Low-orbit satellite communication Doppler frequency offset processing method, system, device and medium

Abstract

This invention relates to the field of satellite communication technology and discloses a method, system, device, and medium for processing Doppler frequency offset in low-Earth orbit (LEO) satellite communication. The method includes: separating a pilot signal from a received signal; using the Doppler frequency as a first system state and the received signal as first observation data, performing Doppler frequency offset estimation on the pilot signal using a frequency offset estimation algorithm to obtain an estimated frequency offset value; using the Doppler frequency as a second system state and the estimated frequency offset value as second observation data, performing Doppler frequency offset tracking on the pilot signal using a frequency offset tracking algorithm to obtain a tracked frequency offset value; and performing frequency offset compensation on the received signal based on the tracked frequency offset value to obtain a frequency offset-compensated received signal. This invention uses Bayesian filtering for streaming processing of the Doppler frequency offset during the estimation and tracking stages, effectively reducing the complexity of data storage and computation, increasing the data update rate, and thus effectively improving the accuracy of LEO satellite data.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and in particular to a method, system, device and medium for processing Doppler frequency offset in low-Earth orbit satellite communication. Background Technology

[0002] The DVB-S2 protocol is widely used in low-Earth orbit (LEO) satellite communications. However, the DVB-S2 standard was originally designed for geostationary orbit. When using the DVB-S2 protocol for LEO satellite communications, the Doppler frequency offset generated by the high-speed relative motion between the LEO satellite and the ground receiver directly causes traditional DVB-S2 receivers to be unable to demodulate, severely affecting the quality of satellite-to-ground signal transmission.

[0003] Pilot-based Doppler frequency offset estimation and compensation is the most common processing method in satellite communications. Currently, common frequency offset estimation methods are generally block-based, requiring the reception of individual pilot blocks for separate estimation. This approach demands highly complex real-time computational capabilities and lacks adaptability to the highly dynamic environment of low Earth orbit. Furthermore, most existing compensation methods directly compensate for frequency offset based on the estimation results, leading to insufficient signal processing accuracy. In addition, there are methods using ephemeris-aided Doppler frequency offset processing, but these methods require prior information about the satellite's trajectory, limiting their applicability. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method, system, device, and medium for processing Doppler frequency offset in low-Earth orbit satellite communication. By employing a streaming estimation and tracking scheme for Doppler frequency offset based on Bayesian filtering, the complexity of data storage and computation is effectively reduced, making it more suitable for processing Doppler frequency offset in rapidly time-varying scenarios of low-Earth orbit satellites.

[0005] In a first aspect, the present invention provides a Doppler frequency offset processing method for low-Earth orbit satellite communication, the method comprising:

[0006] Acquire received signals from low-Earth orbit satellites and extract pilot signals from the received signals;

[0007] Using the Doppler frequency as the first system state and the received signal as the first observation data, a frequency offset estimation algorithm based on Bayesian filtering is used to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate value.

[0008] Using the Doppler frequency as the second system state and the frequency offset estimate as the second observation data, a frequency offset tracking algorithm based on Bayesian filtering is used to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value.

[0009] The received signal is frequency offset compensated based on the frequency offset tracking value to obtain the frequency offset compensated received signal.

[0010] Further, the step of using the Doppler frequency as the first system state, the received signal as the first observation data, and employing a frequency offset estimation algorithm based on Bayesian filtering to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate includes:

[0011] The Doppler frequency and the first and second rates of change of the Doppler frequency are taken as the first system state, and the received signal is taken as the first observation data.

[0012] The pilot signal is non-uniformly sampled to obtain multiple pilot blocks, each containing the same number of sampling times. A state transition matrix is ​​established based on the sampling interval between the current sampling time and the previous sampling time.

[0013] Based on the first system state and the state transition matrix, a first state equation is established, and based on the first system state and the first observation data, a first observation equation is established.

[0014] Based on prior knowledge, the first state estimate and the first covariance matrix of the first system state are initialized to obtain the first initial state estimate and the first initial covariance matrix.

[0015] Based on the first initial state estimate and the first initial covariance matrix, Bayesian filtering is used to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate.

[0016] Furthermore, the step of establishing the state transition matrix based on the sampling interval between the current sampling time and the previous sampling time includes:

[0017] Determine whether the current sampling time and the previous sampling time are in the same pilot block. If they are, establish the first state transition matrix based on the sampling interval within the pilot block. Otherwise, establish the second state transition matrix based on the sampling interval between pilot blocks.

[0018] Furthermore, the step of estimating the Doppler frequency offset of the pilot signal using Bayesian filtering includes:

[0019] Determine whether the current sampling time and the previous sampling time are in the same pilot block. If so, perform Doppler frequency offset estimation on the current sampling time based on the first state estimate, the first covariance matrix, and the first state transition matrix of the previous sampling time. Otherwise, perform Doppler frequency offset estimation on the current sampling time based on the first state estimate, the first covariance matrix, and the second state transition matrix of the previous sampling time.

[0020] Further, the step of using the Doppler frequency as the second system state, the frequency offset estimate as the second observation data, and employing a Bayesian filtering-based frequency offset tracking algorithm to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value includes:

[0021] The Doppler frequency and the first and second rates of change of the Doppler frequency are taken as the second system state, and the frequency offset estimate is taken as the second observation data.

[0022] Based on the second system state and the state transition matrix, a second state equation is established, and based on the second system state and the second observation data, a second observation equation is established.

[0023] Based on the frequency offset estimate, the second state estimate of the second system state is initialized to obtain the second initial state estimate, and based on prior knowledge, the second covariance matrix of the second system state is initialized to obtain the second initial covariance matrix.

