VICTS antenna attitude estimation and correction method and system based on satellite pointing feedback

Abstract

The application relates to the technical field of data processing, and discloses a VICTS antenna attitude estimation and correction method and system based on satellite pointing feedback. A tightly coupled deep fusion system taking an extended Kalman filter as a framework is constructed. On the basis of traditional IMU / GNSS fusion, satellite communication antenna pointing information is introduced as a new observation source. The Jacobian matrix of the observation source for the carrier attitude is derived, effective fusion of the observation source is realized, the attitude estimation precision and robustness of the carrier are reversely corrected and significantly improved, three-axis attitude error is more comprehensively and stably estimated and corrected, the problem of missing roll angle observation is fundamentally solved, different dimension observation information is unified into the same state estimation framework through accurate mathematical modeling, and the precision, autonomy and robustness of the system in a complex environment are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for VICTS antenna attitude estimation and correction based on satellite pointing feedback. Background Technology

[0002] Currently, mobile satellite communication systems such as Satcom-on-the-Move commonly employ inertial measurement units (IMUs) combined with Global Navigation Satellite Systems (GNSS) for vehicle attitude estimation. However, this method has inherent limitations: IMUs (gyroscopes) suffer from drift errors, which can lead to a decrease in attitude estimation accuracy over long periods of operation; while GNSS (such as RTK) can provide an absolute attitude reference, its low update frequency and susceptibility to signal interference make it difficult to meet the requirements of high dynamics and continuous stable tracking. Furthermore, existing methods often treat the satellite communication antenna as a mere payload, using its pointing information only as an output of the control system, failing to effectively utilize its inherent high-precision vehicle attitude observation value. This results in insufficient reliability of attitude estimation in complex electromagnetic environments or during brief GNSS signal interruptions, thus affecting the stability of the communication link. Therefore, there is an urgent need for a new method that can fully utilize existing sensor information, does not rely on additional hardware, and can effectively suppress accumulated errors, improving attitude estimation accuracy and robustness.

[0003] With the rapid development of mobile satellite communication technology, the demand for high-precision and high-stability tracking of mobile communication systems in highly dynamic environments is becoming increasingly urgent. The core challenge in achieving this goal lies in acquiring real-time and accurate carrier attitude information. Traditional fusion methods mainly face the following two technical challenges:

[0004] 1. Inertial navigation error accumulation, external correction lag, and dimension loss issues:

[0005] Existing solutions heavily rely on IMUs (Inertial Measurement Units) for high-frequency attitude estimation, but the inherent bias and drift of their gyroscopes cause three-axis attitude errors to accumulate over time. Although RTK (Real-Time Dynamic Differential Positioning) is introduced to provide an absolute attitude reference for correction, the RTK solution has inherent drawbacks: First, its update frequency is low (typically 1-10Hz) and it is prone to loss of lock when signals are blocked, failing to provide continuous and timely error correction for the IMU, resulting in a rapid decline in attitude estimation accuracy during high maneuvers or GNSS signal interruptions; Second, and more critically, RTK typically only provides observations of the yaw and pitch degrees of freedom, failing to effectively observe and correct the roll angle, forcing the roll angle error to rely solely on short-term IMU integration or static accelerometer observations, resulting in insufficient accuracy and reliability.

[0006] 2. Insufficient utilization of multi-sensor information and the problem of single observation dimension:

[0007] Traditional methods typically treat satellite communication antenna servo systems as purely "actuators" and "consumers," providing attitude commands unidirectionally while completely ignoring their value as high-precision "sensors." After locking onto a target satellite, the precise angular information of the antenna's beam pointing within the carrier's coordinate system contains valuable attitude observations of the carrier relative to the satellite's line-of-sight, information itself encompassing two degrees of freedom. This information, under current technological frameworks, has not been effectively mined and fed back to correct the carrier's own attitude estimation, resulting in a significant waste of information resources. Furthermore, existing RTK schemes can only provide observations of two degrees of freedom, exhibiting a missing dimension and failing to provide complete constraints on the three-axis attitude. Summary of the Invention

[0008] This invention provides a method and system for VICTS antenna attitude estimation and correction based on satellite pointing feedback, aiming to solve at least one of the above-mentioned technical problems.

[0009] To achieve the above objectives, this invention provides a method for VICTS antenna attitude estimation and correction based on satellite pointing feedback, the method comprising the following steps:

[0010] S1: Acquire multi-source attitude-dependent raw data of the VICTS antenna, and define the state vector of the EKF filter based on the multi-source attitude-dependent raw data; wherein, the multi-source attitude-dependent raw data includes IMU output data, RTK output data and satellite pointing parameters;

[0011] S2: Based on the state vector and IMU output data from the previous time step, the prior estimation of the current state vector is completed using the quaternion update formula;

[0012] S3: Based on the state transition Jacobian matrix and the posterior covariance of the previous time step, the prior estimation of the error covariance matrix is ​​completed by combining process noise;

[0013] S4: Construct a multi-source observation model that integrates IMU output data, RTK output data and satellite pointing parameters, solve the Jacobian matrix of each observation component and integrate it into a combined observation Jacobian matrix;

[0014] S5: Based on the prior estimate of the current state vector, the prior estimate of the error covariance matrix, and the combined observation Jacobian matrix, complete the posterior update of the state vector and covariance matrix, and extract the attitude information of the VICTS antenna from the updated posterior state vector.

