A phased array terminal dynamic beam stabilizing method based on vehicle motion prediction

Abstract

This invention relates to a dynamic beam stabilization method for phased array terminals based on vehicle motion prediction, comprising the following steps: real-time synchronous acquisition of vehicle IMU, GNSS, and ephemeris data; construction of a rigid body dynamics model of vehicle motion and a satellite orbital dynamics model; using preprocessed IMU and GNSS data as observations, combined with the rigid body dynamics model of vehicle motion, employing extended Kalman filtering to calculate the vehicle's attitude and position at future moments, and combining ephemeris data and the satellite orbital dynamics model to predict the azimuth and elevation angles of the satellite relative to the array surface, thereby calculating the feedforward compensation predicted phase for each array element and sending it to the array controller, which is triggered at regular intervals based on a PTP synchronization clock; real-time calculation of the prediction error and comparison with a threshold to achieve smooth switching between the prediction feedforward compensation mode and the traditional real-time compensation mode. Compared with existing technologies, this invention has advantages such as improved beam stabilization accuracy under all operating conditions.

Description

Technical Field

[0001] This invention relates to the field of vehicle-mounted phased array control, and in particular to a dynamic beam stabilization method for a phased array terminal based on vehicle motion prediction. Background Technology

[0002] With the rapid development of low-Earth orbit broadband satellite communication, vehicle-to-everything (V2X) communication, autonomous driving, and airborne broadband systems, the communication demands between vehicles and satellites are growing exponentially. To meet the requirements for anti-fading, anti-interference, and anti-dynamic disturbance capabilities under high-speed mobility conditions, vehicle-mounted phased array antennas are gradually becoming the mainstream solution. Phased array antennas achieve electronic scanning capabilities by applying different phase shifts to multiple array elements, enabling satellite tracking without stopping the vehicle or rotating the antenna. However, the environment in which vehicle-mounted terminals operate is complex, and the dynamic disturbances from the vehicle itself are far more severe than those from traditional fixed satellite ground stations. Typical vehicle operation may experience [various disturbances].

[0003] Current beam control primarily relies on attitude compensation from an IMU (Inertial Measurement Unit). However, IMU measurements are subject to several factors: vehicle pitch vibrations at high speeds; longitudinal and lateral angle changes caused by road undulations; dramatic roll angle changes due to lane changes and sharp turns; vehicle tilting during acceleration and braking; and high-frequency attitude disturbances caused by irregular bumps. These dynamic attitudes directly affect the antenna base, causing the antenna's relative pointing angle to constantly change. Phased array antennas inherently require the array steering vector to be precisely aligned with the satellite direction; otherwise, the following problems will occur: 1. Beam main lobe shift causes a decrease in satellite receiving gain; 2. The decline in link quality caused by increased sidelobes and pointing error; 3. Loss of link lock with satellite necessitates reacquiring the signal, which can take several seconds to tens of seconds; 4. Frequent lockouts lead to business interruptions and large fluctuations in throughput.

[0004] Therefore, how to achieve high dynamic, high precision, and low latency phased array pointing stability under conditions of intense vehicle movement is one of the core issues in the field of vehicle-mounted satellite communication.

[0005] Current vehicle-mounted antennas generally rely on IMUs to provide heading, pitch, and roll angles for attitude compensation. However, IMUs themselves have the following problems: attitude lag due to limited sampling frequency; measurement jitter due to acceleration / angular velocity noise; zero-bias instability due to temperature drift; nonlinear response under high-frequency vibration; and filtering delay during the position fusion process with GNSS (Global Navigation Satellite System).

[0006] When a vehicle undergoes rapid and dynamic changes, relying solely on the current vehicle attitude provided by the IMU for array phase compensation cannot meet real-time requirements. Furthermore, phased array antennas inherently have delays: the time required for the control system to calculate the pointing angle; the link for uploading digital beamforming parameters to the array controller; and the response time of the phase shifter / phase shifting network from receiving a command to actual application. These delays typically accumulate to tens to hundreds of milliseconds, while vehicle attitude can change significantly within 100 ms, resulting in compensation lag.

[0007] Chinese patent CN114927884A discloses a dynamic compensation method for improving the performance of vehicle-mounted phased array antennas. It mainly relies on real-time feedback data for dynamic compensation, which may lead to compensation lag and beam instability because real-time processing cannot predict future motion changes of vehicles and satellites, thus introducing errors in fast dynamic environments.