[0024] Based on the second initial state estimate and the second initial covariance matrix, Bayesian filtering is used to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value.

[0025] Furthermore, the step of using Bayesian filtering to perform Doppler frequency offset tracking on the pilot signal includes:

[0026] Determine whether the current sampling time and the previous sampling time are in the same pilot block. If so, perform Doppler frequency offset tracking on the current sampling time based on the second state estimate, the second covariance matrix, and the first state transition matrix of the previous sampling time. Otherwise, perform Doppler frequency offset tracking on the current sampling point based on the second state estimate, the second covariance matrix, and the second state transition matrix of the previous sampling time.

[0027] Further, the step of performing frequency offset compensation on the received signal based on the frequency offset tracking value to obtain the frequency offset compensated received signal includes:

[0028] A frequency offset value opposite to the frequency offset tracking value is introduced into the received signal to achieve frequency offset compensation of the received signal.

[0029] Secondly, the present invention provides a low-Earth orbit satellite communication Doppler frequency offset processing system, the system comprising:

[0030] The pilot extraction module is used to acquire received signals from low-Earth orbit satellites and separate pilot signals from the received signals.

[0031] The frequency offset estimation module is used to take the Doppler frequency as the first system state, the received signal as the first observation data, and use a frequency offset estimation algorithm based on Bayesian filtering to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate value.

[0032] The frequency offset tracking module is used to take the Doppler frequency as the second system state, the frequency offset estimate as the second observation data, and use a frequency offset tracking algorithm based on Bayesian filtering to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value.

[0033] The signal compensation module is used to perform frequency offset compensation on the received signal according to the frequency offset tracking value to obtain the frequency offset compensated received signal.

[0034] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0035] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0036] This invention provides a method, system, device, and medium for processing Doppler frequency offset in low-Earth orbit satellite communication. The invention proposes a Doppler frequency offset processing framework based on Bayesian filtering, which performs streaming processing on Doppler frequency offset during the estimation and tracking stages, effectively reducing the complexity of data storage and computation, improving the data update rate, adapting to rapidly changing Doppler frequency offset in highly dynamic scenarios, and effectively improving the accuracy of low-Earth orbit satellite data. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating the Doppler frequency offset processing method for low-orbit satellite communication in an embodiment of the present invention.

[0038] Figure 2 This is a schematic diagram of the frame structure of the DVB-S2 protocol in an embodiment of the present invention;

[0039] Figure 3 This is a schematic diagram of the estimation results of the UKF frequency offset estimation algorithm in the frequency offset estimation simulation experiment of the present invention;

[0040] Figure 4 This is a schematic diagram of the estimation results of the EKF frequency offset estimation algorithm in the frequency offset estimation simulation experiment of this invention embodiment;

[0041] Figure 5 This is a schematic diagram of the tracking results of the UKF frequency offset tracking algorithm in the frequency offset tracking simulation experiment of this invention embodiment;

[0042] Figure 6 This is a schematic diagram showing the tracking results of the EKF frequency offset tracking algorithm in the frequency offset tracking simulation experiment of this invention, and a comparison of the results with the UKF frequency offset tracking algorithm.

[0043] Figure 7 This is a schematic diagram of the structure of the low-orbit satellite communication Doppler frequency offset processing system in an embodiment of the present invention;

[0044] Figure 8 This is an internal structural diagram of the computer device in an embodiment of the present invention.

[0045] Figure label:

[0046] 10. Pilot extraction module; 20. Frequency offset estimation module; 30. Frequency offset tracking module; 40. Signal compensation module. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] Please see Figure 1 The first embodiment of the present invention proposes a Doppler frequency offset processing method for low-Earth orbit satellite communication, comprising steps S10 to S40:

[0049] Step S10: Acquire the received signal from the low-orbit satellite and separate the pilot signal from the received signal;

[0050] Step S20: Using the Doppler frequency as the first system state and the received signal as the first observation data, the pilot signal is estimated using a frequency offset estimation algorithm based on Bayesian filtering to obtain the frequency offset estimate value.

[0051] Step S30: Using the Doppler frequency as the second system state and the frequency offset estimate as the second observation data, the pilot signal is tracked by a frequency offset tracking algorithm based on Bayesian filtering to obtain the frequency offset tracking value.

[0052] Step S40: Perform frequency offset compensation on the received signal according to the frequency offset tracking value to obtain the frequency offset compensated received signal.

[0053] This invention provides a Doppler frequency offset processing method for low-Earth orbit (LEO) satellite communication systems. First, a receiver receives the satellite signal from the LEO satellite, using this signal as the received signal. Then, a pilot signal is extracted from the received signal. Doppler frequency offset processing of the satellite signal relies on the pilot signal, which is a known signal transmitted simultaneously with the data signal by the transmitter. Its main function is to provide the receiver with synchronization information such as frequency and phase. After timing and frame synchronization, the receiver can determine the pilot position according to the DVB-S2 protocol, facilitating subsequent Doppler frequency offset processing. It should be noted that conventional reception and extraction methods can be used for satellite signal reception and pilot signal extraction, which will not be elaborated upon here.

[0054] For the extracted pilot signals, the commonly used frequency offset estimation method is block processing. Block processing requires receiving independent pilot blocks and performing frequency offset estimation on each block. To reduce the high computational complexity of block processing in real time, this invention designs a Doppler frequency offset streaming estimation and tracking scheme. By performing streaming processing on each pilot block, the complexity of storage and computation is reduced, thereby improving the data update rate. The detailed steps of frequency offset estimation and frequency offset tracking are described below.