[0015] Optionally, in step S1, the expression for the state vector is as follows:

[0016]

[0017] In the formula, Represents the state vector. It is a unit quaternion representing the VICTS antenna attitude, used to describe the three-dimensional attitude of the VICTS antenna. It is the zero bias vector of the gyroscope, representing the inherent bias error of the gyroscope's x, y, and z axes.

[0018] Optionally, step S2: Based on the state vector and IMU output data from the previous time step, the prior prediction of the current time step's state vector is completed using a quaternion update formula, the specific expression of which is:

[0019]

[0020]

[0021]

[0022] In the formula, This represents the prior estimate of the state vector at the current moment. This represents the quaternion prediction value at the current moment. This represents the zero-bias vector at the current moment. This represents the predicted quaternion value at the previous time step. This represents the zero-bias vector at the previous time step. This represents the angular velocity vector of the IMU output data, with components as follows: In , , , This indicates the discretization time step.

[0023] Optionally, step S3: Based on the state transition Jacobian matrix and the posterior covariance of the previous time step, and combined with process noise, a prior estimate of the error covariance matrix is ​​completed. The specific expression is as follows:

[0024]

[0025] In the formula, This represents the prior estimate error covariance matrix at the current time. Let represent the updated posterior covariance matrix from the previous time step. This represents the state transition Jacobian matrix, used to linearize the nonlinear state equations at the previous time step. Represents the process noise covariance matrix. The matrix representing the mapping from process noise to the state space. This indicates transpose.

[0026] Optionally, in step S4, the multi-source observation model specifically includes:

[0027] The gravity observation model corresponding to the IMU output data is expressed as follows:

[0028]

[0029] In the formula, Represents quaternions The obtained rotation matrix from the navigation system to the machine system, This represents the gravity vector in the navigation frame. This represents the predicted accelerometer measurement at time k;

[0030] The RTK observation model corresponding to the RTK output data is expressed as follows:

[0031]

[0032] In the formula, Indicates the yaw angle. Indicates pitch angle, The function representing the solution of quaternions to Euler angles. This represents the predicted RTK measurement value;

[0033] The satellite pointing parameter corresponds to the satellite pointing observation model, and the specific expression is as follows:

[0034]

[0035] In the formula, This represents the unit vector representing the satellite orientation in the navigation system. Represents quaternions The obtained rotation matrix from the navigation system to the machine system, This represents the predicted pointing vector of the machine system satellites.

[0036] Optionally, in step S4, the Jacobian matrix of each observation component is solved, specifically including:

[0037] Jacobian matrix corresponding to the gravity observation model:

[0038]

[0039] In the formula, This represents the Jacobian matrix of the accelerometer's predicted observation vector with respect to the state vector. This represents the partial derivative of the accelerometer prediction observation vector with respect to the quaternion component q. Let denote the partial derivative with respect to zero b, and be an all-zero matrix;

[0040] Jacobian matrix corresponding to the RTK observation model:

[0041]

[0042] In the formula, This represents the Jacobian matrix of the RTK prediction observation vector versus the state vector. They represent the yaw angles respectively. and pitch angle The partial derivative with respect to the quaternion q, Let denote the partial derivative with respect to zero b, and be an all-zero matrix;

[0043] The Jacobian matrix corresponding to the satellite pointing observation model:

[0044]

[0045] In the formula, This represents the Jacobian matrix of the satellite pointing to the predicted observation vector versus the state vector. Let q represent the rotation matrix constructed from the quaternion q, which rotates the navigation frame vector to the machine frame. Represents the rotated vector The partial derivative with respect to the quaternion q describes the sensitivity of changes in the quaternion to satellite pointing prediction observations. Let denote the partial derivative with respect to the zero bias b of the gyroscope, and be an all-zero matrix.

[0046] Optionally, in step S4, the results are integrated into a combined observation Jacobian matrix, the specific expression of which is:

[0047]

[0048] In the formula, This represents the combined observation Jacobian matrix, used for Kalman gain calculation. This indicates the sensitivity of accelerometer predictive observations to quaternions. This represents the sensitivity of accelerometer prediction observations to zero bias, and is an all-zero matrix. This indicates the sensitivity of RTK prediction observations to quaternions. This represents the sensitivity of RTK-predicted observations to zero bias, and is an all-zero matrix. This indicates the sensitivity of satellite pointing prediction observations to quaternions. This represents the sensitivity of satellite pointing prediction observations to zero bias, and is an all-zero matrix.

[0049] Optionally, in step S5, based on the prior estimate of the current state vector, the prior estimate of the error covariance matrix, and the combined observation Jacobian matrix, the posterior update of the state vector and covariance matrix is ​​completed, specifically including:

[0050]

[0051]

[0052]

[0053] In the formula, This represents the Kalman gain at time k. Let the prior error covariance matrix at time k be denoted as . Let Jacobian matrix be the combined observations at time k. Let the posterior error covariance matrix at time k be denoted as . Let k represent the observation noise covariance matrix at time k. Let k represent the posterior state vector at time k. Let k represent the prior state vector at time k. This represents the actual observation vector at time k. The observation function is a combination of IMU, RTK, and satellite pointing functions. Let represent the posterior error covariance matrix at time k. Represents the identity matrix. This indicates transpose.