[0008] Therefore, in the existing technology system, there is no mature solution that can truly solve the core requirement of dynamic beam stabilization of vehicle phased arrays, and it is difficult to avoid short-term loss of lock in complex scenarios such as fast turning and bumpy rides. Summary of the Invention

[0009] The purpose of this invention is to provide a dynamic beam stabilization method for phased array terminals based on vehicle motion prediction. By constructing a vehicle motion dynamics model and combining IMU / GNSS multi-source observation information, the future vehicle attitude and position are predicted by EKF (Extended Kalman Filter). Furthermore, satellite orbital elements are combined to predict the relative motion between the vehicle and the satellite. Finally, based on the prediction results, feedforward compensation of array element phase is performed to achieve dynamically stable beam pointing control.

[0010] The objective of this invention can be achieved through the following technical solutions: A dynamic beam stabilization method for a phased array terminal based on vehicle motion prediction includes the following steps: S1: Real-time synchronous acquisition of vehicle IMU data, GNSS data, and ephemeris data, and preprocessing them; S2: Construct a rigid body dynamics model of vehicle motion to predict the future attitude and position of the vehicle, and construct a satellite orbital dynamics model to predict the future position of the satellite, providing a basis for the joint prediction of the relative motion between the satellite and the ground; S3: Using preprocessed IMU and GNSS data as observations, combined with the vehicle motion rigid body dynamics model, extended Kalman filtering is used to calculate the vehicle's attitude and position at future moments, and combined with ephemeris data and satellite orbit dynamics model, the azimuth and elevation angles of the satellite relative to the array are predicted. S4: Based on the predicted azimuth and elevation angles of the satellite relative to the array surface, combined with the element coordinates and wavenumber of the phased array antenna, the feedforward compensation prediction phase of each element is quickly calculated. S5: The calculated feedforward compensation prediction phase is sent to the array controller of the phased array antenna. The array controller stores the phase parameters in a dedicated buffer and triggers the timing based on the PTP synchronization clock. S6: Calculate the prediction error in real time and compare it with the threshold. Based on the comparison result, smoothly switch between prediction feedforward compensation mode and traditional real-time compensation mode to ensure beam stability. When prediction feedforward compensation mode is selected, continue to execute S1~S5. When traditional real-time compensation mode is selected, perform phase compensation based on the IMU data obtained in real time in S1.

[0011] The preprocessing includes filtering, outlier removal, and coordinate transformation.

[0012] In S2, a state vector-based system is constructed. The state transition model describes the time evolution of the vehicle's motion state, resulting in the rigid body dynamics model of the vehicle's motion: ; in, State vector at time step , For three-axis angles, For triaxial angular velocity, These are the velocities of the three axes; The state transition function for rigid body dynamics is as follows: ; in The IMU sampling period The angular acceleration is a three-axis acceleration, obtained from the difference in the angular velocities of the IMU; The acceleration is a three-axis acceleration, obtained from the difference of GNSS velocities; Let be the system process noise, with a mean of 0 and a covariance of . Gaussian distribution, covariance Tuning based on the measurement accuracy of the IMU and GNSS.

[0013] In step S2, based on the six orbital elements provided by satellite ephemeris data, a two-body orbital model is used to construct a satellite orbital dynamics model to calculate the satellite's future position. Considering the orbital perturbations of low-Earth orbit satellites, an orbital perturbation correction term is introduced, and the solution is obtained. The coordinates of the satellite in the geocentric coordinate system : ; in, The six elements of a satellite orbit; This is the position calculation function for the two-body orbit model; This is the orbital perturbation correction term, obtained from the correction parameters of the satellite ephemeris.

