Feed compensation method, system and product of high-cross-pole-ratio circularly polarized antenna array

Abstract

The invention relates to a feed compensation method, a feed compensation system and a feed compensation product for a high-cross-pole-ratio circularly polarized antenna array in the field of phased-array antennae, and the method comprises the following steps: S1, a feed compensation model training step and S2, a feed signal control step; the method comprises the following steps: S1, in the feed compensation model training step, constructing a training sample library through full-wave electromagnetic simulation software according to a corresponding relation between working parameters and ideal compensation values, dividing samples into a training set and a verification set, and training to obtain a feed compensation model; and S2, in the feed signal control step, a final feed compensation model is deployed in the antenna array control system, a predicted value of an ideal compensation value is obtained after working parameters are obtained, and each antenna unit is adjusted based on the predicted value of the ideal compensation value. According to the antenna array, self-adaptive feed compensation can be realized for the feed signal in the digital domain according to the working parameters of the antenna array, so that the high cross-pole ratio of the antenna under large-angle scanning is realized.

Description

Technical Field

[0001] This application relates to a feeding compensation method, system, and product for high cross-ratio circularly polarized antenna arrays used in the field of phased array antennas. Background Technology

[0002] Phased array antennas form beamforming networks by arranging antenna elements to provide the phase gradient required for beam deflection. In free space, an independent antenna element, completely independent of other elements, has an isolated impedance measured at its feed port, which is the primary design objective for impedance matching. However, in the array environment of a phased array antenna, all elements are simultaneously excited, and the active impedance of an independent antenna element at its feed port is altered by the coupling between these elements. Especially for phased array antennas, as the scan angle changes, the phase relationship between the antenna elements also changes, leading to a corresponding change in the mutual coupling effect between them. This effect superimposes at the feed port, directly causing the active impedance to change with the scan angle. For circularly polarized antennas, during large-angle scans, the resulting mismatch between the active impedance and the feed network leads to a significant decrease in cross-polarization ratio, a decrease in polarization purity, ellipticization of the antenna pattern, and reduced antenna performance.

[0003] Currently, antenna designers primarily achieve high cross-pole ratio array antennas through structural innovations in antenna elements and feed networks, or through the arrangement design of the antenna array. However, complex antenna structures increase the difficulty of antenna manufacturing processes, reduce product yield, and increase production costs. Furthermore, some array arrangement designs, such as sparse arrays, lead to decreased space utilization, making it difficult to achieve antenna miniaturization while meeting gain requirements. Moreover, these solutions inevitably result in a significant decrease in cross-pole ratio at large scanning angles. Summary of the Invention

[0004] The purpose of this application is to overcome the shortcomings of the prior art and provide a feeding compensation method, system and product for a high cross-pole ratio circularly polarized antenna array, which can achieve adaptive feeding compensation of the feeding signal in the digital domain according to the operating parameters of the antenna array, thereby achieving a high cross-pole ratio of the antenna under large-angle scanning.

[0005] Firstly, this application provides a feeding compensation method for a high cross-polarity circularly polarized antenna array, wherein the antenna array is a phased array antenna composed of an antenna element array, and adopts the following technical solution: This includes the S1 power supply compensation model training steps and the S2 power supply signal control steps; The training steps for the S1 power supply compensation model specifically include the following steps: S101. For the target antenna array, different operating parameters are preset, and simulation experiments are conducted using full-wave electromagnetic simulation software to obtain the ideal compensation value of each antenna element in the state of the highest cross-pole ratio of the antenna pattern when the operating parameters are constant. A training sample library is constructed based on the correspondence between the operating parameters and the ideal compensation value, and the samples are divided into training set and validation set. S102, construct neural networks with encoder-decoder LSTM architectures with different hyperparameters. Use the training set as input to the working parameters and the prediction results of the ideal compensation value as output to train each neural network and obtain different power supply compensation models. S103, the power supply compensation model trained under different hyperparameters is validated through the validation set, and the power supply compensation model with the smallest deviation between the predicted result of the ideal compensation value and the ideal compensation value is selected. S104. Randomly set the working parameters and input them into the power supply compensation model and the full-wave electromagnetic simulation software respectively. Compare the predicted results of the ideal compensation value with the ideal compensation value. If the deviation is lower than the threshold, it is confirmed as the final power supply compensation model. The S2 power supply signal control steps specifically include the following steps: S201, Deploy the final feed compensation model in the antenna array control system; S202, After the antenna array control system acquires the operating parameters, it synchronously inputs the final feed compensation model to obtain the predicted value of the ideal compensation value; S203, the antenna array control system adjusts each antenna element based on the predicted value of the ideal compensation value.