[0055] In the Doppler frequency offset estimation stage, this embodiment, based on the idea of ​​Bayesian filtering, establishes a second-order Doppler frequency offset variation model using the Doppler frequency as the system state and designs a Doppler frequency state estimator using the received signal as the observation. The main difference between this stage and existing technologies is that, in the Bayesian filtering observation model, this embodiment does not use the Doppler frequency as the observation, but rather the received signal. That is, this embodiment applies Bayesian filtering to the signal processing stage, rather than the conventional data processing stage. Furthermore, this observation model is more suitable for the non-uniform pilot structure under the DVB-S2 protocol. Based on this, this embodiment provides a frequency offset estimation algorithm based on the idea of ​​Bayesian filtering to update the frequency offset in real time at each pilot sampling moment. Specific steps include:

[0056] The Doppler frequency and the first and second rates of change of the Doppler frequency are taken as the first system state, and the received signal is taken as the first observation data.

[0057] The pilot signal is non-uniformly sampled to obtain multiple pilot blocks, each containing the same number of sampling times. A state transition matrix is ​​established based on the sampling interval between the current sampling time and the previous sampling time.

[0058] Based on the first system state and the state transition matrix, a first state equation is established, and based on the first system state and the first observation data, a first observation equation is established.

[0059] Based on prior knowledge, the first state estimate and the first covariance matrix of the first system state are initialized to obtain the first initial state estimate and the first initial covariance matrix.

[0060] Based on the first initial state estimate and the first initial covariance matrix, Bayesian filtering is used to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate.

[0061] The frequency offset estimation algorithm in this embodiment adopts the Bayesian filtering concept. The Bayesian filter is a framework for handling uncertain state estimation. Its core is to describe the uncertainty of the state using a probability distribution through a recursive "prediction-update" process, and continuously optimize the estimation based on Bayes' theorem. The process of Bayesian filter handling uncertain state estimation is as follows:

[0062] 1) Initial state probability: The probability of the system at the initial time ( (in state) probability .

[0063] 2) State transition probability: The system in a given state transition state... Moment State Under the condition of transitioning to the state at time k probability .

[0064] 3) Observation probability: The state of the system at time k is known. Under these conditions, the observed data probability .

[0065] 4) Prediction (Prior Estimation): The system is based on the posterior probability at time k-1. Given the state transition probability at time k, predict the prior probability at time k. The formula is as follows:

[0066]

[0067] in Let represent all observed data from time 1 to k-1. When the state transition equation is nonlinear, the above equation has no closed-form solution. In this case, a nonlinear Kalman filter can be used to solve it, such as the Unscented Kalman Filter (UKF) or the Extended Kalman Filter (EKF). The basic idea of ​​UKF is to select a set of Sigma points and perform nonlinear transformation on these points in the state transition to approximate the probability distribution. The basic idea of ​​EKF is to perform a first-order Taylor expansion of the state equation at the current estimated point and then use a linear Kalman filter for processing.

[0068] 5) Update (posterior estimation): The system combines the observations at time k. Based on Bayes' theorem, the prior probability is updated to obtain the posterior probability at time k. The formula is as follows:

[0069]

[0070] in This is a normalization constant to ensure that the sum of posterior probabilities is 1. Similar to the prediction step, when estimating the posterior probability, the original Bayesian filter update formula has no closed-form solution, and it is usually difficult to solve directly when the measurement equation is nonlinear. Therefore, a nonlinear Kalman filter can also be used to solve it, with similar ideas. For example, UKF reconstructs the mean and covariance based on Sigma points, while EKF approximates it based on Taylor expansion.

[0071] This embodiment, based on conventional Bayesian filtering for estimation, designs a streaming processing method for pilot signals, according to... Figure 2 As can be seen from the frame structure of the DVB-S2 protocol, the positions of the pilot signals extracted from the received signal exhibit a non-uniform distribution. Therefore, this embodiment uses a non-uniform sampling method to sample the pilot signals. Let the sampling time be k. Here, m is any integer. This can be understood as obtaining m pilot blocks through non-uniform sampling, with each pilot block containing 36 sampling times. Two adjacent sampling times may or may not be in the same pilot block. This embodiment does not process each pilot block individually, but rather processes the sampling times across all pilot blocks continuously.

[0072] In continuous Bayesian filtering, the system's state equation is established based on state variables and the state transition matrix. This embodiment models the state transition matrix differently depending on whether the two sampling times are in the same pilot block. The specific steps include:

[0073] Determine whether the current sampling time and the previous sampling time are in the same pilot block. If they are, establish the first state transition matrix based on the sampling interval within the pilot block. Otherwise, establish the second state transition matrix based on the sampling interval between pilot blocks.

[0074] In this embodiment, it is assumed that the sampling interval within the pilot block is The number of symbols between the two pilot blocks is If the current sampling time and the previous sampling time are in the same pilot block, then the first state transition matrix F1 is established based on the sampling interval within the pilot block:

[0075]

[0076] If the current sampling time and the previous sampling time are not in the same pilot block, then the current sampling time is the first sampling time in that pilot block, and the previous sampling time is the last sampling time in the previous pilot block. In this case, the second state transition matrix F2 is established based on the time interval between the two pilot blocks.

[0077]

[0078] Since the interval between the two pilot blocks is There are symbols, and during sampling, the sampling interval between two symbols is . Therefore, the time interval between the two pilot blocks is As can be seen, the second state transition matrix mentioned above is established based on the time interval between the two pilot blocks.

