p Line 1 q Column elements, for t Time of the first The interpolation interval corresponding to each frequency domain data block.

[0035] S2, divide the resampled signal into blocks according to the overlap retention method, and execute the process sequentially for each overlap retention data block. N Point fast Fourier transform, frequency domain matched filtering, and offset-timing phase correction are used to obtain the frequency domain signal of the ground receiving channel, which may include: The resampled signal is divided into blocks according to the overlap retention method to obtain overlap retained data blocks; Convert each overlapping reserved data block intoN The point frequency domain signal is passed through a root-raised cosine matched filter to obtain the joint frequency response of the shaping filter and the matched filter. Offset-timing phase correction is then performed using the frequency domain rotation coefficient generated by the first feedback path to obtain the frequency domain signal for the ground receiving channel.

[0036] The embodiments of the present invention use an overlap-preservation method to resample the signal. Extracting from the middle N Point data block and execute N Point FFT. The overlap length is set according to the matched filter order. After IFFT, samples contaminated by circular convolution are discarded, retaining only the valid interval. The overlap-preserving structure allows matched filtering, timing bias detection, and frequency domain correction to share the same FFT / IFFT processing chain, facilitating high-speed parallel implementation.

[0037] Suppose that the first phase after matched filtering and bias-timing phase correction is... p Each frequency domain data block is Under double oversampling conditions, the positive and negative roll-off regions of the root-raised cosine filter are mirror images of each other with respect to half an FFT period, with the two intervals being... N The spectral component of / 2 contains conjugate rotation information of the symbol timing phase. The first feedback generates frequency domain rotation coefficients, which are then used by the bias-timing phase corrector to cancel the residual timing phase deviation of the current segment in the frequency domain.

[0038] Convert each overlapping reserved data block into N Point frequency domain signal The joint frequency response of the shaping filter and the matched filter is obtained through a root-raised cosine matched filter, and the residual timing phase deviation of the current data block is compensated using the following frequency domain rotation coefficient generated by the first feedback: ; in, For the first m The ground receiving channel corresponds to the first... κ The frequency domain rotation coefficients of the estimated segment. For the first m The ground receiving channel corresponds to the first... κ Smoothed timing phase deviation of each estimated segment.

[0039] After completing the compensation, the first m Frequency domain signals of each ground receiving channel .

[0040] S3, extract the phase gap from the positive and negative roll-off regions of the frequency domain signal. N The paired spectra of two frequency points are used to perform conjugate deweighted accumulation on the paired spectra, and the timing phase deviation estimate of the ground receiving channel is obtained based on the phase angle of the accumulation result.

[0041] Determine the effective roll-off half-width , Within this discrete frequency index range, select respectively i + N / 4 and i +3 N Paired spectrum at position / 4. p The estimated time, the first i Timing phase deviation at each roll-off position Written as: ; in, To obtain a complex phase angle, For the first m The ground receiving channel is in the first p The frequency domain signal at time t.

[0042] Single frequency point estimation is easily affected by noise and the randomness of input symbols, therefore, it is necessary to perform cross-frequency estimation within a Cartesian coordinate system. L At any moment, 2 The effective frequency points are accumulated. The conjugate unweighted estimation method used in this embodiment of the invention yields the first... m The ground receiving channel corresponds to the first... κ The estimated timing phase deviation of each estimated segment The expression is as follows: ; in, L The number of frequency domain data blocks included in a single timing deviation estimate. , For the effective roll-off half-width, The root-raised cosine roll-off factor, For the first m The ground receiving channel is in the first κ In the estimated segment, the first... l Frequency domain signal at time 10:00 It is a complex conjugate.

[0043] Traditional conjugate normalized weighting requires dividing the product of each pair of spectra by its magnitude to give different frequency points the same weight. Considering that the larger the spectral amplitude, the smaller the phase disturbance caused by noise, the expression for the timing phase deviation estimate eliminates magnitude normalization, making strong frequencies contribute more to the accumulation direction and naturally reducing the influence of weak frequencies. This processing, on the one hand, creates an inverse weighted gain oriented towards low signal-to-noise ratios, and on the other hand, eliminates the modulus calculation and division, reducing hardware complexity.

[0044] The above estimates hold true under the following approximate conditions: First, the distance between adjacent satellite-borne transmit antennas is much smaller than the satellite-to-ground transmission distance, and the Doppler frequency deviation when the same ground station receives data from different transmit channels is negligible; second, when the number of FFT points is finite, the effect of frequency offset on frequency response can be treated as the nearest frequency point; third, the transmission delay difference from multiple transmit channels to the same ground station varies slowly within an estimated segment. These conditions are consistent with the actual geometric and processing scale of E-band low-Earth orbit satellite data transmission.