[0006] By adopting the above technical solution, this application constructs a training sample library based on the correspondence between the operating parameters of the antenna array and the ideal compensation values ​​of each antenna element. It then trains a feed compensation model based on an encoder-decoder LSTM neural network, thereby automatically obtaining the predicted compensation values ​​of the feed signals of each antenna element under different operating conditions based on the input operating parameters, achieving adaptive dynamic compensation of the feed signals. Using these predicted compensation values ​​for amplitude and phase compensation of the feed signals of each antenna element can achieve a high crossover ratio at all scanning angles.

[0007] Preferably, in S101, the ideal compensation values ​​of each antenna element in the state of maximum cross-pole ratio of the antenna pattern when the operating parameters are constant are obtained using a genetic algorithm, specifically including: S1011, Construct an antenna array model composed of antenna elements in full-wave electromagnetic simulation software; S1012, Set the operating frequency of the antenna array in the current simulation test; S1013, Set the normal direction of the antenna array as the initial scanning angle for the simulation experiment; S1014, with the compensation value of the antenna element being 0 as the initial compensation value, an initial population of individuals with randomly distributed compensation values ​​for each antenna element is constructed based on the initial compensation value. S1015, using the current population individuals to conduct simulation experiments at the working frequency and scanning angle, to obtain the antenna array cross-pole ratio corresponding to the current population individuals; S1016, select the compensation values ​​of each antenna element in several groups with the highest cross-pole ratio as parent individuals; cross-recombine the compensation values ​​of antenna elements corresponding to different parent individuals to form cross individuals; randomly select and mutate the compensation values ​​of each antenna element in the parent individuals to form mutated individuals. S1017: Generate a new population through parent individuals, crossover individuals and mutated individuals, and repeat S1015-S1016 until the optimal solution for an individual is obtained. The optimal solution for an individual is the ideal compensation value for the current working frequency and scanning angle. S1018, adjust the scanning angle to the next angle setting value, repeat S1013-S1017 to obtain the ideal compensation value of each preset angle at the current working frequency; S1019, adjust the working frequency to the next frequency setting value, repeat S1012-S1018 to obtain the ideal compensation value for each preset angle at each working frequency, thereby constructing a training sample library.

[0008] By adopting the above technical solution, the antenna array is simulated under preset scanning angle and operating frequency conditions using full-wave electromagnetic simulation software. Different radiation pattern performances are obtained under different compensation values ​​for each antenna element. With the highest cross-pole ratio as the goal, the ideal compensation values ​​for different preset scanning angles and operating frequencies are obtained through genetic algorithm to construct a training sample library.

[0009] Preferably, the antenna array is a one-dimensional phased array antenna. In S1018, the scanning angle is the azimuth angle, and the angle setting value is a number of azimuth angles offset to one side from the normal direction of the antenna array.

[0010] Preferably, the antenna array is a two-dimensional phased array antenna. In S1018, the scanning angle is a combination of azimuth and elevation angles. The angle setting value is a combination of several preset azimuth and elevation angles offset from the normal direction of the antenna array in one direction of change of azimuth and one direction of change of elevation angle.

[0011] By adopting the above technical solution, one-dimensional phased array antennas and two-dimensional phased array antennas are distinguished, and a specific method for setting the scanning angle in antenna array simulation experiments is provided. Simultaneously, the solution utilizes the symmetry of the antenna to reduce the number of angle setting values, thereby significantly reducing the computational load of the algorithm. For a one-dimensional antenna, the scanning angle is the azimuth angle, and the angle setting value is a number of set angles offset from the normal direction towards the positive (or negative) azimuth angle. For a two-dimensional antenna, the scanning angle is a combination of azimuth and elevation angles, and the angle setting value is a combination of azimuth and elevation angles offset from the normal direction towards the positive (or negative) azimuth angle and towards the positive (or negative) elevation angle.

[0012] Preferably, the scanning angle setting value gradually increases from the normal direction of the antenna array. For two adjacent angle settings, the initial compensation value of the later angle setting value is the ideal compensation value of the earlier angle setting value.

[0013] By adopting the above technical solution, the scanning angle of the simulation experiment gradually increases from the normal direction and the experiment is carried out in sequence. The ideal compensation value of the adjacent previous angle setting value is used as the starting point of the genetic algorithm for the later angle setting value, which can greatly reduce the computational load of the genetic algorithm.

[0014] Preferably, the density of the distribution of the angle setting values ​​increases accordingly with the increase of the scanning angle.

[0015] By adopting the above technical solution, when the active impedance of the antenna element changes more drastically with the scanning angle at large scanning angles, increasing the density of angle setting values ​​within the large scanning angle range can provide greater data support for the large angle range, thereby enabling the construction of more complex feed compensation models.

[0016] Preferably, the ideal compensation value is the compensation value for the real-time amplitude and phase of the feed signal of each antenna element.