[0054] Optionally, in step S5, the attitude information of the VICTS antenna is extracted from the updated posterior state vector, specifically including:

[0055] S51: From the updated posterior state vector Extracting unit quaternions The unit quaternion is converted into the three-dimensional orientation of the VICTS antenna to correct the orientation of the VICTS antenna.

[0056] S52: Transform the posterior state vector With the posterior error covariance matrix As the initial state for the next time step k+1, it helps to complete the prior prediction of the state vector for the next time step k+1 and the prediction of the prior error covariance matrix for the next time step, thus realizing continuous attitude estimation.

[0057] Furthermore, to achieve the above objectives, the present invention also provides a VICTS antenna attitude estimation and correction system based on satellite pointing feedback, comprising:

[0058] A definition module is used to acquire multi-source attitude-dependent raw data of VICTS antennas and define the state vector of EKF filter based on the multi-source attitude-dependent raw data; wherein, the multi-source attitude-dependent raw data includes IMU output data, RTK output data and satellite pointing parameters;

[0059] The first prediction module is used to perform prior estimation of the current state vector based on the state vector of the previous time step and the IMU output data, using a quaternion update formula.

[0060] The second prediction module is used to perform prior estimation of the error covariance matrix based on the state transition Jacobian matrix and the posterior covariance of the previous time step, combined with process noise.

[0061] The solver module is used to construct a multi-source observation model that integrates IMU output data, RTK output data and satellite pointing parameters, solve the Jacobian matrix of each observation component and integrate it into a combined observation Jacobian matrix;

[0062] The update module is used to complete the posterior update of the state vector and covariance matrix based on the prior estimation of the current state vector, the prior estimation of the error covariance matrix, and the combined observation Jacobian matrix, and extract the attitude information of the VICTS antenna from the updated posterior state vector.

[0063] The beneficial effects of this invention are as follows: It proposes a VICTS antenna attitude estimation and correction method and system based on satellite pointing feedback. By constructing a tightly coupled deep fusion system with extended Kalman filtering as the framework, it innovatively introduces satellite antenna pointing information as a new observation source on the basis of traditional IMU / GNSS fusion. By deriving its Jacobian matrix for the carrier attitude, it achieves effective fusion of this observation, thereby correcting and significantly improving the accuracy and robustness of the carrier attitude estimation. It achieves a more comprehensive and robust estimation and correction of three-axis attitude errors, fundamentally solving the problem of missing roll angle observations. Through precise mathematical modeling, it unifies observation information from different dimensions into the same state estimation framework, greatly improving the system's accuracy, autonomy, and robustness in complex environments. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating the VICTS antenna attitude estimation and correction method based on satellite pointing feedback, according to an embodiment of the present invention.

[0065] Figure 2 This is a block diagram of the attitude estimation system in an embodiment of the present invention;

[0066] Figure 3 This is a flowchart of the EKF fusion algorithm in an embodiment of the present invention;

[0067] Figure 4 This is a schematic diagram of the structure of the VICTS antenna attitude estimation and correction system based on satellite pointing feedback according to an embodiment of the present invention. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0069] This invention provides a method for VICTS antenna attitude estimation and correction based on satellite pointing feedback, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the VICTS antenna attitude estimation and correction method based on satellite pointing feedback, as described in an embodiment of the present invention.

[0070] In this embodiment, a VICTS antenna attitude estimation and correction method based on satellite pointing feedback is provided, the method comprising the following steps:

[0071] S1: Acquire multi-source attitude-dependent raw data of the VICTS antenna, and define the state vector of the EKF filter based on the multi-source attitude-dependent raw data; wherein, the multi-source attitude-dependent raw data includes IMU output data, RTK output data and satellite pointing parameters;

[0072] S2: Based on the state vector and IMU output data from the previous time step, the prior estimation of the current state vector is completed using the quaternion update formula;

[0073] S3: Based on the state transition Jacobian matrix and the posterior covariance of the previous time step, the prior estimation of the error covariance matrix is ​​completed by combining process noise;

[0074] S4: Construct a multi-source observation model that integrates IMU output data, RTK output data and satellite pointing parameters, solve the Jacobian matrix of each observation component and integrate it into a combined observation Jacobian matrix;

[0075] S5: Based on the prior estimate of the current state vector, the prior estimate of the error covariance matrix, and the combined observation Jacobian matrix, complete the posterior update of the state vector and covariance matrix, and extract the attitude information of the VICTS antenna from the updated posterior state vector.

[0076] It should be noted that with the rapid development of mobile satellite communication technology, the demand for high-precision and high-stability tracking of mobile communication systems in highly dynamic environments is becoming increasingly urgent. The core challenge in achieving this goal lies in acquiring real-time and accurate carrier attitude information. Traditional fusion methods mainly face the following two technical challenges: the accumulation of inertial navigation errors and the lag in external correction, as well as the problem of missing dimensions; and the problem of insufficient utilization of multi-sensor information and the problem of single observation dimension.

