A method, system, and media for on-orbit beam pointing prediction of a radar antenna

Abstract

The application provides a method, system and medium for on-orbit beam pointing prediction of a radar antenna, comprising: obtaining test data of a satellite ground test, including thermal load state, radar antenna array normal pointing and star sensor optical axis pointing; establishing a multi-layer neural network according to the ground test data; after the satellite is launched into orbit, recording the on-orbit thermal load state, radar antenna beam pointing information and star sensor observation results at the time of carrying out external calibration; using the foregoing multi-layer neural network and the on-orbit thermal load state during the external calibration, predicting the radar antenna array normal pointing and the star sensor optical axis pointing; using the radar antenna beam pointing information, the star sensor observation results and the predicted array normal pointing and star sensor optical axis pointing, correcting the foregoing multi-layer neural network; when the satellite runs to other positions, using the corrected neural network to predict the radar antenna beam pointing at the time according to the on-orbit thermal load state of the satellite at the position.

Description

Technical Field

[0001] This invention relates to the field of aerospace satellite technology, and to a method for predicting the beam pointing of a phased array radar antenna during long-term on-orbit operation of a satellite, based on a modified multilayer neural network established by ground experiments and on-orbit calibration. Specifically, it relates to a method for predicting the on-orbit beam pointing of a satellite phased array radar antenna, and more particularly to a method, system, and medium for predicting the on-orbit beam pointing of a radar antenna. Background Technology

[0002] Synthetic Aperture Radar (SAR) is an active Earth observation system that can be installed on aircraft, satellites, and other flight platforms to conduct all-weather, 24 / 7 Earth observation. In recent years, with the rapid development of aerospace technology, SAR has been increasingly applied to spaceborne Earth observation, with spaceborne radar antennas in the form of planar phased arrays being the most widely used. Its basic working principle involves transmitting directional microwaves towards a predetermined target and receiving the echo to image that target. Therefore, improving the accuracy of the microwave beam pointing emitted by the antenna can significantly reduce the deviation angle between the beam pointing and the expected pointing, enhance the radiation intensity of the predetermined target, and thus improve the observation and imaging performance. In practical engineering applications, the microwave beam pointing emitted by the radar antenna is usually defined as the microwave beam pointing under the satellite's flight reference. Since satellites use star sensors as attitude measurement and control sensors to measure the satellite's position and attitude in space, for radar satellites, the microwave beam pointing information emitted by the radar antenna can ultimately be described as the radar antenna microwave beam pointing vector in the satellite's star sensor coordinate system.

[0003] Existing phased array radar antennas require deployment via a unfolding mechanism after entering orbit to form a complete antenna. During this process, factors such as the vibration environment during satellite launch, the unfolding mechanism, and the orbital thermal environment all affect the microwave beam pointing of the radar antenna, causing a deviation between the on-orbit beam pointing and the beam pointing obtained from ground tests, thus impacting the actual performance of the radar antenna during its on-orbit operation. It is well known that radar satellites employing large phased array antennas require deployment via a mechanism after entering orbit to form a complete antenna. During this process, factors such as the vibration environment during satellite launch, the unfolding accuracy of the mechanism, and the orbital thermal environment all affect the shape and position accuracy of the radar antenna, thereby influencing its beam pointing.

[0004] Regarding how to obtain the on-orbit beam pointing information of radar antennas, existing research and engineering practices mainly rely on controlling the antenna's shape and position accuracy to minimize the impact of the aforementioned transmission process and on-orbit environment on the mechanical performance of the radar antenna. Typically, the beam pointing obtained from ground tests is used to approximate the beam pointing under on-orbit conditions, or the method of using satellite-borne measurement equipment is employed for on-orbit beam measurement.