[0045] The conjugate unweighted estimation does not perform modulus normalization on each pair of spectra, so that the spectral components with larger amplitudes naturally receive higher weights in the Cartesian coordinate accumulation, thereby reducing the impact of noise, while eliminating the need for modulus calculation and division operations.

[0046] S4. The timing phase deviation estimate is smoothed by first-order infinite impulse response to obtain the smoothed timing phase deviation.

[0047] In order to avoid significantly increasing the estimated segment length L To suppress high-frequency noise, a first-order infinite impulse response smoothing is performed on each estimated timing phase deviation in the complex domain according to the following formula: ; in, For the first m The ground receiving channel corresponds to the first... κ Smooth timing phase deviation of each estimated segment, Forward coefficients, This is the feedback coefficient. Usually, the sum of the two is set to 1. A smaller one... It can suppress high-frequency noise more effectively, but it will prolong the initial convergence time. The smoothing of timing phase deviation is performed by vector recursion in the complex plane, avoiding the erroneous averaging caused by direct angle filtering when the timing phase crosses the period boundary.

[0048] In steady state and when adjacent estimates are approximately independent, the smoothed estimate variance satisfies: ; in, This is the variance operator.

[0049] The first-order infinite impulse response smoothing recursively applies the unit complex vector corresponding to the timing phase deviation in the complex domain, rather than directly applying linear filtering to the angle value, in order to avoid erroneous transitions in the timing phase at the period boundary.

[0050] S5 divides the smoothed timing phase deviation into two feedback paths, feeding them back to the bias-timing phase correction and the digitally controlled oscillator respectively, to form a dual feedback closed loop, which may include: The smooth timing phase deviation is divided into two feedback paths. The first feedback path is converted into a frequency domain rotation coefficient and fed back to the bias-timing phase correction. The second feedback path is filtered through a loop to extract the timing frequency deviation and fed back to the numerically controlled oscillator, thus forming a dual feedback closed loop that combines fine frequency domain correction with continuous time domain tracking.

[0051] The first feedback is used to compensate for the residual timing phase deviation in the current estimation segment, and the second feedback is used to correct the sampling rate and continuously track the timing frequency deviation; the two feedbacks complement each other in terms of accuracy, convergence speed and dynamic tracking capability.

[0052] Multi-receiver branch dual-feedback symbol timing recovery architecture diagram, as follows: Figure 4 As shown, each receiving branch achieves fine frequency domain correction and continuous time domain tracking through two feedback paths.

[0053] Specifically, the first feedback path converts the smoothed timing phase deviation into a frequency domain rotation coefficient according to the following formula: ; in, For the first m The ground receiving channel corresponds to the first... κ The frequency domain rotation coefficients of the estimated segment.

[0054] The frequency domain rotation coefficients are directly fed back to the bias-timing phase corrector. This branch performs fine-tuned timing phase compensation for each estimated segment in the frequency domain, which mainly determines the steady-state correction accuracy.

[0055] The second feedback loop can employ a first-order proportional-integral loop filter to extract the timing frequency deviation from changes in adjacent estimates; specifically, the first... m Timing frequency deviation of each ground receiving channel: ; in, For the first m The ground receiving channel corresponds to the first... κ The timing frequency deviation of each estimated segment, and These are the proportional branch coefficient and the integral branch coefficient, respectively. If To normalize the loop bandwidth, the proportional branch coefficient and the integral branch coefficient can be selected according to the following formula: .

[0056] The loop filter output is used to correct the numerically controlled oscillator control word, so that the time-domain resampling follows the long-term drift of the received sampling rate.

[0057] In this embodiment of the invention, the resampler for interpolation resampling can be a Lagrange interpolator with a Farrow structure, generally in the form of: ; in, U These are even-numbered tap parameters related to the order of the interpolation filter. i This represents the local frequency index of the roll-off region. for t The interpolation reference position corresponding to the given time. This is the impulse response of the digital interpolation filter. Using the fourth-order Lagrange coefficient matrix, it can be written as: ; This is for rounding down.

[0058] No. m One ground receiving channel t Resampled signal at time 1 The expression is as follows: ; in, For the first m The input digital baseband signal of each receiving channel As the interpolation reference position, The first Lagrange interpolation coefficient matrix p Line 1 q Column elements, for t Time of the first The interpolation interval corresponding to each frequency domain data block.