[0014] The S3 is divided into three stages: initialization, prediction, and update. Specifically, it includes the following steps: Initialization: Take the multi-source preprocessed data at the initial moment as the initial value and set the initial state vector. ,in, , , These are the initial attitude, angular velocity, and linear velocity; the initial error covariance matrix is ​​set. This is a diagonal matrix, with diagonal elements representing the measurement error variances of each state variable; the process noise covariance matrix is ​​set. and observation noise covariance matrix ; Prediction phase: based on Optimal state estimate at time t And vehicle motion rigid body dynamics model, predict Prior state vector at time step and prior error covariance matrix : ; ; in, State transition function The Jacobian matrix is ​​obtained by taking the partial derivative of the rigid body dynamics model, from the prior state vector. Mid-calculation of vehicle attitude prediction and predicted location ; Update phase: with The IMU and GNSS preprocessed data at time 1 are the observations. Construct observation equations Calculate Kalman gain And update the prior state and covariance to obtain Optimal state estimate at time t and the optimal error covariance matrix : ; ; ; in, For the observation matrix, For state vectors, To observe the noise vector, It is the identity matrix; The satellite-to-ground relative azimuth calculation is based on the predicted vehicle position. and satellite positions predicted based on satellite orbital dynamics models Calculate the position vector of the satellite relative to the vehicle array. ; this position vector Transform from the Earth-centered, Earth-fixed coordinate system to the local coordinate system of the front surface, and calculate the corresponding azimuth angle. and pitch angle The formula is: ; ; in, , , They are respectively Three-axis components in the local coordinate system of the array surface This indicates the magnitude of the position vector.

[0015] Specifically, S4 is: Based on the predicted azimuth angle of the satellite relative to the array Pitch angle Combined with the element coordinates of the phased array antenna and wave number Fast calculation of feedforward compensation prediction phase for each array element : ; in, , , The element coordinates of the phased array antenna are respectively The three-axis components; This is the reference phase.

[0016] In step S5, the array controller, based on the PTP clock, The feedforward compensation prediction phase from the buffer is precisely loaded into the phase shifters of each array element at every moment, with the loading completion time perfectly aligned with the actual attitude change time of the vehicle. The feedforward compensation prediction phase is based on... The vehicle's attitude and position at any given time, as well as the relative orientation between the satellite and the ground, are calculated.

[0017] Specifically, S6 is: Calculate prediction error: Data collection Real-world attitude measurement of the IMU at any given moment , and predicted posture Compare and calculate the prediction error. Take the maximum value As a basis for error judgment; when When the condition is determined to be a normal dynamic operating condition, the predictive feedforward compensation mode is maintained, and S1~S5 are continued. The preset threshold; when When the condition is determined to be an ultra-high dynamic condition, a smooth switch is immediately triggered, switching from predictive feedforward compensation mode to traditional real-time compensation mode. Phase compensation is performed using the measured attitude of the IMU, and the process noise covariance matrix of the extended Kalman filter is readjusted.

[0018] The method further includes: The signal-to-interference-plus-noise ratio (SIR) of the received signal from the phased array antenna and the satellite receiving gain are collected as feedback evaluation indicators for beam pointing accuracy. If the SIR of the received signal continues to decrease or the satellite receiving gain is lower than the threshold, the feedforward compensation prediction phase is corrected based on the change in the feedback evaluation indicators.

[0019] The method further includes: Based on the vehicle's dynamic operating conditions, the process noise covariance matrix of the extended Kalman filter is adaptively tuned: the process noise covariance matrix is ​​increased under ultra-high dynamic conditions, and decreased under normal dynamic conditions. The actual position of the satellite is collected in real time and compared with the predicted position to update the orbital perturbation correction term of the satellite orbital dynamics model in real time.

[0020] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention constructs a rigid body dynamics model of the vehicle and a satellite orbital dynamics model, and uses an extended Kalman filter to predict the attitude and position of the vehicle at future moments. It takes into account the superposition effect of motion between the satellite and the ground, and improves the accuracy of the calculation of the relative position between the satellite and the ground.

[0021] (2) Based on the predicted azimuth and elevation angles of the satellite relative to the array surface, this invention combines the array element coordinates and wavenumbers to directly and quickly calculate the feedforward compensation phase of each array element through geometric relationships. It has low computational complexity and short calculation delay, which is suitable for the computing power and real-time requirements of vehicle-mounted terminals.

[0022] (3) The present invention calculates the prediction error in real time and compares it with a preset threshold. Based on the comparison result, it automatically switches between prediction feedforward compensation mode and traditional real-time compensation mode. When the prediction accuracy is high, feedforward compensation is used to obtain lower latency and higher dynamic performance. When the prediction error is large, it seamlessly switches back to real-time compensation, ensuring that the system can maintain a stable beam pointing under different degrees of motion intensity, avoiding phase jump and algorithm divergence. It can effectively cope with dynamic disturbances such as vehicle pitch vibration, road surface undulation, lane change sharp turn and high frequency bumps, and improve robustness.