[0079] In this embodiment, when performing Doppler frequency offset estimation, the Doppler frequency and its first and second rates of change are taken as the system state, and the received signal is taken as the observation data. State equations and observation equations are established. During sampling Bayesian filtering for frequency offset estimation, the previous sampling time is used to estimate the frequency offset of the current sampling time, regardless of whether the two sampling times are in the same pilot block. This can be understood as follows: if the current sampling time is the first sampling time in the current pilot block, then the previous sampling time is the last sampling time in the previous pilot block. The state transition matrix used in the estimation process needs to be selected based on whether the current sampling time and the previous sampling time are in the same pilot block.

[0080] The following section uses the frequency offset recursive algorithms established by the Unscented Kalman Filter (UKF) and Extended Kalman Filter (EKF), which are based on the Bayesian filtering idea, as examples to illustrate the Doppler frequency offset estimation process in the above embodiments.

[0081] When performing Doppler frequency offset estimation, the Doppler frequency offset is... First-order rate of change of Doppler frequency offset and its second rate of change Using the received signal as the system's state variable and the observed data as the observation data, state equations and observation equations are established. As the state equation, where for The state variable at time t is the first system state; F is the state transition matrix. In the subsequent estimation process, the state transition matrix F is selected as either F1 or F2 based on the interval between the current sampling point and the previous sampling point.

[0082] by Let be the observation equation, where, For state functions, , Sampling time, j H represents the imaginary unit, and H is the measurement matrix. , for The covariance matrix of the Gaussian white noise at time t is: , for Observational data at any given time. Frequency offset estimation stage. Values The received signal at any given moment.

[0083] During the algorithm initialization phase, the initial state estimate of the first system state is set. and the initial covariance matrix The initial state estimate is set as the value obtained from prior knowledge, including the frequency offset and its rate of change, plus the initial frequency shift, i.e.:

[0084]

[0085] In the formula, This represents the initial estimate of the Doppler frequency offset. This represents the initial estimate of the first-order rate of change of the Doppler frequency offset. This represents the initial estimate of the second-order rate of change of Doppler frequency offset.

[0086] The initial covariance matrix is ​​set according to the magnitude of the initial estimation error. It is usually set as a diagonal matrix, and its diagonal elements reflect the uncertainty of the initial estimation of each state variable.

[0087] Based on the above settings, the frequency offset estimation steps based on Unscented Kalman Filter (UKF) and Extended Kalman Filter (EKF) are described below. Since the estimation and tracking of the current sampling time are calculated based on the relevant parameters of the previous sampling time, and the two sampling times may be in the same pilot block or belong to two different pilot blocks, for clarity, the case where the two sampling times are in the same pilot block is denoted as the current time being adjacent to the previous pilot time, and the case where the two sampling times are in different pilot blocks is denoted as the current time not being adjacent to the previous pilot time. Assuming the current time is k, if the current time is adjacent to the previous pilot time, then the previous sampling time is k-1; if not adjacent, then the previous sampling time is k-1. In subsequent embodiments, these two situations will be described separately.

[0088] 1) Doppler frequency offset estimation based on UKF

[0089] The unscented Kalman filter is based on the unscented transform (UT) and adopts the Kalman filtering framework. It is a filter that directly approximates the posterior probability density distribution of the state by obtaining sampling points (usually called Sigma points) through a deterministic sampling strategy.

[0090] When using UKF to update the frequency offset estimate at each pilot sampling time, the prediction update process is as follows:

[0091] A. Calculate the Sigma point:

[0092] If the current time is adjacent to the previous pilot time, then the state estimate from the previous time is used. Covariance Matrix 2 using the Sigma point sampling strategy to obtain Sigma points (n is the dimension of the state vector, preferably) The calculation method is as follows:

[0093]

[0094] If the current time is not adjacent to the previous pilot time, then the state estimate is based on the last sampling time of the previous pilot block. Covariance Matrix 2 using the Sigma point sampling strategy to obtain Sigma points (n is the dimension of the state vector, here) The formula is:

[0095]

[0096] In the formula, This represents the i-th sampling point at time k-1. This represents the initial sampling point at time k-1. Indicates the first The i-th sampling point at time i, Indicates the first The initial sampling point at time 1. Indicates control parameters, , Parameters for controlling the distribution range of Sigma points, for One of the values, preferably, is taken. , This is a minor parameter, usually set to 0. It should be noted that in this embodiment and subsequent embodiments, the formula expressions in different stages of the calculation process may have the same symbol representation and parameter meaning, but their values ​​will differ depending on the processing stage. Therefore, we will not distinguish between their parameter representations here; the specific value can be selected according to the current processing stage.

[0097] B. Predicting the state of the Sigma point: The Sigma point is passed through the state equation to obtain the predicted Sigma point:

[0098] If the current time is adjacent to the previous pilot time, the Sigma points are obtained using the Sigma point sampling strategy. Through state equations Obtain the predicted sigma point :

[0099]

[0100] In the formula, This represents the i-th predicted sampling point at time k in the k-1 time interval.

[0101] If the current time is not adjacent to the previous pilot time, the Sigma points are obtained using the Sigma point sampling strategy. Through state equations Obtain the predicted Sigma point ,Right now:

[0102]

[0103] in, Indicates in The i-th prediction sampling point at time k. The number of symbols representing the interval.

[0104] C. Calculate the state prediction value and covariance prediction value:

[0105] If the current time is adjacent to the previous pilot time, use the predicted Sigma point. Calculate the state prediction value at the current time. Covariance Predictions :

[0106]

[0107] If the current time is not adjacent to the previous pilot time, use the predicted Sigma point. Calculate the state prediction value at the current time. Covariance Predictions :

[0108]

[0109] in, It is a zero-mean Gaussian white noise process. The covariance matrix.