[0005] For example, the invention patent with publication number CN108152787B discloses a method for accurately obtaining the beam pointing of a satellite radar antenna, including the following: (1) The satellite is equipped with a microwave transmitting antenna, a laser transmitter, and a star sensor, wherein the transmitting antenna and the laser transmitter have a common reference reference, and the transmitting antenna is located in front of the satellite radar antenna and points towards the radar antenna; (2) The transmitting antenna transmits microwaves to the radar antenna, the radar antenna receives and obtains the microwave beam direction in the coordinate system of the transmitting antenna, and inverts to obtain the beam pointing of the radar antenna when transmitting microwaves in the coordinate system of the transmitting antenna; (3) At the same time as (2), the laser transmitter transmits laser light into the star sensor, and the star sensor obtains the beam pointing and the position of the laser transmitter in the coordinate system of the star sensor. This patent belongs to the method of obtaining beam pointing by on-orbit measurement, which is fundamentally different from the method of this invention, which uses ground test data and on-orbit operation to obtain thermal load and beam pointing during external calibration, and then uses a multi-layer neural network method to predict the beam pointing of the satellite when it is running in other positions.

[0006] Chinese patent application CN105526879A discloses an on-orbit measurement system and method for satellite large-array antenna deformation based on fiber optic gratings. The system includes a light generator, a fiber optic cable, multiple grating measurement points, a demodulator, and an information processor, all connected sequentially via a conductive fiber. The light generator comprises a main light source and a beam splitter, providing the light waves required for multiple conductive fibers. Multiple grating measurement points are arranged on the satellite large-array antenna surface to form a sensing network. The demodulator demodulates the collected light waves to obtain the strain and temperature of each grating measurement point. The information processor calculates the strain and temperature of each grating measurement point to obtain the antenna array surface deformation parameters. This patent primarily calculates the antenna array surface deformation parameters based on the strain and temperature of each grating measurement point, which is fundamentally different from the method of this invention, which uses ground test data and on-orbit external calibration to obtain thermal load and beam pointing, and then uses a multi-layer neural network to predict the beam pointing of the satellite at other locations.

[0007] Chinese patent application CN105444669B discloses a measurement system and method for measuring changes in the pointing of a large plane. The system includes a linear laser emitter, one-dimensional PSD measurement points, a measurement controller, and an information processor. The information processor is data-connected to the measurement controller, which is control-connected to the linear laser emitter. Multiple one-dimensional PSD measurement points are provided, each equipped with a one-dimensional PSD sensor. The light emitted by the linear laser emitter is transmitted to the one-dimensional PSD sensor, which is data-connected to the measurement controller. These multiple one-dimensional PSD measurement points are arranged on a large plane. The patent also provides a measurement method for the system. This patent primarily uses measured data from several measurement points to fit a plane, thereby obtaining the plane's changes. This method differs fundamentally from the present invention's method, which uses ground-based experimental data and data obtained during on-orbit calibration to acquire thermal load and beam pointing, and then uses a multi-layer neural network to predict the beam pointing of the satellite at other locations. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a method, system, and medium for predicting the on-orbit beam pointing of a radar antenna.

[0009] According to the present invention, a method, system, and medium for predicting the on-orbit beam pointing of a radar antenna are provided, the solution of which is as follows:

[0010] In a first aspect, a method for on-orbit beam pointing prediction of a radar antenna is provided, the method comprising:

[0011] Step S1: During the ground test, acquire ground test data including the satellite's test thermal load status, the phased array radar antenna array normal direction, and the star-sensor optical axis direction;

[0012] Step S2: Establish a multi-layer neural network based on the ground test data;

[0013] Step S3: After the satellite is launched into orbit, during the external calibration of the radar system using the ground calibration field, the on-orbit thermal load status of the satellite, the beam pointing information of the phased array radar antenna, and the star-sensor observation results at the time of external calibration are recorded simultaneously.

[0014] Step S4: Using the multilayer neural network obtained from the ground test results and the thermal load state obtained during the external calibration, predict the normal pointing of the phased array radar antenna array and the star-sensitive optical axis pointing during the external calibration.

[0015] Step S5: Use the phased array radar antenna beam pointing information, star-sensor observation results, and predicted phased array radar antenna array surface normal pointing and star-sensor optical axis pointing obtained during external calibration to correct the multilayer neural network;

[0016] Step S6: When the satellite moves to another location, the phased array radar antenna beam pointing at that location is predicted using a modified multilayer neural network based on the satellite's thermal load status.