[0077] To address the aforementioned problems, this embodiment proposes a novel key technology solution:

[0078] (1) Constructing a tightly coupled deep fusion framework for multi-source heterogeneous observations:

[0079] A novel state and observation model is designed, integrating azimuth / elevation angle observations provided by RTK with beam pointing vector (the satellite's direction in the carrier coordinate system) observations provided by the satellite antenna as observations, and deeply fusing them with the IMU. The innovation of this framework lies in the collaborative fusion of absolute angle observations from RTK and relative vector observations from satellite pointing. Although RTK and satellite pointing can only correct two degrees of freedom when used individually, their physical sources and mathematical expressions differ. When both are simultaneously introduced into the filter, it is equivalent to providing observation information for four degrees of freedom, forming a redundant observation system for the carrier's three-axis attitude (three degrees of freedom). This framework can continuously suppress IMU drift using high-frequency satellite pointing information even when RTK data is intermittently updated or even briefly lost. More importantly, even when RTK is available, the joint observation can achieve a more comprehensive and robust estimation and correction of three-axis attitude errors by establishing indirect but effective observation constraints on the roll angle, fundamentally solving the problem of missing roll angle observations.

[0080] (2) Establish and integrate the satellite pointing backward correction model:

[0081] The core of this approach lies in deriving the analytical Jacobian matrix of the carrier attitude state (quaternion) based on satellite pointing observations, and efficiently incorporating this observation into the EKF update process. This enables the system to utilize feedback information from the antenna servo system to reverse-calculate and correct the carrier's global attitude, transforming "one-way control" into "two-way feedback" and turning the antenna system from a passive execution unit into an active observation sensor. This model forms the basis for achieving synergistic effects between RTK observations and satellite pointing observations. Through precise mathematical modeling, it unifies observation information from different dimensions into a single state estimation framework, significantly improving the system's accuracy, autonomy, and robustness in complex environments.

[0082] Specifically, this invention provides a method for on-the-move carrier attitude estimation based on multi-sensor information fusion. The core of this method lies in constructing a tightly coupled deep fusion system based on the Extended Kalman Filter (EKF). Building upon traditional IMU / GNSS fusion, satellite antenna pointing information is introduced as a new observation source. By deriving its Jacobian matrix for carrier attitude, effective fusion of this observation is achieved, thereby inversely correcting and significantly improving the accuracy and robustness of carrier attitude estimation. The system collects raw data from IMU, RTK, and satellite antennas, preprocesses it, and then inputs it into the core EKF filter for fusion processing. The EKF utilizes the IMU gyroscope for high-frequency state prediction and updates the state using the three observation sources (accelerometer, RTK, and satellite pointing vector), ultimately outputting a high-precision carrier attitude estimation result.

[0083] like Figure 2 and Figure 3 The figures shown are a block diagram of the attitude estimation system of the present invention and a flowchart of the EKF fusion algorithm, respectively. A VICTS antenna attitude estimation and correction method based on satellite pointing feedback includes the following specific implementation steps:

[0084] Step 1: Define the state vector.

[0085] The state vector X of the EKF filter is defined as:

[0086]

[0087] The state vector has 7 dimensions, where:

[0088] X is a state vector, representing the overall state of the system. It is a unit quaternion representing the attitude of the carrier, used to describe the three-dimensional attitude of the carrier (yaw angle, pitch angle, roll angle). It is the zero bias vector of the gyroscope, representing the inherent bias error of the gyroscope's x, y, and z axes.

[0089] Step 2: State prediction.

[0090] Gyroscope angular velocity measurement is Discretization time step is .

[0091]

[0092] in The quaternion angular velocity matrix:

[0093]

[0094] The complete formula for predicting the state is:

[0095]

[0096] Here It refers to the prior estimate of the state variable at the current moment, conforming to the standard form of discretization and prediction. Where:

[0097] : The predicted value of the quaternion at the current moment. : Quaternion value at the previous time step. : The gyroscope's zero-bias prediction value at the current moment. The zero bias value of the gyroscope at the previous moment. The angular velocity vector (unit: rad / s) measured by the gyroscope has the following components: In , , .

[0098] Step 3: Covariance prediction.

[0099]

[0100] The recursive formula for the prediction error covariance matrix is ​​derived by estimating the covariance matrix based on the state at the previous time step, linearizing the nonlinear state equation using the state transition Jacobian matrix, and adding the process noise covariance to obtain the prior covariance matrix at the current time step. Where:

[0101] The prior estimation error covariance matrix (prediction covariance) at the current moment describes the uncertainty of the state prediction. : The posterior covariance matrix updated in the previous time step. The state transition Jacobian matrix linearizes the nonlinear state equation at the previous time step. The process noise covariance matrix reflects the contribution of system model errors or noise from sensors such as gyroscopes and accelerometers. The mapping matrix from process noise to state space. This indicates transpose.

[0102] Step 4: Observation Model.