[0110] and These are all weighting coefficients, calculated as follows:

[0111]

[0112] The noise figure is used to introduce prior knowledge about the noise distribution; for a Gaussian distribution, The optimal time.

[0113] Then, the state is updated based on the measurements, as follows:

[0114] A. Calculate the Sigma point:

[0115] If the current time is adjacent to the previous pilot time, use the state prediction value of the current time. Covariance Predictions 2 using the Sigma point sampling strategy to obtain Sigma points (n is the dimension of the state vector, here) ):

[0116]

[0117] If the current time is not adjacent to the previous pilot time, use the state prediction value of the current time. Covariance Predictions 2 using the Sigma point sampling strategy to obtain Sigma points (n is the dimension of the state vector, here) ):

[0118]

[0119] In the formula, This represents the initial sampling point at time k, at time k-1. This represents the i-th sampling point at time k-1. Indicates in At this moment k The initial sampling point at time 1. Indicates in k-l At this moment k The i-th sampling point at time t.

[0120] B. Calculate and measure the Sigma point:

[0121] If the current time is adjacent to the previous pilot time, the 2 obtained using the Sigma point sampling strategy Sigma points The measurement Sigma point at the current moment is obtained by passing the observation equation. :

[0122]

[0123] If the current time is not adjacent to the previous pilot time, the 2 obtained using the Sigma point sampling strategy Sigma points The measurement Sigma point at the current moment is obtained by passing the observation equation. :

[0124]

[0125] In the formula, This represents the i-th measurement sampling point at time k, at time k-1. Indicates in At this moment k The i-th measurement sampling point at time i.

[0126] C. Calculate the measured predicted values ​​and covariance:

[0127] If the current time is adjacent to the previous pilot time, this algorithm uses the measurement Sigma point at the current time. Calculate the predicted measurement value at the current moment. Covariance matrix covariance The cross-covariance matrix of the state vector and the observation vector is represented by . .

[0128]

[0129]

[0130]

[0131] If the current time is not adjacent to the previous pilot time, use the measurement Sigma point at the current time. Calculate the predicted measurement value at the current moment. Observation covariance matrix The cross-covariance matrix of the state vector and the observation vector is represented by . .

[0132]

[0133]

[0134]

[0135] In the formula, This represents the predicted value of the measurement at time k, given that the value is at time k-1. Indicates in The predicted value of the measurement at time k is given by time k. Let k represent the observation covariance matrix at time k. This represents the cross-covariance matrix.

[0136] D. Calculate the Kalman gain:

[0137] The Kalman gain is calculated using the following formula:

[0138]

[0139] In the formula, This represents the Kalman gain at time k.

[0140] E. Update the state estimate and covariance:

[0141] If the current time is adjacent to the previous pilot time, the state of the current time is calculated according to the following formula. Covariance Matrix .

[0142]

[0143]

[0144] If the current time is not adjacent to the previous pilot time, the state estimate for the current time is calculated according to the following formula. Covariance Matrix .

[0145]

[0146]

[0147] At this time and In step A of the prediction update, it will only be used for the calculation of the Sigma point at the next time step.

[0148] Through the above-described time update and measurement update process, the UKF implementation in this embodiment can perform non-uniform estimation of Doppler frequency offset and its first-order and second-order rates of change, thereby obtaining more accurate estimates, effectively suppressing the effects of noise and multipath fading, and improving the accuracy and stability of frequency offset estimation.

[0149] 2) Doppler frequency offset estimation based on EKF

[0150] When EKF is used to update the frequency offset estimate at each pilot sampling time, the prediction process is as follows:

[0151] A. State Prediction:

[0152] If the current time is adjacent to the previous pilot time, then the state estimate is based on the previous pilot time. The state estimate for the current moment is obtained through the state equation. Right now:

[0153]

[0154] If the current time is not adjacent to the previous pilot time, the state estimate is based on the last pilot time of the previous pilot block. The state estimate for the current moment is obtained through the state equation. Right now:

[0155]

[0156] B. Covariance prediction:

[0157] If the current time is adjacent to the previous pilot time, use the covariance matrix of the previous pilot time. Calculate the prediction covariance matrix at the current time. The calculation method is as follows:

[0158]

[0159] If the current time is not adjacent to the previous pilot time, use the covariance matrix of the last pilot time of the previous pilot block. Calculate the prediction covariance matrix at the current time. The calculation method is as follows:

[0160]

[0161] in, It is the covariance matrix of the zero-mean Gaussian white noise process at time k-1. yes The covariance matrix of the zero-mean Gaussian white noise process at time t. , It is a zero-mean Gaussian white noise process.

[0162] Then, the state is updated based on the prediction process, as follows:

[0163] A. Calculate observation and prediction:

[0164] If the current time is adjacent to the previous pilot time, use the predicted state estimate at the current time. As a function The independent variable is used to calculate the observation prediction matrix at the current time through the observation equation. The calculation method is as follows:

[0165]

[0166] If the current time is not adjacent to the previous pilot time, use the predicted state estimate at the current time. As a function The independent variable is used to calculate the observation prediction matrix at the current time through the observation equation. The calculation method is as follows:

[0167]

[0168] B. Calculate the observation Jacobian matrix:

[0169] If the current time is adjacent to the previous pilot time, the observation prediction matrix for the current time is calculated using the following formula. Perform the derivative operation to calculate the observation Jacobian matrix at the current time. .

[0170]

[0171] If the current time is not adjacent to the previous pilot time, the observation prediction matrix for the current time is calculated using the following formula. Perform the derivative operation to calculate the observation Jacobian matrix at the current time. .