[0059] The numerically controlled oscillator is driven by a combination of a reference control word and loop feedback: .

[0060] The phase decrement register and interpolation validity flag are updated according to the following formula: ; ; in, for t The interpolation validity flag corresponding to the time step. For symbolic functions, This is an XOR operation.

[0061] When the interpolation validity flag is 1, the interpolation interval is calculated based on the similar triangle relationship: ; in, for t The interpolation interval corresponding to the time.

[0062] When the interpolation validity flag is 0, there is no need to calculate the interpolation interval. Therefore, the second feedback can dynamically increase or decrease the valid output samples, achieving continuous compensation for timing frequency deviations.

[0063] S6 performs an inverse fast Fourier transform on the frequency domain signal that has completed bias-timing phase correction to extract the effective part of the overlapping and retained data block, and obtains the output signal that has completed symbol timing recovery.

[0064] Perform frequency domain matched filtering and bias-timing phase correction on the signal. N Point IFFT removes invalid samples introduced by overlap preservation, resulting in an output that completes symbol timing recovery. Each receiving branch independently executes steps S1 to S6, and the outputs are aligned in the centralized processing node before entering the channel estimation, equalization, and MIMO detection modules.

[0065] The embodiments of the present invention are aimed at M r Each non-similar ground receiving channel is equipped with an independent resampler, timing bias detector, smoothing filter, loop filter, and numerically controlled oscillator, and independently generates timing phase feedback and timing frequency feedback for each channel; the output of each channel after symbol timing recovery is used in the centralized processing node for subsequent channel estimation, channel equalization, and multiple-input multiple-output detection.

[0066] In one alternative implementation, the estimated segment length is selected based on the loop feedback delay and the estimation variance. L Let the number of parallel paths be... N P Feedback latency is T F The slope of the S-curve is detected periodically. K d Then the normalized locking constraint of the loop can be written as: ; Normalized locking constraints indicate that increasing L It can reduce estimation variance and improve noise and ICI immunity, but it will reduce the timing frequency deviation range that the loop can track; reduce L It can improve dynamic tracking capabilities, but it increases residual timing jitter. Therefore, it is possible to select a method based on a combination of transmission delay difference, signal-to-noise ratio, and target timing frequency deviation (TFD) range. L , and loop bandwidth.

[0067] The technical effects of the dual-feedback symbol timing recovery method provided in the embodiments of the present invention will be illustrated below through simulation.

[0068] Simulation conditions: A 4×4 LoS-MIMO model for E-band low-Earth orbit satellites was established, with the onboard transmission array and ground station both employing a uniform linear configuration. The main geometric parameters, communication parameters, and representative instantaneous states are shown in Table 1. This represents the geocentric angle corresponding to the satellite and the link from the first launch channel to the first receiving channel.

[0069] Table 1 Simulation Parameters of the Satellite-Ground LoS-MIMO System

[0070] Timing Deviation Detection Simulation: set up The estimated mean and variance were verified through Monte Carlo simulation. Figure 5 The simulation results show the estimated mean and error of the timing deviation detection algorithm under 4×4 MIMO conditions. Figure 5 The estimated mean and estimation error under 4×4 MIMO are given. The phase of the correlation coefficient caused by the transmission delay difference is mainly manifested as the estimated median shift. Even with ICI, the closed loop can still converge around the new zero point.

[0071] Figure 6 The plot shows the estimated variance of different timing deviation detection methods under 4×4 MIMO conditions. Figure 6 The variances of the conjugate normalized weighted estimation and the conjugate unweighted estimation proposed in this embodiment of the invention were compared. As the transmission delay difference increases and the channel correlation coefficient decreases, the estimation variance increases. Under the same conditions, the conjugate unweighted estimation proposed in this embodiment of the invention reduces the variance by approximately 4 dB in a noise-free environment compared to the traditional method, and further reduces it by 0 dB and... It maintains a stable trend even at a signal-to-noise ratio of 5dB.