[0017] By adopting the above technical solution, amplitude and phase adjustment compensation of the feed signal of each antenna element can be achieved.

[0018] As a preferred option, in S102, the specific method for constructing a neural network with an encoder-decoder LSTM architecture to obtain the power supply compensation model is as follows: S1021, Set the hyperparameters of the encoder-decoder LSTM architecture neural network, including setting model structure parameters and training process parameters; S1022, using the change of the working parameters as a time step, the training sample library is serialized according to the order of the change of the working parameters; S1023 employs an encoder consisting of one or more LSTM layers to read the input sequence of working parameters, update the hidden state according to the input at each time step, and output the final hidden state encoded as a context vector. S1024 employs a decoder consisting of one or more LSTM layers. The input context vector is used as the initial hidden state. Based on the input at each time step, the output vector is obtained. The output vector is then passed through a fully connected layer to generate a sequence of predicted values ​​for the ideal compensation values ​​step by step. S1025 compares the predicted value of the ideal compensation value with the ideal compensation value item by item using the loss function, calculates the total loss value of the ideal compensation value, and feeds the total loss value back to the decoder and encoder through the backpropagation algorithm for parameter update.

[0019] By adopting the above technical solution, the neural network of the encoder-decoder LSTM architecture can be trained to obtain a feed compensation model based on the training sample library with the preset frequency values ​​and scanning angles. The input of the feed compensation model is the working parameters of the antenna array, and the output of the antenna array is the predicted value of the ideal compensation value of each antenna element.

[0020] Secondly, this application provides a feed compensation system for a high cross-pole ratio circularly polarized antenna array, which applies the above-mentioned feed compensation method and adopts the following technical solution: Includes an antenna array control module, a feed compensation prediction module, a feed adjustment module, and a phased array antenna; The antenna array control module acquires the operating parameters of the antenna array, outputs the operating parameters to the feed compensation prediction module, acquires the predicted value of the ideal compensation value fed back by the feed compensation prediction module, superimposes the predicted value of the ideal compensation value with the scanning feed signal, and outputs the superimposed signal to the feed adjustment module. The feed compensation prediction module is equipped with a final feed compensation model. Based on the input operating parameters, it obtains the predicted value of the ideal compensation value and returns it to the antenna array control module. The antenna array is a phased array antenna, which is composed of an array of antenna elements. Each antenna element has a feed signal inlet, which is connected to the feed adjustment module. The feed adjustment module adjusts the feed signal transmitted to each antenna element based on the superimposed signal.

[0021] Thirdly, the computer program product provided in this application employs a technical solution including a computer program or instructions, which enables the computer program or instructions to implement the above-mentioned feeding compensation method for a high cross-pole ratio circularly polarized antenna array.

[0022] In summary, this application includes at least one of the following beneficial technical effects: This application enables adaptive compensation of antenna element feed signals at different frequencies and scanning angles in a phased array antenna, optimizes the radiation pattern of the antenna array, improves the signal transmission quality of the antenna array, and achieves flexible beam control.

[0023] This application facilitates the matching of the active impedance of each antenna element with the impedance of the feed network at large scanning angles, achieves a higher cross-polarization ratio at large angles, improves circular polarization purity, thereby obtaining a larger scanning angle and reducing interference between orthogonal polarization channels.

[0024] 3. This application constructs a compensation prediction model to achieve intelligent online compensation of the feed signal in the digital domain, which can improve the control accuracy and control flexibility of the antenna system. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a feeding compensation method for a high cross-pole ratio circularly polarized antenna array according to an embodiment of this application. Figure 2 This is a flowchart illustrating S1 in an embodiment of this application; Figure 3 This is a flowchart illustrating S101 in an embodiment of this application; Figure 4 This is a flowchart illustrating S102 in an embodiment of this application; Figure 5 This is a flowchart illustrating S2 in an embodiment of this application; Figure 6 This is a schematic diagram of the architecture of a feeding compensation system for a high cross-pole ratio circularly polarized antenna array according to an embodiment of this application; Figure 7 This is a schematic diagram of the architecture of an exemplary computer device in an embodiment of this application. Detailed Implementation

[0026] This specific embodiment is merely an explanation of this application and is not intended to limit it. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they are within the scope of this application.