[0103] The key to observational updates lies in constructing an observational model that integrates multiple observational sources. The EKF observational model is as follows:

[0104]

[0105] : Represents the actual measured observation vector. : Represents the accelerometer measurement vector, used to reflect the gravitational acceleration component of the carrier in the machine system. : Indicates the yaw and pitch angles measured by the dual-antenna RTK. : Represents the satellite direction vector obtained from satellite observation. : The prediction function from state to observation. Observation noise reflects the measurement error of the sensor.

[0106] Observation function It consists of the following components:

[0107] 1. Gravity observation model (gravity vector constraint).

[0108] An airborne triaxial accelerometer is used to measure the gravitational acceleration component acting on the carrier. The vector measured by the accelerometer is used to constrain the carrier's attitude, ensuring that the attitude calculation remains consistent with the direction of gravity.

[0109] 2. RTK observation model.

[0110] Using a dual-antenna RTK receiver, the yaw and pitch angles of the carrier are calculated via a baseline direction. The measurement results are used as the observation input EKF to achieve attitude correction.

[0111] 3. Satellite observation model.

[0112] By using satellite information as an observation sensor, the satellite orientation vector in the navigation system is obtained. A rotation matrix is ​​used to map the navigation system satellite vector to the machine system, which is then compared with the predicted machine system orientation vector. This process corrects the vehicle attitude in reverse, improving the accuracy and robustness of attitude estimation.

[0113] The following is a detailed explanation of each observation component.

[0114] 1. Gravity observation model (accelerometer):

[0115] (1) Formula for predicting observed values:

[0116]

[0117] The gravity vector of the navigation system is transformed into the mechanical system through a rotation matrix to obtain the predicted accelerometer measurement value. : by quaternions The obtained rotation matrix from the navigation system to the machine system. : Gravity vector in the navigation frame (positive downwards, unit m / s²). : The accelerometer measurement predicted at time k.

[0118] (2) Corresponding Jacobian matrix:

[0119]

[0120] : Represents the Jacobian matrix of the accelerometer's predicted observation vector to the state vector. : represents the partial derivative of the accelerometer predictive observation vector with respect to the quaternion component q. : represents the partial derivative with respect to zero bias b. Since the observation does not depend on zero bias, it is all zero.

[0121] 2. RTK Attitude Observation Model:

[0122] (1) Formula for predicting observed values:

[0123]

[0124] The attitude quaternion at the current moment The calculated yaw angle and pitch angle , : Expression for yaw angle. : Pitch angle expression. : The function for solving quaternions into Euler angles. : Predicted RTK measurement value.

[0125] (2) Corresponding Jacobian matrix:

[0126]

[0127] : Represents the Jacobian matrix of the RTK prediction observation vector to the state vector. : represent yaw angles respectively and pitch angle The partial derivative with respect to the quaternion q. : denotes the partial derivative with respect to zero bias b, all zero.

[0128] 3. Satellite pointing observation model:

[0129] (1) Formula for predicting observed values:

[0130]

[0131] In this invention, the satellite orientation unit vector in the navigation coordinate system is first... By using the quaternion of the current predicted attitude of the system Constructed rotation matrix Perform a coordinate transformation to convert it from the navigation system to the aircraft system, and obtain the aircraft system's predicted satellite pointing vector at the current moment. Subsequently, the predicted vector is compared with the actual measured machine-system satellite orientation vector, and the difference between the two is calculated as the observation residual. This residual information is then used to adjust the state vector. Corrections and updates are made to achieve accurate estimation of the carrier attitude and gyroscope zero bias.

[0132] : Satellite orientation unit vector in the navigation system. : by quaternions The obtained rotation matrix from the navigation system to the machine system. : Predicted pointing vector of the machine system satellites.

[0133] (2) Corresponding Jacobian matrix:

[0134]

[0135] In the Jacobian matrix of the satellite pointing prediction observation vector to the state vector, the left-hand submatrix about quaternions is the core analytical expression derived in this scheme. It accurately describes the sensitivity of minute changes in quaternions to the predicted satellite pointing vector. Using this Jacobian matrix, the observed actual satellite pointing information is compared with the predicted value. The residual is then applied inversely to the state vector, achieving precise correction of the carrier attitude. This inverse correction method based on the satellite orientation vector significantly improves attitude estimation accuracy, enhances the stability and reliability of the system under multi-sensor fusion, and effectively suppresses accumulated errors caused by gyroscope drift or zero bias, making the carrier attitude calculation more accurate and robust.

[0136] : Represents the Jacobian matrix of the satellite pointing to the predicted observation vector on the state vector. : Represents a rotation matrix constructed from quaternions q, which rotates the navigation system vector to the machine system. : Represents the rotated vector The partial derivative with respect to the quaternion q describes the sensitivity of changes in the quaternion to satellite pointing prediction observations. : represents the partial derivative with respect to the zero bias b of the gyroscope. Since satellite pointing observation does not depend on the zero bias, it is an all-zero matrix.