[0017] Preferably, the direction of the radar antenna array normal is a description of the outer normal vector of the phased array radar antenna array in the whole satellite mechanical reference coordinate system, and the direction of the star sensor optical axis is a description of the angle between the star sensor optical axis vector and the mechanical reference coordinate system.

[0018] Preferably, the multilayer neural network established using ground test data takes the test thermal load state as input and the phased array radar antenna normal pointing and star-sensor optical axis pointing as output; during the initial external calibration of the radar system in orbit, the on-orbit thermal load state, phased array radar antenna beam pointing information and star-sensor observation results at the external calibration time are obtained.

[0019] Preferably, the deviation values ​​between the phased array radar antenna array surface normal direction and star-sensor optical axis direction predicted in step S4 and the radar antenna beam pointing information and star-sensor observation results are obtained respectively, and then superimposed onto the multi-layer neural network to correct the multi-layer neural network.

[0020] Secondly, a system for on-orbit beam pointing prediction of a radar antenna is provided, the system comprising:

[0021] Module M1: Acquires ground test data during ground tests, including the satellite's test thermal load status, the phased array radar antenna array normal direction, and the star-sensor optical axis direction.

[0022] Module M2: Establishes a multi-layer neural network based on the ground test data;

[0023] Module M3: After the satellite is launched into orbit, during the external calibration of the radar system using the ground calibration field, the on-orbit thermal load status of the satellite, the beam pointing information of the phased array radar antenna, and the star-sensitive observation results are recorded simultaneously at the time of external calibration.

[0024] Module M4: Utilizes the multilayer neural network obtained from ground test results and the thermal load state obtained during external calibration to predict the normal pointing of the phased array radar antenna surface and the star-sensitive optical axis pointing during external calibration;

[0025] Module M5: The multilayer neural network is corrected by using the phased array radar antenna beam pointing information, star-sensor observation results, and predicted phased array radar antenna array normal pointing and star-sensor optical axis pointing obtained during external calibration.

[0026] Module M6: When the satellite moves to another location, it uses a modified multilayer neural network to predict the beam pointing of the phased array radar antenna when the satellite is in that location, based on the satellite's thermal load status.

[0027] Preferably, the direction of the radar antenna array normal is a description of the outer normal vector of the phased array radar antenna array in the whole satellite mechanical reference coordinate system, and the direction of the star sensor optical axis is a description of the angle between the star sensor optical axis vector and the mechanical reference coordinate system.

[0028] Preferably, the multilayer neural network established using ground test data takes the test thermal load state as input and the phased array radar antenna normal pointing and star-sensor optical axis pointing as output; during the initial external calibration of the radar system in orbit, the on-orbit thermal load state, phased array radar antenna beam pointing information and star-sensor observation results at the external calibration time are obtained.

[0029] Preferably, the deviation values ​​between the phased array radar antenna array surface normal direction and star-sensor optical axis direction predicted in module M4 and the radar antenna beam pointing information and star-sensor observation results are obtained respectively, and then superimposed onto the multi-layer neural network to correct the multi-layer neural network.

[0030] Thirdly, a computer-readable storage medium is provided that stores a computer program, which, when executed by a processor, implements the steps of the method.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] This invention leverages the unique specificity of a physical satellite, meaning that once the satellite's physical state is established, its mechanical, electrical, and thermal properties are solidified as inherent characteristics. Under these conditions, ground-based tests are used to obtain the satellite's thermal load state, the phased array radar antenna array normal pointing test results, and the star sensor optical axis pointing test results. A multi-layer neural network is established, connecting "thermal load → phased array radar antenna array normal pointing + star sensor optical axis pointing." This comprehensively considers multiple coupling factors, including the uncertainty and discreteness of the performance parameters of the satellite's physical components and the complex connections between these components, thus forming a transmission relationship between the thermal load and "phased array radar antenna array normal pointing + star sensor optical axis pointing."