[0023] (4) This invention accurately matches the vehicle attitude prediction step size with the total delay of the phased array control link, collects IMU, GNSS and ephemeris data in real time, and sends the calculated predicted phase to the array controller and stores it in a dedicated buffer. It is triggered at a time based on the PTP synchronization clock, which ensures that the timing of phase loading is strictly aligned with the timing of predicted attitude, further eliminating asynchronous errors in data transmission and execution, and enhancing the determinism and reliability of beam compensation. Attached Figure Description

[0024] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0025] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0026] The core objective of this embodiment is to address three major problems in vehicle-mounted phased arrays: IMU attitude measurement delay and compensation lag caused by noise; beam aging caused by phased array control link delay; and rapid changes in relative azimuth caused by the superposition of vehicle motion and satellite motion. A dynamic beam stabilization method for phased array terminals based on vehicle motion prediction is proposed, primarily solving three key technical problems: 1. Vehicle motion prediction model construction: How to use IMU, vehicle speed, angular velocity and GNSS information to construct a vehicle state dynamic model with practical engineering significance, and use EKF for short-term prediction? 2. Phased array feedforward phase solution: After the future attitude of the vehicle and the future orientation of the satellite are known, how to quickly calculate the array steering vector to reduce the complexity and time delay of the feedforward control calculation? 3. Dynamic switching mechanism: In high dynamic conditions, there may be errors in vehicle attitude prediction. How to design a beamforming scheme to avoid phase jumps, divergence or loss of lock? This embodiment uses multi-source data acquisition and preprocessing, vehicle motion model construction, EKF satellite-to-ground joint prediction of motion, feedforward phase compensation calculation, phase feedforward execution, dynamic fault-tolerant switching, and closed-loop feedback correction as its core processes. All algorithms are implemented on the industrial-grade main control board of the vehicle-mounted phased array terminal. The total latency of single-frame data calculation is ≤10ms, which is much smaller than the total latency of the phased array control link, meeting the real-time requirements of vehicle-mounted communication on the move. Specifically, for example... Figure 1 As shown, the method includes the following steps: S1: Real-time synchronous acquisition of vehicle IMU data, GNSS data, and ephemeris data, followed by filtering, outlier removal, and coordinate transformation preprocessing to eliminate noise and system errors, providing a clean and effective data foundation for subsequent model building and prediction.

[0027] S2: Construct a rigid body dynamics model of vehicle motion to predict the future attitude and position of the vehicle, and construct a satellite orbital dynamics model to predict the future position of the satellite, providing a basis for the joint prediction of the relative motion between the satellite and the ground.

[0028] Constructing a state vector-based system The state transition model describes the time evolution of the vehicle's motion state, resulting in the rigid body dynamics model of the vehicle's motion: ; in, State vector at time step , For three-axis angles, For triaxial angular velocity, These are the velocities of the three axes; The state transition function for rigid body dynamics is as follows: ; in The IMU sampling period The angular acceleration is a three-axis acceleration, obtained from the difference in the angular velocities of the IMU; The acceleration is a three-axis acceleration, obtained from the difference of GNSS velocities; Let be the system process noise, with a mean of 0 and a covariance of . Gaussian distribution, covariance Tuning based on the measurement accuracy of the IMU and GNSS.

[0029] Based on the six orbital elements provided by satellite ephemeris data, a two-body orbital model is used to construct a satellite orbital dynamics model to calculate the satellite's future position. Considering the orbital perturbations of low-Earth orbit satellites, an orbital perturbation correction term is introduced for the solution. The coordinates of the time satellite in ECEF (Earth-Centered Earth-Fixed) coordinate system : ; in, The six elements of a satellite orbit; This is the position calculation function for the two-body orbit model; This is an orbital perturbation correction term, caused by factors such as irregularities in the Earth's gravitational field and atmospheric drag, and is obtained through correction parameters from satellite ephemeris.

[0030] S3: Using preprocessed IMU and GNSS data as observations, combined with the vehicle motion rigid body dynamics model, extended Kalman filtering is used to calculate the vehicle's attitude and position at future moments. Combined with ephemeris data and satellite orbit dynamics model, the azimuth and elevation angles of the satellite relative to the array are predicted.