[0137] 4. Combined observation Jacobian matrix:

[0138]

[0139] : Represents the combined observation Jacobian matrix, directly used for Kalman gain calculation. Block matrix explanation: : Indicates the sensitivity of accelerometer predictive observations to quaternions. : Indicates the sensitivity of accelerometer prediction observations to zero bias (all-zero matrix). : Indicates the sensitivity of RTK prediction observations to quaternions. : Indicates the sensitivity of RTK prediction observations to zero bias (all-zero matrix). : Indicates the sensitivity of satellite pointing prediction observations to quaternions. : Indicates the sensitivity of satellite pointing prediction observations to zero bias (all-zero matrix).

[0140] Step 5: Observation update (measurement update).

[0141] 1. Kalman gain calculation:

[0142]

[0143] : Represents the Kalman gain at time k. : Represents the prior error covariance matrix at time k. : Represents the Jacobian matrix of the combined observations at time k. : Represents the posterior error covariance matrix at time k. : Represents the observation noise covariance matrix (measurement noise) at time k. This indicates transpose.

[0144] 2. State update equation:

[0145]

[0146] : Represents the posterior (updated) state vector at time k. : Represents the prior (predicted) state vector at time k, obtained in step two. : Represents the actual observation vector (measurement) at time k, obtained in step four. : Represents the observation function (mapping the state to the observation space), which in this system is a combination function of gravity / RTK / satellite pointing.

[0147] 3. Covariance Update Equation (Joseph Form)

[0148]

[0149] : Represents the posterior error covariance matrix at time k. : Represents the identity matrix. This indicates transpose.

[0150] Compared with existing technologies, the mobile carrier attitude estimation method based on satellite pointing information reverse correction provided by this invention has the following significant advantages:

[0151] 1. Significantly improves the long-term accuracy and stability of attitude estimation.

[0152] Traditional methods heavily rely on gyroscopes for high-frequency attitude estimation, and their inherent bias and drift cause errors to accumulate over time. This invention creatively uses the satellite pointing vector provided by the satellite communication antenna as a new type of observation and deeply integrates it into the EKF framework. This observation provides the system with a high-precision external attitude reference independent of the IMU and GNSS. Even during RTK signal loss or update intervals, the system can still continuously suppress and correct IMU drift using high-frequency (servo system control frequency, up to 100Hz or higher) satellite pointing observations. This effectively solves the problem of rapidly decreasing attitude estimation accuracy due to GNSS signal interruption or low update rate in traditional methods, achieving stable attitude output with high precision over long periods.

[0153] 2. Provides multi-dimensional global observation redundancy, significantly enhances the accuracy of ROLL angle estimation, and makes up for the structural defects of RTK.

[0154] While existing technologies like RTK (Real-Time Kinematics) can provide high-precision yaw and pitch observations, their observation mechanisms limit their ability to directly provide absolute roll observations. In dynamic scenarios such as UAVs, roll angle estimation typically relies on accelerometers, but during continuous linear acceleration / deceleration or turning, non-gravitational acceleration disturbances introduce significant errors, leading to distorted roll angle estimations. This invention introduces a satellite pointing vector as another global observation source. From an observation dimension perspective, RTK provides observations for two degrees of freedom: yaw and pitch, while satellite pointing provides directional observations for the other two degrees of freedom. The two sets of observations overlap but have different sensitive axes. Through optimized fusion, satellite pointing provides crucial observational constraints for the roll angle, which is missing in RTK, jointly constructing a complete three-dimensional attitude observation system. This fundamentally solves the problem of missing roll dimension observations in a single RTK, significantly improving the accuracy and reliability of full attitude estimation in dynamic environments.

[0155] 3. Greatly enhances the robustness and reliability of the system in complex environments.

[0156] Existing technologies suffer from drastic performance degradation in scenarios where GNSS signals are easily obstructed or interfered with, such as urban canyons, tunnels, and tree-lined roads. This invention introduces satellite pointing observations to construct a three-source redundant fusion architecture of IMU / GNSS / satellite pointing. The three observation sources complement each other in terms of characteristics: IMU is high-frequency but divergent, GNSS is absolute but low-frequency and volatile, and satellite pointing is high-frequency and absolute. When the GNSS (RTK) signal temporarily fails, the system can automatically degrade to the IMU / satellite pointing fusion mode, still maintaining high-precision attitude tracking, unlike traditional solutions that rely solely on continuously divergent pure inertial navigation. This "never-disconnect" capability greatly enhances the adaptability and mission reliability of the mobile communication system in complex and harsh environments.

[0157] 4. Fully explore and utilize the potential of existing hardware to achieve "sensor reuse," which is highly economical.

[0158] This invention eliminates the need for any additional hardware sensors (such as more IMUs, magnetometers, etc.). Instead, it innovatively transforms the satellite communication antenna servo system from a simple "actuator" into a dual-function "attitude sensor." Through innovation at the software algorithm level, it deeply mines and utilizes previously idle and wasted information resources (the antenna's precise pointing angle) within the system, achieving "value-added" and "reuse" of hardware resources. This not only solves the technical challenges but also avoids the problems of increased costs, increased system complexity, and reduced reliability caused by adding hardware. It achieves a leap in system performance at minimal cost, demonstrating extremely high cost-effectiveness and engineering application value.

[0159] 5. It has achieved a leap from "one-way open-loop" to "two-way closed-loop" attitude estimation, thereby improving the system's autonomy.