[0072] Symbolic timing recovery (STR) loop simulation: Configure 4×4 MIMO and a feedback delay of approximately 40 processing clock cycles. Figure 7 The graph shows the impact of transmission delay difference on the symbol timing recovery (STR) loop's resistance to timing frequency deviation and timing jitter variance. Figure 7 The anti-TFD capability and residual timing jitter variance are presented under different transmission delay differences and estimated segment lengths. L When the value is increased, the estimated average gain improves and the residual jitter decreases, but the dynamic tracking range shrinks; a larger value... L It also allows the loop to remain locked at lower channel correlation coefficients. For example... L When the correlation coefficient is 256 and the magnitude of the correlation coefficient decreases to 0.1, the loop can still resist approximately 7 ppm of TFD and maintain 10 -2 Scale of timing jitter variance.

[0073] The simulation results show that, without relying on pilot signals, the embodiment of the present invention decomposes the LoS-MIMO multi-station non-same-source timing recovery into independent closed loops for each receiving branch, and utilizes frequency domain precise phase compensation and time domain sampling rate compensation to work together; under typical parameters, it can track sampling deviations on the order of thousands of ppm while maintaining low residual timing jitter, making it suitable for E-band ultra-high-speed satellite-to-ground data transmission.

[0074] The dual-feedback symbol timing recovery method for non-homogeneous satellite-to-ground line-of-sight MIMO systems provided in this invention is designed for satellite-to-ground LoS-MIMO scenarios where multiple ground stations cannot share a clock. It sets up an independent closed loop for each receiving branch, eliminating the need for cross-site co-source clock deployment to complete timing calibration before joint reception. This dual-feedback symbol timing recovery method proposes a conjugate unweighted estimation method that, compared to traditional conjugate normalized weighted estimation, eliminates modulus and division operations. Furthermore, it improves estimation gain by retaining strong spectral components under low signal-to-noise ratio conditions, reducing the estimation variance by approximately 4 dB in noise-free scenarios during simulation. Employing a dual-feedback structure of frequency-domain fine correction and time-domain resampling, it can continuously track timing frequency deviations on the order of kiloppm under reasonable estimation segmentation and loop parameters, while reducing residual timing jitter. This dual-feedback symbol timing recovery method utilizes Fast Fourier Transform, parallel frequency point accumulation, Farrow interpolation, and independent receiving branches, making it suitable for parallel implementation on FPGAs or dedicated integrated circuits. It can serve E-band broadband satellite data transmission above 3Gsym / s.

[0075] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any equivalent substitutions or improvements made within the spirit and principles of the present invention to the number of FFT points, estimated segment length, smoothing and loop coefficients, interpolation filter order, number of parallel paths, and implementation platform should be included within the scope of protection of the present invention.

[0076] Secondly, corresponding to the above method embodiments, this invention also provides an electronic device, such as... Figure 8 As shown, it includes a processor 001, a communication interface 002, a memory 003, and a communication bus 004, wherein the processor 001, the communication interface 002, and the memory 003 communicate with each other through the communication bus 004. The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of any of the dual feedback symbol timing recovery methods for non-same-source satellite-to-ground line-of-sight MIMO systems provided in the first aspect of the present invention.

[0077] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0078] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0079] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0080] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0081] The method provided in this invention can be applied to electronic devices. Specifically, the electronic device can be a desktop computer, a portable computer, a smart mobile terminal, a server, etc. No limitation is made herein; any electronic device that can implement this invention falls within the protection scope of this invention.

[0082] Thirdly, corresponding to the dual feedback symbol timing recovery method for non-similar satellite-to-ground line-of-sight MIMO systems provided in the first aspect, this embodiment of the invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps of any of the dual feedback symbol timing recovery methods for non-similar satellite-to-ground line-of-sight MIMO systems provided in the first aspect of this invention.

[0083] For the electronic device / storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and relevant details can be found in the description of the method embodiments.

[0084] It should be noted that the electronic device and storage medium in the embodiments of the present invention are respectively electronic devices and storage media that apply the above-described dual feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems. Therefore, all embodiments of the above-described dual feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems are applicable to the electronic device and storage medium, and can achieve the same or similar beneficial effects.