[0031] This process is divided into three stages: initialization, prediction, and update. Specifically, it includes the following steps: Initialization: Take the multi-source preprocessed data at the initial moment as the initial value and set the initial state vector. ,in, , , These are the initial attitude, angular velocity, and linear velocity; the initial error covariance matrix is ​​set. This is a diagonal matrix, with diagonal elements representing the measurement error variances of each state variable; the process noise covariance matrix is ​​set. and observation noise covariance matrix ; Prediction phase: based on Optimal state estimate at time t And vehicle motion rigid body dynamics model, predict Prior state vector at time step and prior error covariance matrix : ; ; in, State transition function The Jacobian matrix is ​​obtained by taking the partial derivative of the rigid body dynamics model, from the prior state vector. Mid-calculation of vehicle attitude prediction and predicted location ; Update phase: with The IMU and GNSS preprocessed data at time 1 are the observations. Construct observation equations Calculate Kalman gain And update the prior state and covariance to obtain Optimal state estimate at time t and the optimal error covariance matrix : ; ; ; in, For the observation matrix, For state vectors, To observe the noise vector, It is the identity matrix; The satellite-to-ground relative azimuth calculation is based on the predicted vehicle position. and satellite positions predicted based on satellite orbital dynamics models Calculate the position vector of the satellite relative to the vehicle array. ; this position vector Transform from the Earth-centered, Earth-fixed coordinate system to the local coordinate system of the front surface, and calculate the corresponding azimuth angle. and pitch angle The formula is: ; ; in, , , They are respectively Three-axis components in the local coordinate system of the array surface This indicates the magnitude of the position vector.

[0032] S4: Based on the predicted azimuth and elevation angles of the satellite relative to the array surface, combined with the element coordinates and wavenumbers of the phased array antenna, the feedforward compensation predicted phase of each element is quickly calculated.

[0033] Based on the predicted azimuth angle of the satellite relative to the array Pitch angle Combined with the element coordinates of the phased array antenna and wave number Fast calculation of feedforward compensation prediction phase for each array element : ; in, , , The element coordinates of the phased array antenna are respectively The three-axis components; This is the reference phase.

[0034] S5: The calculated feedforward compensation prediction phase is sent to the array controller of the phased array antenna. The array controller stores the phase parameters in a dedicated buffer and triggers the timing based on the PTP synchronization clock.

[0035] exist At any given time, the main control board transmits the feedforward compensation prediction phase of each array element via the CAN FD high-speed communication link. The phase parameters are sent to the array controller of the phased array antenna in one go. The array controller stores the phase parameters in a dedicated buffer and triggers the timing based on the PTP synchronization clock. Phase transmission: Communication link transmission delay ≤ 5ms, At +5ms, all phase parameters are cached to the array controller, with no packet loss and no errors. Precise loading: The array controller loads precisely according to the PTP clock. At +5ms, the feedforward compensation phase of the buffer is precisely loaded into the phase shifter of each array element. The loading completion time is perfectly aligned with the actual attitude change time of the vehicle, achieving precise matching of compensation delay and fundamentally solving the compensation lag problem of traditional solutions.

[0036] S6: Calculate the prediction error in real time and compare it with the threshold. Based on the comparison result, smoothly switch between prediction feedforward compensation mode and traditional real-time compensation mode to ensure beam stability. When prediction feedforward compensation mode is selected, continue to execute S1~S5. When traditional real-time compensation mode is selected, perform phase compensation based on the IMU data obtained in real time in S1.

[0037] First, calculate the prediction error: data collection Real-world attitude measurement of the IMU at any given moment , and predicted posture Compare and calculate the prediction error. Take the maximum value As a basis for error judgment; when When the condition is determined to be a normal dynamic operating condition, the predictive feedforward compensation mode is maintained, and S1~S5 are continued. The preset threshold; when When the condition is determined to be an ultra-high dynamic condition, a smooth switch is immediately triggered, switching from predictive feedforward compensation mode to traditional real-time compensation mode. Phase compensation is performed using the measured attitude of the IMU, and the process noise covariance matrix of the extended Kalman filter is readjusted.

[0038] S7: Collect the signal-to-interference-plus-noise ratio (SINR) and satellite receiver gain of the phased array antenna. As a feedback evaluation index for beam pointing accuracy, if the received signal-to-interference-plus-noise ratio (SINR) continues to decrease or the satellite receiving gain... If the value is below the threshold, then the feedforward compensation prediction phase is adjusted based on the change in the feedback evaluation index. Make corrections; Based on the vehicle's dynamic operating conditions, the process noise covariance matrix of the extended Kalman filter is adaptively tuned. Increasing the process noise covariance matrix under ultra-high dynamic conditions Reducing the process noise covariance matrix under normal dynamic operating conditions ; Real-time acquisition of the satellite's actual position, comparison with the predicted position, and adjustment of the orbital perturbation term in the satellite's orbital dynamics model. Real-time updates are performed to improve the accuracy of satellite position prediction.