[0172]

[0173] C. Calculate the Kalman gain:

[0174] If the current time is adjacent to the previous pilot time, the prediction covariance matrix of the current time is used. And the observed Jacobian matrix at the current moment The Kalman gain at the current moment is calculated using the following formula. .

[0175]

[0176] If the current time is not adjacent to the previous pilot time, the prediction covariance matrix of the current time is used. And the observed Jacobian matrix at the current moment The Kalman gain at the current moment is calculated using the following formula. .

[0177]

[0178] D. State update and covariance update:

[0179] If the current time is adjacent to the previous pilot time, calculate the state of the current time using the following formula. Covariance Matrix .

[0180]

[0181]

[0182] If the current time is not adjacent to the previous pilot time, calculate the state of the current time using the following formula. Covariance Matrix .

[0183]

[0184]

[0185] Where I represents the identity matrix.

[0186] At this time and It is only used for state prediction and covariance prediction in the next adjacent pilot time step during the prediction process.

[0187] Through the above prediction and update process, the EKF embodiment in this example can estimate the Doppler frequency offset and its first-order and second-order rates of change, obtain accurate estimates, suppress the effects of noise and multipath fading, and improve the accuracy and stability of frequency offset estimation.

[0188] To verify the effectiveness, the UKF and EKF frequency offset estimation algorithms are simulated and validated using the MATLAB platform. This experiment focuses on signal-to-noise ratio... Under the given conditions, Doppler frequency offset estimation is performed on the signal extracted by the pilot extraction module. A non-uniform sampling method is adopted, and the sampling time k is set to... ,in .

[0189] The simulation results of the two experiments are as follows: Figure 3 and Figure 4 As shown, from Figure 3 and Figure 4 As can be seen, both the UKF frequency offset estimation algorithm and the EKF frequency offset estimation algorithm have high estimation accuracy, good performance in real-time estimation of Doppler frequency offset, and can perform estimation over a large time span.

[0190] In the frequency offset tracking stage, the pilot signal is frequency offset tracked based on the frequency offset estimate obtained in the frequency offset estimation stage, thereby obtaining a more accurate frequency offset tracking value. During frequency offset tracking, the Doppler frequency is still used as the system state, and a second-order Doppler frequency offset variation model is established, but the frequency offset estimate obtained in the frequency offset estimation stage is used as the observation data. In the frequency offset tracking stage, the Bayesian filtering idea is still employed, and a non-uniform sampling method is used to track the pilot signal extracted by the pilot extraction module. Specific steps include:

[0191] The Doppler frequency and the first and second rates of change of the Doppler frequency are taken as the second system state, and the frequency offset estimate is taken as the second observation data.

[0192] Based on the second system state and the state transition matrix, a second state equation is established, and based on the second system state and the second observation data, a second observation equation is established.

[0193] Based on the frequency offset estimate, the second state estimate of the second system state is initialized to obtain the second initial state estimate, and based on prior knowledge, the second covariance matrix of the second system state is initialized to obtain the second initial covariance matrix.

[0194] Based on the second initial state estimate and the second initial covariance matrix, Bayesian filtering is used to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value.

[0195] The frequency offset tracking process in this embodiment is similar to the frequency offset estimation process described above. Both are based on the Bayesian filtering algorithm. In the frequency offset tracking algorithm, the system state is still based on the Doppler frequency, first-order rate of change, and second-order rate of change. Different state transition matrices F are modeled depending on whether the sampling times of the pilot signals are in the same pilot block. That is, F=F1 when they are in the same pilot block, and F=F2 otherwise. There are two main differences between frequency offset tracking and the frequency offset estimation described above. One difference is that the observation data used in frequency offset tracking is different from that used in frequency offset estimation. The other difference is that the expression of the observation equation established in frequency offset tracking is different from that in frequency offset estimation.

[0196] Specifically, in the frequency offset tracking phase, the frequency offset estimate obtained in the frequency offset estimation phase is used as the observation data. Based on the system state and the observation data, state equations and observation equations are established. The state equation is:

[0197]

[0198] In the formula, for The state variable at time t, F G is the state transition matrix, and G is the process noise driving matrix. For a zero-mean Gaussian white noise process, its covariance matrix is: .

[0199] The observation equation is:

[0200]

[0201] In the formula, Let H be the observation data at time k, and H be the measurement matrix. Since the observation data during the tracking phase is the frequency offset estimate output from the estimation phase, it includes the Doppler frequency offset and its first and second rate of change information. , The covariance matrix of zero-mean Gaussian white noise is: The observation data here are frequency offset estimates obtained during the frequency offset estimation stage.

[0202] During the algorithm initialization phase, the initial state estimate of the system state is taken from the initial frequency offset estimate and the initial estimate of its rate of change output by the Doppler frequency offset estimate. The initial covariance matrix is ​​set according to the magnitude of the initial estimation error, and can usually be set as a diagonal matrix, whose diagonal elements reflect the uncertainty of the initial estimate of each state variable.

[0203] As can be seen, the state equation in the tracking phase has an additional noise term compared to the state equation in the estimation phase. The difference between the observation equation in the tracking phase and the observation equation in the estimation phase lies in the different values ​​of the observation data.

[0204] When updating the frequency offset tracking value using a frequency offset tracking algorithm based on Bayesian principles, the appropriate state transition matrix is ​​selected based on whether the current time and the previous pilot time are adjacent. Data is then updated according to the aforementioned state equation and observation equation to obtain a more accurate frequency offset tracking value. When using a nonlinear Kalman filter (such as UKF or EKF) as the frequency offset tracking algorithm, the specific tracking process can be referenced in the two embodiments of the frequency offset estimation stage. The difference lies in the different values ​​of the observed data and the different expressions of the state equations; the specific calculation process is the same and will not be repeated here.