[0033] During the initial orbital period after satellite insertion, when the satellite conducts external calibration of the radar system using a ground calibration field, the on-orbit thermal load status is acquired through onboard temperature measurement equipment. Using the aforementioned multi-layer neural network established based on ground test data and the current on-orbit thermal load status, the pointing of the phased array radar antenna array normal and the pointing of the star sensor optical axis can be predicted. The predicted results are compared with the radar antenna beam pointing information and star sensor observation results obtained from external calibration to identify the deviations in the phased array radar antenna array normal and antenna beam pointing, and the deviations in the star sensor optical axis pointing from the actual observation results. Based on these deviations, the aforementioned multi-layer neural network established based on ground test data is corrected to obtain the transmission relationship of "thermal load → antenna beam pointing information + star sensor observation results".

[0034] During the long-term operation of subsequent satellites in orbit, their on-orbit thermal load states change compared to when they were operating in the calibration field, as they move to different locations. Using the aforementioned corrected transfer relationship, the beam pointing information of the phased array radar antenna at the current location can be predicted. In this way, the beam pointing of the phased array radar antenna can be directly predicted when the satellite is operating in a region other than the calibration field, providing predictable support for the actual beam pointing and imaging capabilities of the phased array radar antenna at different orbital positions during the long-term operation of radar satellites. Attached Figure Description

[0035] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0036] Figure 1 This is a flowchart of the present invention;

[0037] Figure 2 A schematic diagram of a multilayer neural network built based on ground test data;

[0038] Figure 3 A schematic diagram for predicting the orientation of the phased array radar antenna surface normal and the orientation of the star-sensitive optical axis. Detailed Implementation

[0039] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0040] The accuracy of beam pointing at a specific target is important for radar systems. However, as satellites operate in orbit for extended periods, the effects of the long-term, periodically changing orbital thermal environment on radar systems will gradually accumulate. This will cause significant deviations in the beam pointing correction obtained from external calibration when the radar system images areas other than the external calibration field, ultimately affecting the imaging performance of the radar system.

[0041] Considering the limitations of external calibration of satellite radar systems and the fact that the long-term changing orbital thermal environment is a key factor affecting the beam pointing of radar antennas during long-term on-orbit operation, this invention provides a method for predicting the on-orbit beam pointing of radar antennas, referring to... Figure 1 As shown, the method includes:

[0042] Step S1: Obtain the experimental thermal load state of the satellite during ground testing (denoted as T). 地面 ), the direction of the phased array radar antenna array normal (denoted as P) 地面-阵面-法线指向 ), Star-sensitive optical axis orientation (denoted as P) 地面-星敏-光轴指向 Ground-based experimental data, including those from [unclear - likely referring to a specific experiment or study].

[0043] Step S2: Based on ground test data, establish a multi-layer neural network consisting of "thermal load → phased array radar antenna array normal pointing + star-sensitive optical axis pointing", denoted as T. 地面 →P 地面-阵面-法线指向 P 地面-星敏-光轴指向 . Reference Figure 2 As shown, a multi-layer neural network model is established with thermal load as input and the normal direction of the phased array radar antenna surface and the direction of the star-sensitive optical axis as output. By learning from several sets of data obtained from ground tests, the intrinsic relationship between the thermal load applied during ground tests and the normal direction of the phased array radar antenna surface and the direction of the star-sensitive optical axis is obtained.

[0044] Step S3: During the initial orbit insertion phase, while conducting external calibration of the radar system using the ground calibration field, simultaneously record the satellite's on-orbit thermal payload status at that moment (denoted as T). 在轨D Radar antenna beam pointing information (denoted as P) 在轨-阵面D-波束指向 ), star-sensitive observation results (denoted as P) 在轨-星敏D-观测结果 ).

[0045] Step S4: Using the multilayer neural network obtained in step S2 and the on-orbit thermal load state obtained in step S3, refer to... Figure 3 As shown, based on the aforementioned intrinsic relationship between the applied thermal load and the normal pointing of the phased array radar antenna surface and the star-sensitive optical axis obtained from ground test data, the normal pointing of the radar antenna surface during external calibration (denoted as P) is predicted. 地面-阵面D-法线指向预测 ), Star-sensitive optical axis orientation (denoted as P) 地面-星敏D-光轴指向预测 ).