[0085] It should be noted that, in the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0086] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems, applicable to systems with... N b A single-origin spaceborne launch channel M r A satellite-to-ground line-of-sight multiple-input multiple-output demodulation system with three non-homogeneous ground receiving channels and a centralized processing node, characterized in that... Dual-feedback symbol timing recovery method, including: S1, generate the control word for the numerically controlled oscillator based on the symbol rate, input sampling rate and timing frequency deviation of the loop feedback, and use the control word to interpolate and resample the baseband signal of the ground receiving channel to generate a resampled signal; S2, the resampled signal is divided into blocks according to the overlap-preservation method, and each overlap-preservation data block is executed sequentially. N The frequency domain signal of the ground receiving channel is obtained by performing point fast Fourier transform, frequency domain matched filtering, and offset-timing phase correction. S3, extract the phase gap from the positive and negative roll-off regions of the frequency domain signal. N The pair of spectra at two frequency points are combined, and the pair of spectra are conjugate unweighted accumulation. The timing phase deviation estimate of the ground receiving channel is obtained based on the phase angle of the accumulation result. S4, the estimated timing phase deviation is smoothed by first-order infinite impulse response to obtain a smoothed timing phase deviation; S5, the smoothed timing phase deviation is divided into two feedback paths, which are fed back to the bias-timing phase correction and the numerically controlled oscillator respectively, to form a dual feedback closed loop; S6 performs an inverse fast Fourier transform on the frequency domain signal that has completed bias-timing phase correction to extract the effective part of the overlapping and retained data block, and obtains the output signal that has completed symbol timing recovery.

2. The dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems according to claim 1, characterized in that, M r Each non-similar ground receiving channel is formed by ground stations that are separate from each other and do not share a sampling clock. Each ground station transmits the baseband signal to the centralized processing node. No. m The ground receiving channel receives the first n Normalized timing phase deviation of signals from each satellite-borne transmission channel satisfy: ； in, For the first m The normalized timing phase deviation of the signal received by each ground receiving channel from the first satellite-borne transmitting channel. These are the discrete time points after resampling. For mold taking operation, For the first n From the first satellite launch channel to the second... m The transmission delay difference of each ground receiving channel This is the sampling rate corresponding to twice the symbol rate.

3. The dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems according to claim 1, characterized in that, The interpolation resampling process is completed using a fourth-order Lagrange interpolator with a Farrow structure; m One ground receiving channel t Resampled signal at time 1 The expression is as follows: ； in, For the first m The input digital baseband signal of each receiving channel As the interpolation reference position, The first Lagrange interpolation coefficient matrix p Line 1 q Column elements, for t Time of the first The interpolation interval corresponding to each frequency domain data block.

4. The dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems according to claim 1, characterized in that, S2 includes: The resampled signal is divided into blocks according to the overlap-preservation method to obtain overlap-preserved data blocks; Convert each overlapping reserved data block into N The point frequency domain signal is passed through a root-raised cosine matched filter to obtain the joint frequency response of the shaping filter and the matched filter. Offset-timing phase correction is then performed using the frequency domain rotation coefficient generated by the first feedback path to obtain the frequency domain signal for the ground receiving channel.

5. The dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems according to claim 1, characterized in that, No. m The ground receiving channel corresponds to the first... κ The estimated timing phase deviation of each estimated segment The expression is as follows: ； in, L The number of frequency domain data blocks included in a single timing deviation estimate. , For the effective roll-off half-width, The root-raised cosine roll-off factor, For the first m The ground receiving channel is in the first κ In the estimated segment, the first... l Frequency domain signal at time 10:00 It is a complex conjugate.

6. The dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems according to claim 1, characterized in that, The smoothed timing phase deviation is divided into two feedback paths, which are respectively fed back to the bias-timing phase correction and the numerically controlled oscillator to form a dual feedback closed loop, including: The smooth timing phase deviation is divided into two feedback paths. The first feedback path is converted into a frequency domain rotation coefficient and fed back to the bias-timing phase correction. The second feedback path is filtered through a loop to extract the timing frequency deviation and fed back to the numerically controlled oscillator, thus forming a dual feedback closed loop that combines fine frequency domain correction with continuous time domain tracking.

7. The dual-feedback symbol timing recovery method for non-co-origin satellite-to-ground line-of-sight MIMO systems according to claim 1, characterized in that, The loop filter adopts a proportional-integral structure and updates the first loop according to the following formula: m Timing frequency deviation of each ground receiving channel: ； in, For the first m The ground receiving channel corresponds to the first... κ The timing frequency deviation of each estimated segment, For the first m The ground receiving channel corresponds to the first... The timing frequency deviation of each estimated segment, This is the proportional branch coefficient. For the first m The ground receiving channel corresponds to the first... κ Smooth timing phase deviation of each estimated segment, For the first m The ground receiving channel corresponds to the first... Smooth timing phase deviation of each estimated segment, This is the integral branch coefficient.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of the dual feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the dual-feedback symbol timing recovery method for non-same-source satellite-to-ground line-of-sight MIMO systems as described in any one of claims 1-7.

Images