[0205] This embodiment uses a frequency offset tracking algorithm based on nonlinear Kalman filtering to track the frequency offset estimate in real time under the condition that the pilot position is non-uniformly distributed. It continuously optimizes the state estimation, thereby obtaining a more accurate frequency offset tracking value, effectively suppressing the influence of noise and multipath fading, and improving the accuracy and stability of frequency offset estimation.

[0206] To verify the tracking performance of frequency offset tracking, the UKF and EKF frequency offset tracking algorithms were simulated and verified using the MATLAB platform. The simulations used the Doppler frequency offset output during the Doppler frequency estimation stage. First-order rate of change Second-order rate of change As the initial value for tracking, the algorithm is initialized, with the signal-to-noise ratio... Under the given conditions, Doppler frequency offset tracking is performed on the extracted pilot signal. The tracking results of the UKF frequency offset tracking algorithm are as follows: Figure 5 As shown, the tracking results of EKF and the comparison results of tracking errors with UKF are as follows. Figure 6 As shown.

[0207] from Figure 5 It can be seen that when the signal-to-noise ratio (SNR) is 10 dB, the tracking process enters a steady state at 0.0003 s, and the maximum errors in frequency offset and first- and second-order rates of change thereafter are approximately and This demonstrates that the UKF tracking algorithm can track the target signal extremely quickly throughout the carrier tracking process, making it highly suitable for systems with second-order or higher rate of change in frequency offset. It effectively balances tracking accuracy and stability when tracking signals. Low-Earth orbit satellites operate in highly dynamic and complex environments, and their Doppler frequency offset often exhibits second-order or higher rate of change. Figure 6It can be seen that the EKF tracking algorithm and the UKF tracking algorithm have similar tracking capabilities for Doppler frequency offset. However, the EKF tracking algorithm is inferior to the UKF tracking algorithm in tracking the first and second rates of change of Doppler frequency offset, resulting in poorer tracking performance. Simulation results indicate that either the EKF or UKF estimation algorithm is preferred in the frequency offset estimation stage, while the UKF tracking algorithm is preferred in the frequency offset tracking stage, as these methods yield better estimation and tracking results.

[0208] After obtaining the frequency offset tracking value through the above steps, frequency offset compensation can be performed on the received signal based on the frequency offset tracking value. In this embodiment, the basic idea of ​​frequency offset compensation is to introduce a frequency offset opposite to the Doppler frequency offset tracking value into the received signal to cancel the frequency offset in the received signal. Specifically, a time-domain compensation method is adopted, which multiplies the signal by a complex exponential signal generated based on the frequency offset tracking value in the time domain to achieve frequency offset compensation, so that the carrier frequency of the signal is restored to the frequency of the local oscillator. The time-domain compensation method provided in this embodiment has the advantages of low computational complexity and good real-time performance, and can achieve accurate and efficient frequency offset compensation.

[0209] This embodiment provides a Doppler frequency offset processing method for low-Earth orbit satellite communication. It proposes a Doppler frequency offset processing framework based on the Bayesian filtering concept. Bayesian filtering is used for processing in both the estimation and tracking stages, which effectively reduces the complexity of data storage and computation, improves the update rate, can adapt to the rapidly time-varying Doppler frequency offset in highly dynamic scenarios, and improves the accuracy of low-Earth orbit satellite data.

[0210] Please see Figure 7 Based on the same inventive concept, the second embodiment of this invention proposes a low-orbit satellite communication Doppler frequency offset processing system, comprising:

[0211] Pilot extraction module 10 is used to acquire received signals from low-orbit satellites and separate pilot signals from the received signals;

[0212] The frequency offset estimation module 20 is used to take the Doppler frequency as the first system state, the received signal as the first observation data, and use a frequency offset estimation algorithm based on Bayesian filtering to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimation value.

[0213] The frequency offset tracking module 30 is used to take the Doppler frequency as the second system state, take the frequency offset estimate as the second observation data, and use a frequency offset tracking algorithm based on Bayesian filtering to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value.

[0214] The signal compensation module 40 is used to perform frequency offset compensation on the received signal according to the frequency offset tracking value to obtain the frequency offset compensated received signal.

[0215] The technical features and effects of the low-Earth orbit satellite communication Doppler frequency offset processing system proposed in this embodiment are the same as those of the method proposed in this embodiment, and will not be repeated here. Each module in the above-mentioned low-Earth orbit satellite communication Doppler frequency offset processing system can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in a computer device in hardware form, or it can be stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0216] Furthermore, embodiments of the present invention also propose a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0217] Please see Figure 8 The diagram illustrates the internal structure of a computer device in one embodiment. This computer device can specifically be a terminal or a server. The computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a Doppler frequency offset processing method for low-Earth orbit satellite communication. The display screen of the computer device can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0218] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0219] Furthermore, embodiments of the present invention also propose a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0220] In summary, this invention proposes a method, system, device, and medium for processing Doppler frequency offset in low-Earth orbit (LEO) satellite communication. The method involves acquiring a received signal from a LEO satellite and separating a pilot signal from it. Using the Doppler frequency as a first system state and the received signal as first observation data, a Bayesian filtering-based frequency offset estimation algorithm is used to estimate the Doppler frequency offset of the pilot signal, yielding an estimated frequency offset value. Using the Doppler frequency as a second system state and the estimated frequency offset value as second observation data, a Bayesian filtering-based frequency offset tracking algorithm is used to track the Doppler frequency offset of the pilot signal, yielding a tracked frequency offset value. The received signal is then compensated for the frequency offset based on the tracked value, resulting in a frequency offset-compensated received signal. This invention proposes a Doppler frequency offset processing framework based on Bayesian filtering, performing streaming processing of the Doppler frequency offset during the estimation and tracking stages. This effectively reduces the complexity of data storage and computation, improves the data update rate, and can adapt to rapidly changing Doppler frequency offsets in highly dynamic scenarios, effectively improving the accuracy of LEO satellite data.