[0046] Step S5: Using the radar antenna beam pointing information, star-sensor observation results, and prediction results obtained in step S3, the aforementioned multilayer neural network is corrected.

[0047] △P 阵面 =P 在轨-阵面D-波束指向 -P 在轨-阵面D-法线指向预测

[0048] △P 星敏 =P 在轨-星敏D-观测结果 -P 在轨-星敏D-光轴指向预测

[0049] Step S6: When the satellite moves to another location, the radar antenna beam pointing at the current location is predicted using a modified multilayer neural network based on the current on-orbit thermal load status.

[0050] Specifically, the direction of the radar antenna array normal is a description of the outer normal vector of the phased array radar antenna array in the whole satellite mechanical reference coordinate system, and the direction of the star sensor optical axis is a description of the angle between the star sensor optical axis vector and the mechanical reference coordinate system.

[0051] The multilayer neural network established using ground test data takes the test thermal load state as input and the phased array radar antenna array normal direction and star-sensor optical axis direction as output. During the initial external calibration of the radar system in the early stage of orbit insertion, the on-orbit thermal load state, phased array radar antenna beam pointing information and star-sensor observation results at the time of external calibration are obtained.

[0052] By utilizing this multi-layer neural network and the radar system during the initial orbital calibration, the thermal load state of the phased array radar antenna surface normal and the star-sensitive optical axis direction under the on-orbit thermal load state during the external calibration period are predicted.

[0053] The phased array antenna beam pointing information can be obtained through external calibration of the radar system. At the same time as external calibration, star-sensor observations are performed to obtain the star-sensor observation results. The predicted radar antenna array normal pointing and star-sensor optical axis pointing are compared with the radar antenna beam pointing information and star-sensor observation results obtained from external calibration. The deviation values ​​of the two are superimposed on the multi-layer neural network to correct the aforementioned multi-layer neural network and obtain the transmission relationship of "thermal load → antenna beam pointing information + star-sensor observation results".

[0054] This invention leverages the unique nature of physical satellites; once the physical state of a satellite is established, its mechanical, electrical, and thermal properties are fixed as inherent characteristics. It simplifies coupling factors such as the uncertainty and discreteness of performance parameters of satellite components and the complexity of component connections, directly integrating these factors into a multi-layered network for comprehensive consideration. Using the thermal load state obtained from ground tests, the phased array radar antenna array normal pointing test results, and the star-sensitive optical axis pointing test results, it establishes the transmission relationship between the thermal load and the "phased array radar antenna array normal pointing + star-sensitive optical axis pointing".

[0055] During the initial orbital period after satellite insertion, when the satellite conducts external calibration of the radar system using a ground calibration field, the on-orbit thermal load status is acquired through onboard temperature measurement equipment. Using the aforementioned multi-layer neural network established based on ground test data and the current on-orbit thermal load status, the pointing of the phased array radar antenna array normal and the pointing of the star sensor optical axis can be predicted. The predicted results are compared with the radar antenna beam pointing information and star sensor observation results obtained from external calibration to identify the deviations in the phased array radar antenna array normal and antenna beam pointing, and the deviations in the star sensor optical axis pointing from the actual observation results. Based on these deviations, the aforementioned multi-layer neural network established based on ground test data is corrected to obtain the transmission relationship of "thermal load → antenna beam pointing information + star sensor observation results".

[0056] During the long-term operation of subsequent satellites in orbit, their on-orbit thermal load states change compared to when they were operating in the calibration field, as they move to different locations. Using the aforementioned corrected transfer relationship, the beam pointing information of the phased array radar antenna at the current location can be predicted. In this way, the beam pointing of the phased array radar antenna can be directly predicted when the satellite is operating in a region other than the calibration field, providing predictable support for the actual beam pointing and imaging capabilities of the phased array radar antenna at different orbital positions during the long-term operation of radar satellites.