[0039] The above methods revolve around the core requirement of vehicle-mounted phased array communication on the move, achieving a technological breakthrough from passive real-time compensation to active predictive feedforward compensation. The key points are as follows: 1. Delay matching and prediction mechanism for vehicle-mounted mobile communication: Innovatively, the vehicle attitude prediction step size is accurately matched with the total delay of the phased array control link to achieve accurate delay compensation; 2. Joint prediction model of relative motion between satellite and ground: It integrates the rigid body dynamics model of vehicle motion and the orbital dynamics model of satellite to achieve joint prediction of vehicle attitude and position and satellite position. It takes into account the superposition effect of motion between satellite and ground, and improves the accuracy of the calculation of relative position between satellite and ground. 3. Low-complexity fast feedforward phase solution algorithm: Based on the array steering vector method, the array element phase solution formula is constructed, and the feedforward compensation phase is directly solved through geometric relationship. It has low computational complexity and short solution latency, which is suitable for the computing power and real-time requirements of vehicle terminals. 4. Dynamic fault tolerance and smooth switching mechanism for high dynamic operating conditions: A mode switching strategy based on the prediction error threshold is designed, which combines phase interpolation to achieve a smooth transition between prediction feedforward compensation and traditional real-time compensation, avoiding phase jumps and algorithm divergence, and improving the robustness of the solution. 5. Hybrid compensation architecture of open-loop feedforward and closed-loop feedback: Based on open-loop predictive feedforward compensation, closed-loop feedback correction based on received signal quality is added, while adaptive optimization of EKF model parameters is realized, improving beam stabilization accuracy under all operating conditions.

[0040] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction, characterized in that, Includes the following steps: S1: Real-time synchronous acquisition of vehicle IMU data, GNSS data, and ephemeris data, and preprocessing of them; S2: Construct a rigid body dynamics model of vehicle motion to predict the future attitude and position of the vehicle, and construct a satellite orbital dynamics model to predict the future position of the satellite, providing a basis for the joint prediction of the relative motion between the satellite and the ground; S3: Using preprocessed IMU and GNSS data as observations, combined with the vehicle motion rigid body dynamics model, extended Kalman filtering is used to calculate the vehicle's attitude and position at future moments, and combined with ephemeris data and satellite orbit dynamics model, the azimuth and elevation angles of the satellite relative to the array are predicted. S4: Based on the predicted azimuth and elevation angles of the satellite relative to the array surface, combined with the element coordinates and wavenumber of the phased array antenna, the feedforward compensation prediction phase of each element is quickly calculated. S5: The calculated feedforward compensation prediction phase is sent to the array controller of the phased array antenna. The array controller stores the phase parameters in a dedicated buffer and triggers the timing based on the PTP synchronization clock. S6: Calculate the prediction error in real time and compare it with the threshold. Based on the comparison result, smoothly switch between prediction feedforward compensation mode and traditional real-time compensation mode to ensure beam stability. When prediction feedforward compensation mode is selected, continue to execute S1~S5. When traditional real-time compensation mode is selected, perform phase compensation based on the IMU data obtained in real time in S1.

2. The method of claim 1, wherein, The preprocessing includes filtering, outlier removal, and coordinate transformation.

3. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, In S2, a state vector-based system is constructed. The state transition model describes the time evolution of the vehicle's motion state, resulting in the rigid body dynamics model of the vehicle's motion: ； in, State vector at time step , For three-axis angles, For triaxial angular velocity, These are the velocities of the three axes; Here is the state transition function for rigid body dynamics, specifically: ； in The IMU sampling period The angular acceleration is a three-axis acceleration, obtained from the difference in the angular velocities of the IMU; The acceleration is a three-axis acceleration, obtained from the difference of GNSS velocities; Let be the system process noise, with a mean of 0 and a covariance of . Gaussian distribution, covariance Tuning based on the measurement accuracy of the IMU and GNSS.

4. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, In step S2, based on the six orbital elements provided by satellite ephemeris data, a two-body orbital model is used to construct a satellite orbital dynamics model to calculate the satellite's future position. Considering the orbital perturbations of low-Earth orbit satellites, an orbital perturbation correction term is introduced, and the solution is obtained. The coordinates of the satellite in the geocentric coordinate system : ； in, The six elements of a satellite orbit; This is the position calculation function for the two-body orbit model; This is the orbital perturbation correction term, obtained from the correction parameters of the satellite ephemeris.

5. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 3, characterized in that, The S3 is divided into three stages: initialization, prediction, and update. Specifically, it includes the following steps: Initialization: Take the multi-source preprocessed data at the initial moment as the initial value and set the initial state vector. ,in, , , These are the initial attitude, angular velocity, and linear velocity; the initial error covariance matrix is ​​set. This is a diagonal matrix, with diagonal elements representing the measurement error variances of each state variable; the process noise covariance matrix is ​​set. and observation noise covariance matrix ; Prediction phase: based on Optimal state estimate at time t And vehicle motion rigid body dynamics model, predict Prior state vector at time step and prior error covariance matrix : ； ； in, State transition function The Jacobian matrix is ​​obtained by taking the partial derivative of the rigid body dynamics model, from the prior state vector. Mid-calculation of vehicle attitude prediction and predicted location ; Update phase: with The IMU and GNSS preprocessed data at time 1 are the observations. Construct observation equations Calculate Kalman gain And update the prior state and covariance to obtain Optimal state estimate at time t and the optimal error covariance matrix : ； ； ； in, For the observation matrix, For state vectors, To observe the noise vector, It is the identity matrix; The satellite-to-ground relative azimuth calculation is based on the predicted vehicle position. and satellite positions predicted based on satellite orbital dynamics models Calculate the position vector of the satellite relative to the vehicle array. ; this position vector Transform from the Earth-centered, Earth-fixed coordinate system to the local coordinate system of the front surface, and calculate the corresponding azimuth angle. and pitch angle The formula is: ； ； in, , , They are respectively Three-axis components in the local coordinate system of the array surface This indicates the magnitude of the position vector.

6. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, Specifically, S4 is: Based on the predicted azimuth angle of the satellite relative to the array Pitch angle Combined with the element coordinates of the phased array antenna and wave number Fast calculation of feedforward compensation prediction phase for each array element : ； in, , , The element coordinates of the phased array antenna are respectively The three-axis components; This is the reference phase.

7. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, In step S5, the array controller, based on the PTP clock, The feedforward compensation prediction phase from the buffer is precisely loaded into the phase shifters of each array element at every moment, with the loading completion time perfectly aligned with the actual attitude change time of the vehicle. The feedforward compensation prediction phase is based on... The vehicle's attitude and position at any given time, as well as the relative orientation between the satellite and the ground, are calculated.

8. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, Specifically, S6 is: Calculate prediction error: Data collection Real-world attitude of the IMU at any given moment , and predicted posture Compare and calculate the prediction error. Take the maximum value As a basis for error judgment; when When the condition is determined to be a normal dynamic operating condition, the predictive feedforward compensation mode is maintained, and S1~S5 are continued. The preset threshold; when When the condition is determined to be an ultra-high dynamic condition, a smooth switch is immediately triggered, switching from predictive feedforward compensation mode to traditional real-time compensation mode. Phase compensation is performed using the measured attitude of the IMU, and the process noise covariance matrix of the extended Kalman filter is readjusted.

9. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, The method further includes: The signal-to-interference-plus-noise ratio (SIR) of the received signal from the phased array antenna and the satellite receiving gain are collected as feedback evaluation indicators for beam pointing accuracy. If the SIR of the received signal continues to decrease or the satellite receiving gain is lower than the threshold, the feedforward compensation prediction phase is corrected based on the change in the feedback evaluation indicators.

10. The method for dynamic beam stabilization of a phased array terminal based on vehicle motion prediction according to claim 1, characterized in that, The method further includes: Based on the vehicle's dynamic operating conditions, the process noise covariance matrix of the extended Kalman filter is adaptively tuned: the process noise covariance matrix is ​​increased under ultra-high dynamic conditions, and decreased under normal dynamic conditions. The actual position of the satellite is collected in real time and compared with the predicted position to update the orbital perturbation correction term of the satellite orbital dynamics model in real time.

Images