[0221] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0222] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A method for processing Doppler frequency offset in low-Earth orbit satellite communication, characterized in that, include: Acquire received signals from low-Earth orbit satellites and extract pilot signals from the received signals; Using the Doppler frequency as the first system state and the received signal as the first observation data, a frequency offset estimation algorithm based on Bayesian filtering is used to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate value. Using the Doppler frequency as the second system state and the frequency offset estimate as the second observation data, a frequency offset tracking algorithm based on Bayesian filtering is used to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value. The received signal is frequency offset compensated according to the frequency offset tracking value to obtain the frequency offset compensated received signal. The step of using the Doppler frequency as the first system state, the received signal as the first observation data, and employing a Bayesian filtering-based frequency offset estimation algorithm to estimate the Doppler frequency offset of the pilot signal to obtain the frequency offset estimate includes: The Doppler frequency and the first and second rates of change of the Doppler frequency are taken as the first system state, and the received signal is taken as the first observation data. The pilot signal is non-uniformly sampled to obtain multiple pilot blocks, each containing the same number of sampling times. A state transition matrix is ​​established based on the sampling interval between the current sampling time and the previous sampling time. Based on the first system state and the state transition matrix, a first state equation is established, and based on the first system state and the first observation data, a first observation equation is established. Based on prior knowledge, the first state estimate and the first covariance matrix of the first system state are initialized to obtain the first initial state estimate and the first initial covariance matrix. Based on the first initial state estimate and the first initial covariance matrix, Bayesian filtering is used to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate. The steps of using the Doppler frequency as the second system state, the frequency offset estimate as the second observation data, and employing a Bayesian filtering-based frequency offset tracking algorithm to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value include: The Doppler frequency and the first and second rates of change of the Doppler frequency are taken as the second system state, and the frequency offset estimate is taken as the second observation data. Based on the second system state and the state transition matrix, a second state equation is established, and based on the second system state and the second observation data, a second observation equation is established. Based on the frequency offset estimate, the second state estimate of the second system state is initialized to obtain the second initial state estimate, and based on prior knowledge, the second covariance matrix of the second system state is initialized to obtain the second initial covariance matrix. Based on the second initial state estimate and the second initial covariance matrix, Bayesian filtering is used to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value.

2. The low-Earth orbit satellite communication Doppler frequency offset processing method according to claim 1, characterized in that, The step of establishing the state transition matrix based on the sampling interval between the current sampling time and the previous sampling time includes: Determine whether the current sampling time and the previous sampling time are in the same pilot block. If they are, establish the first state transition matrix based on the sampling interval within the pilot block. Otherwise, establish the second state transition matrix based on the sampling interval between pilot blocks.

3. The low-Earth orbit satellite communication Doppler frequency offset processing method according to claim 2, characterized in that, The step of estimating the Doppler frequency offset of the pilot signal using Bayesian filtering includes: Determine whether the current sampling time and the previous sampling time are in the same pilot block. If so, perform Doppler frequency offset estimation on the current sampling time based on the first state estimate, the first covariance matrix, and the first state transition matrix of the previous sampling time. Otherwise, perform Doppler frequency offset estimation on the current sampling time based on the first state estimate, the first covariance matrix, and the second state transition matrix of the previous sampling time.

4. The low-Earth orbit satellite communication Doppler frequency offset processing method according to claim 2, characterized in that, The step of performing Doppler frequency offset tracking on the pilot signal using Bayesian filtering includes: Determine whether the current sampling time and the previous sampling time are in the same pilot block. If so, perform Doppler frequency offset tracking on the current sampling time based on the second state estimate, the second covariance matrix, and the first state transition matrix of the previous sampling time. Otherwise, perform Doppler frequency offset tracking on the current sampling point based on the second state estimate, the second covariance matrix, and the second state transition matrix of the previous sampling time.

5. The low-Earth orbit satellite communication Doppler frequency offset processing method according to claim 1, characterized in that, The step of performing frequency offset compensation on the received signal based on the frequency offset tracking value to obtain the frequency offset compensated received signal includes: A frequency offset value opposite to the frequency offset tracking value is introduced into the received signal to achieve frequency offset compensation of the received signal.

6. A low-Earth orbit satellite communication Doppler frequency offset processing system, characterized in that, The system is applied to the method as described in any one of claims 1 to 5, comprising: The pilot extraction module is used to acquire received signals from low-Earth orbit satellites and separate pilot signals from the received signals. The frequency offset estimation module is used to take the Doppler frequency as the first system state, the received signal as the first observation data, and use a frequency offset estimation algorithm based on Bayesian filtering to perform Doppler frequency offset estimation on the pilot signal to obtain the frequency offset estimate value. The frequency offset tracking module is used to take the Doppler frequency as the second system state, the frequency offset estimate as the second observation data, and use a frequency offset tracking algorithm based on Bayesian filtering to perform Doppler frequency offset tracking on the pilot signal to obtain the frequency offset tracking value. The signal compensation module is used to perform frequency offset compensation on the received signal according to the frequency offset tracking value to obtain the frequency offset compensated received signal.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images