[0057] This invention provides a method, system, and medium for predicting the on-orbit beam pointing of a radar antenna. It utilizes the thermal load state obtained from ground tests, star-sensor optical axis pointing test results, and phased array radar antenna array surface normal pointing test results to establish a multi-layer neural network representing "thermal load → phased array radar antenna array surface normal pointing + star-sensor optical axis pointing." Using this neural network and the on-orbit thermal load state during external calibration, the antenna array surface normal pointing and star-sensor optical axis pointing during external calibration of the radar system are predicted. The predicted results are compared with the beam pointing and star-sensor observation results obtained from external calibration. Based on the deviation between the two, the neural network is corrected to obtain the transmission relationship of "thermal load → antenna beam pointing information + star-sensor observation result prediction." During long-term on-orbit operation, when the satellite moves to a different location, its on-orbit thermal load state has changed compared to when it was operating in the calibration field. The aforementioned corrected multi-layer neural network can then be used to predict the phased array radar antenna beam pointing at the current location.

[0058] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0059] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for predicting the on-orbit beam pointing of a radar antenna, characterized in that, include: Step S1: During the ground test, acquire ground test data including the satellite's test thermal load status, the phased array radar antenna array normal direction, and the star-sensor optical axis direction; Step S2: Establish a multi-layer neural network based on the ground test data; Step S3: After the satellite is launched into orbit, during the external calibration of the radar system using the ground calibration field, the on-orbit thermal load status of the satellite, the beam pointing information of the phased array radar antenna, and the star-sensor observation results at the time of external calibration are recorded simultaneously. Step S4: Using the multilayer neural network obtained from the ground test results and the thermal load state obtained during the external calibration, predict the normal pointing of the phased array radar antenna array and the star-sensitive optical axis pointing during the external calibration. Step S5: Use the phased array radar antenna beam pointing information, star-sensor observation results, and predicted phased array radar antenna array surface normal pointing and star-sensor optical axis pointing obtained during external calibration to correct the multilayer neural network; Step S6: When the satellite moves to another location, based on the thermal load state of the satellite, the phased array radar antenna beam pointing at that location is predicted using a modified multilayer neural network. After obtaining thermal load and beam pointing using ground test data and during external calibration in orbit, the beam pointing of the satellite in other locations is predicted using a multi-layer neural network method. A multi-layer neural network was established by using ground tests to obtain the satellite's thermal load status, the phased array radar antenna array surface normal pointing test results, and the star sensor optical axis pointing test results. By comprehensively considering the uncertainty and discreteness of the performance parameters of the satellite's physical components, as well as the complex connection relationships between the physical components and multiple coupling factors, a transmission relationship is formed between the thermal load and the normal direction of the phased array radar antenna surface and the direction of the star-sensitive optical axis. In the initial stage of satellite insertion into orbit, when the satellite uses the ground calibration field to carry out external calibration of the radar system, the on-orbit thermal load status at this time is obtained through the on-board temperature measurement equipment. Then, by using a multi-layer neural network established based on ground test data and the on-orbit thermal load status at this time, the normal pointing of the phased array radar antenna surface and the star-sensitive optical axis pointing are predicted. The predicted results are compared with the radar antenna beam pointing information and star sensor observation results obtained by external calibration. The results are the deviations of the phased array radar antenna array normal pointing and antenna electronic beam pointing, and the deviations of the star sensor optical axis pointing and the actual observation results, respectively. Based on the obtained deviations, the multilayer neural network established based on the ground test data is corrected to obtain the transmission relationship between the thermal load and the phased array radar antenna array normal pointing and star sensor optical axis pointing. During the subsequent long-term operation of the satellite in orbit, when the satellite moves to other locations, its on-orbit thermal load state has changed compared to when it was operating in the calibration field; Using the modified transmission relationship obtained above, the beam pointing information of the phased array radar antenna at the current position is predicted.

2. The method for predicting the on-orbit beam pointing of a radar antenna according to claim 1, characterized in that, The direction of the radar antenna array normal is a description of the outer normal vector of the phased array radar antenna array in the whole satellite mechanical reference coordinate system, and the direction of the star sensor optical axis is a description of the angle between the star sensor optical axis vector and the mechanical reference coordinate system.