[0160] In traditional solutions, the attitude estimation system unidirectionally outputs control commands to the antenna servo system, forming an open-loop chain. This invention corrects the carrier attitude by feeding back antenna pointing information, creating a negative feedback closed-loop control system of "attitude estimation - antenna pointing - feedback correction." This bidirectional closed-loop architecture significantly improves the system's automation level and robustness. The system can achieve dynamic self-calibration and online error correction, reducing absolute dependence on a single external sensor such as GNSS, thereby maintaining more stable and reliable service in complex environments.

[0161] Therefore, through innovative design at the algorithm level, this invention effectively solves the long-standing problem of high-precision and high-reliability attitude estimation in the field of mobile communication. It has the advantages of high accuracy, strong robustness, low cost, and easy engineering implementation, and has broad prospects for promotion and application.

[0162] Reference Figure 4 , Figure 4This is a schematic diagram of the structure of the VICTS antenna attitude estimation and correction system based on satellite pointing feedback according to an embodiment of the present invention.

[0163] like Figure 4 As shown, the VICTS antenna attitude estimation and correction system based on satellite pointing feedback proposed in this embodiment of the invention includes:

[0164] The definition module 10 is used to acquire multi-source attitude-dependent raw data of the VICTS antenna and define the state vector of the EKF filter based on the multi-source attitude-dependent raw data; wherein, the multi-source attitude-dependent raw data includes IMU output data, RTK output data and satellite pointing parameters;

[0165] The first prediction module 20 is used to perform prior estimation of the current state vector based on the state vector of the previous time step and the IMU output data, using a quaternion update formula.

[0166] The second prediction module 30 is used to perform prior estimation of the error covariance matrix based on the state transition Jacobian matrix and the posterior covariance of the previous time step, combined with process noise.

[0167] Solver module 40 is used to construct a multi-source observation model that integrates IMU output data, RTK output data and satellite pointing parameters, solve the Jacobian matrix of each observation component and integrate it into a combined observation Jacobian matrix;

[0168] The update module 50 is used to complete the posterior update of the state vector and covariance matrix based on the prior estimation of the current state vector, the prior estimation of the error covariance matrix, and the combined observation Jacobian matrix, and to extract the attitude information of the VICTS antenna from the updated posterior state vector.

[0169] Other embodiments or specific implementations of the VICTS antenna attitude estimation and correction system based on satellite pointing feedback of the present invention can be referred to the above-described method embodiments, and will not be repeated here.

[0170] It is understood that in the description of this specification, references to terms such as "one embodiment," "another embodiment," "other embodiments," or "first embodiment to Nth embodiment," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0171] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0172] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for VICTS antenna attitude estimation and correction based on satellite pointing feedback, characterized in that, The method includes the following steps: S1: Acquire multi-source attitude-dependent raw data of the VICTS antenna, and define the state vector of the EKF filter based on the multi-source attitude-dependent raw data; wherein, the multi-source attitude-dependent raw data includes IMU output data, RTK output data and satellite pointing parameters; S2: Based on the state vector and IMU output data from the previous time step, the prior estimation of the current state vector is completed using the quaternion update formula; S3: Based on the state transition Jacobian matrix and the posterior covariance of the previous time step, the prior estimation of the error covariance matrix is ​​completed by combining process noise; S4: Construct a multi-source observation model that integrates IMU output data, RTK output data and satellite pointing parameters, solve the Jacobian matrix of each observation component and integrate it into a combined observation Jacobian matrix; S5: Based on the prior estimate of the current state vector, the prior estimate of the error covariance matrix, and the combined observation Jacobian matrix, complete the posterior update of the state vector and covariance matrix, and extract the attitude information of the VICTS antenna from the updated posterior state vector.

2. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 1, characterized in that, In step S1: the expression for the state vector is as follows: In the formula, Represents the state vector. It is a unit quaternion representing the VICTS antenna attitude, used to describe the three-dimensional attitude of the VICTS antenna. It is the zero bias vector of the gyroscope, representing the inherent bias error of the gyroscope's x, y, and z axes.

3. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 1, characterized in that, Step S2: Based on the state vector and IMU output data from the previous time step, the prior prediction of the current state vector is completed using the quaternion update formula. The specific expression is as follows: In the formula, This represents the prior estimate of the state vector at the current moment. This represents the quaternion prediction value at the current moment. This represents the zero-bias vector at the current moment. This represents the predicted quaternion value at the previous time step. This represents the zero-bias vector at the previous time step. This represents the angular velocity vector of the IMU output data, with components as follows: In , , , This indicates the discretization time step.

4. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 1, characterized in that, Step S3: Based on the state transition Jacobian matrix and the posterior covariance of the previous time step, and combined with process noise, complete the prior estimation of the error covariance matrix. The specific expression is as follows: In the formula, This represents the prior estimate error covariance matrix at the current time. Let represent the updated posterior covariance matrix from the previous time step. This represents the state transition Jacobian matrix, used to linearize the nonlinear state equations at the previous time step. Represents the process noise covariance matrix. The matrix representing the mapping from process noise to the state space. This indicates transpose.

5. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 1, characterized in that, In step S4, the multi-source observation model specifically includes: The gravity observation model corresponding to the IMU output data is expressed as follows: In the formula, Represents quaternions The obtained rotation matrix from the navigation system to the machine system, This represents the gravity vector in the navigation frame. This represents the predicted accelerometer measurement at time k; The RTK observation model corresponding to the RTK output data is expressed as follows: In the formula, This represents the yaw angle at time k. This represents the pitch angle at time k. The function representing the solution of quaternions to Euler angles. This represents the predicted RTK measurement value; The satellite pointing parameter corresponds to the satellite pointing observation model, and the specific expression is as follows: In the formula, This represents the unit vector representing the satellite orientation in the navigation system. Describe the quaternion at time k. The obtained rotation matrix from the navigation system to the machine system, This represents the predicted pointing vector of the machine system satellites.

6. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 5, characterized in that, In step S4, the Jacobian matrix of each observation component is solved, specifically including: Jacobian matrix corresponding to the gravity observation model: In the formula, This represents the Jacobian matrix of the accelerometer's predicted observation vector with respect to the state vector. This represents the partial derivative of the accelerometer prediction observation vector with respect to the quaternion q. Let denote the partial derivative with respect to zero b, and be an all-zero matrix; Jacobian matrix corresponding to the RTK observation model: In the formula, This represents the Jacobian matrix of the RTK prediction observation vector versus the state vector. They represent the yaw angles respectively. and pitch angle The partial derivative with respect to the quaternion q, Let denote the partial derivative with respect to zero b, and be an all-zero matrix; The Jacobian matrix corresponding to the satellite pointing observation model: In the formula, This represents the Jacobian matrix of the satellite pointing to the predicted observation vector versus the state vector. Let q represent the rotation matrix constructed from the quaternion q, which rotates the navigation frame vector to the machine frame. Represents the rotated vector The partial derivative with respect to the quaternion q describes the sensitivity of changes in the quaternion to satellite pointing prediction observations. Let denote the partial derivative with respect to the zero bias b of the gyroscope, and be an all-zero matrix.

7. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 6, characterized in that, In step S4, the results are integrated into a combined observation Jacobian matrix, the specific expression of which is: In the formula, This represents the combined observation Jacobian matrix, used for Kalman gain calculation. This indicates the sensitivity of accelerometer predictive observations to quaternions. This represents the sensitivity of accelerometer prediction observations to zero bias, and is an all-zero matrix. This indicates the sensitivity of RTK prediction observations to quaternions. This represents the sensitivity of RTK-predicted observations to zero bias, and is an all-zero matrix. This indicates the sensitivity of satellite pointing prediction observations to quaternions. This represents the sensitivity of satellite pointing prediction observations to zero bias, and is an all-zero matrix.

8. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 1, characterized in that, In step S5, based on the prior estimate of the current state vector, the prior estimate of the error covariance matrix, and the combined observation Jacobian matrix, the posterior update of the state vector and covariance matrix is ​​completed, specifically including: In the formula, This represents the Kalman gain at time k. Let the prior error covariance matrix at time k be denoted as . Let Jacobian matrix be the combined observations at time k. Let represent the posterior error covariance matrix at time k. Let k represent the observation noise covariance matrix at time k. Let k represent the posterior state vector at time k. Let k represent the prior state vector at time k. This represents the actual observation vector at time k. The observation function is a combination function of IMU, RTK, and satellite pointing. Let represent the posterior error covariance matrix at time k. Represents the identity matrix. This indicates transpose.

9. The VICTS antenna attitude estimation and correction method based on satellite pointing feedback as described in claim 1, characterized in that, In step S5, the attitude information of the VICTS antenna is extracted from the updated posterior state vector, specifically including: S51: From the updated posterior state vector Extract the quaternion at time k. The quaternion Convert to the three-dimensional orientation of the VICTS antenna to correct the orientation of the VICTS antenna; S52: Transform the posterior state vector With the posterior error covariance matrix As the initial state for the next time step k+1, it helps to complete the prior prediction of the state vector for the next time step k+1 and the prediction of the prior error covariance matrix for the next time step, thus realizing continuous attitude estimation.

10. A VICTS antenna attitude estimation and correction system based on satellite pointing feedback, characterized in that, include: A definition module is used to acquire multi-source attitude-dependent raw data of VICTS antennas and define the state vector of EKF filter based on the multi-source attitude-dependent raw data; wherein, the multi-source attitude-dependent raw data includes IMU output data, RTK output data and satellite pointing parameters; The first prediction module is used to perform prior estimation of the current state vector based on the state vector of the previous time step and the IMU output data, using a quaternion update formula. The second prediction module is used to perform prior estimation of the error covariance matrix based on the state transition Jacobian matrix and the posterior covariance of the previous time step, combined with process noise. The solver module is used to construct a multi-source observation model that integrates IMU output data, RTK output data and satellite pointing parameters, solve the Jacobian matrix of each observation component and integrate it into a combined observation Jacobian matrix; The update module is used to complete the posterior update of the state vector and covariance matrix based on the prior estimation of the current state vector, the prior estimation of the error covariance matrix, and the combined observation Jacobian matrix, and extract the attitude information of the VICTS antenna from the updated posterior state vector.

Images