3. The method for predicting the on-orbit beam pointing of a radar antenna according to claim 1, characterized in that, The multilayer neural network established using ground test data takes the test thermal load state as input and the phased array radar antenna array normal direction and star-sensor optical axis direction as output. During the initial external calibration of the radar system in orbit, the on-orbit thermal load state, phased array radar antenna beam pointing information and star-sensor observation results are obtained.

4. A system for predicting the on-orbit beam pointing of a radar antenna, characterized in that, include: Module M1: Acquires ground test data during ground tests, including the satellite's test thermal load status, the phased array radar antenna array normal direction, and the star-sensor optical axis direction. Module M2: Establishes a multi-layer neural network based on the ground test data; Module M3: After the satellite is launched into orbit, during the external calibration of the radar system using the ground calibration field, the on-orbit thermal load status of the satellite, the beam pointing information of the phased array radar antenna, and the star-sensitive observation results are recorded simultaneously at the time of external calibration. Module M4: Utilizes the multilayer neural network obtained from ground test results and the thermal load state obtained during external calibration to predict the normal pointing of the phased array radar antenna surface and the star-sensitive optical axis pointing during external calibration; Module M5: The multilayer neural network is corrected by using the phased array radar antenna beam pointing information, star-sensor observation results, and predicted phased array radar antenna array normal pointing and star-sensor optical axis pointing obtained during external calibration. Module M6: When the satellite moves to another location, based on the satellite's thermal load status, a modified multilayer neural network is used to predict the beam pointing of the phased array radar antenna when the satellite is in that location; After obtaining thermal load and beam pointing using ground test data and during external calibration in orbit, the beam pointing of the satellite in other locations is predicted using a multi-layer neural network method. A multi-layer neural network was established by using ground tests to obtain the satellite's thermal load status, the phased array radar antenna array surface normal pointing test results, and the star sensor optical axis pointing test results. By comprehensively considering the uncertainty and discreteness of the performance parameters of the satellite's physical components, as well as the complex connection relationships between the physical components and multiple coupling factors, a transmission relationship is formed between the thermal load and the normal direction of the phased array radar antenna surface and the direction of the star-sensitive optical axis. In the initial stage of satellite insertion into orbit, when the satellite uses the ground calibration field to carry out external calibration of the radar system, the on-orbit thermal load status at this time is obtained through the on-board temperature measurement equipment. Then, by using a multi-layer neural network established based on ground test data and the on-orbit thermal load status at this time, the normal pointing of the phased array radar antenna surface and the star-sensitive optical axis pointing are predicted. The predicted results are compared with the radar antenna beam pointing information and star sensor observation results obtained by external calibration. The results are the deviations of the phased array radar antenna array normal pointing and antenna electronic beam pointing, and the deviations of the star sensor optical axis pointing and the actual observation results, respectively. Based on the obtained deviations, the multilayer neural network established based on the ground test data is corrected to obtain the transmission relationship between the thermal load and the phased array radar antenna array normal pointing and star sensor optical axis pointing. During the subsequent long-term operation of the satellite in orbit, when the satellite moves to other locations, its on-orbit thermal load state has changed compared to when it was operating in the calibration field; Using the modified transmission relationship obtained above, the beam pointing information of the phased array radar antenna at the current position is predicted.

5. The system for predicting the on-orbit beam pointing of a radar antenna according to claim 4, characterized in that, The direction of the radar antenna array normal is a description of the outer normal vector of the phased array radar antenna array in the whole satellite mechanical reference coordinate system, and the direction of the star sensor optical axis is a description of the angle between the star sensor optical axis vector and the mechanical reference coordinate system.

6. The system for predicting the on-orbit beam pointing of a radar antenna according to claim 4, characterized in that, The multilayer neural network established using ground test data takes the test thermal load state as input and the phased array radar antenna normal pointing and star-sensor optical axis pointing as output. During the initial external calibration of the radar system in the early stage of orbit insertion, the on-orbit thermal load state, phased array radar antenna beam pointing information and star-sensor observation results at the time of external calibration are obtained.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 3.

Images