[0027] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that in the optional embodiments of this application, the object information and other related data involved require the permission or consent of the object when the embodiments of this application are applied to specific products or technologies, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. That is to say, if the embodiments of this application involve data related to the object, it needs to be obtained with the authorization and consent of the object, the authorization and consent of the relevant departments, and in compliance with the relevant laws, regulations, and standards of the country and region. If personal information is involved in the embodiments, the acquisition of all personal information requires the consent of the individual. If sensitive information is involved, the separate consent of the information subject is required. The embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0028] This application aims to train and deploy a feed compensation model so that, by inputting the operating parameters of a phased array antenna array, the predicted value of the ideal compensation value of the feed signal of each antenna element can be automatically obtained. This allows the feed signal of the antenna array to be superimposed with the predicted value of the ideal compensation value to obtain a superimposed feed signal. Based on this superimposed feed signal, each antenna element can optimize its antenna pattern, thereby achieving a high cross-polarity ratio at a large scanning angle, and thus achieving high circular polarization purity and effective channel isolation.

[0029] Please see Figure 1 The present application provides a method for feeding compensation of a high cross-pole ratio circularly polarized antenna array, which includes an S1 feeding compensation model training step and an S2 feeding signal control step.

[0030] Please see Figure 2 The training steps for the S1 power supply compensation model include the following steps: S101. For the target antenna array, different operating parameters are preset, and simulation experiments are conducted using full-wave electromagnetic simulation software to obtain the ideal compensation value of each antenna element in the state of the highest cross-pole ratio of the antenna pattern when the operating parameters are constant. A training sample library is constructed based on the correspondence between the operating parameters and the ideal compensation value, and the samples are divided into training set and validation set.

[0031] The operating parameters refer to the operating frequency and scanning angle of the antenna array, while the compensation value for each antenna element refers to the feed compensation for the amplitude and phase of the feed signal of each antenna element. It should be noted that for a specific operating parameter, the feed compensation of each antenna element in the antenna array is not exactly the same, and the collection of the feed compensations of all antenna elements constitutes the compensation value.

[0032] For more details, please see Figure 3 In S101, a genetic algorithm is used to obtain the ideal compensation values ​​of each antenna element in the state of maximum cross-pole ratio of the antenna pattern when the operating parameters are constant. Specifically, this includes: S1011, construct an antenna array model composed of antenna elements in a full-wave electromagnetic simulation software.

[0033] S1012, Set the operating frequency of the antenna array in the current simulation test.

[0034] The operating frequency needs to match the design operating frequency of the antenna array model. For example, for a KA band antenna, the operating frequency range is 26.5-40 GHz.

[0035] S1013, set the normal direction of the antenna array as the initial scanning angle for the simulation experiment.

[0036] S1014, with the compensation value of the antenna element being 0 as the initial compensation value, an initial population of individuals with randomly distributed compensation values ​​for each antenna element is constructed based on the initial compensation value.

[0037] The specific method for obtaining the initial population is as follows: based on the initial compensation value and within a preset threshold range, a Gaussian distribution strategy is used to randomly generate random individuals with amplitude and phase for each antenna unit, and these random individuals are used to obtain the initial population. The population size is defined according to the antenna element size of the antenna array.

[0038] S1015, using the current population individuals to conduct simulation experiments at the operating frequency and scanning angle, to obtain the antenna array cross-pole ratio corresponding to the current population individuals.

[0039] S1016: Select the compensation values ​​of several groups of antenna elements with the highest cross-pole ratio as parent individuals. Cross-recombine the compensation values ​​of antenna elements corresponding to different parent individuals to form cross individuals. Randomly select and mutate the compensation values ​​of each antenna element in the parent individuals to form mutated individuals.

[0040] In the embodiments of this application, the crossover rate Pc = 0.85 and the mutation rate Pm = 0.04.

[0041] In step S1017, a new population is generated through parent individuals, crossover individuals, and mutated individuals, and steps S1015-S1016 are repeated. The genetic algorithm terminates when the antenna array cross-pole ratio obtained from the individual's optimal solution reaches a threshold, or when the number of iterations reaches the maximum preset value G=300. The individual's optimal solution represents the ideal compensation value for the current operating frequency and scanning angle.

[0042] S1018: After completing the construction of the training sample library for the current scanning angle, adjust the scanning angle to the next angle setting value, and repeat S1013-S1017 to obtain the ideal compensation value of each preset angle under the current working frequency.

[0043] Antenna arrays are divided into one-dimensional phased array antennas and two-dimensional phased array antennas. For a one-dimensional phased array antenna, its scanning angle is the azimuth angle, while for a two-dimensional phased array antenna, its scanning angle is a combination of the azimuth and elevation angles. To fully utilize the antenna's symmetry, avoid redundant calculations, and improve algorithm efficiency, the ideal compensation value for the scanning angle on one side of the symmetrical direction is calculated only. The ideal compensation value on the other side can be obtained by using equal amplitude and the same or opposite phase.

[0044] For example, in the embodiments of this application, for a one-dimensional phased array antenna, the angle setting value is a number of azimuth angles offset from the normal direction of the antenna array in the positive direction; for a two-dimensional phased array antenna, the angle setting value is a combination of a number of preset azimuth angles and elevation angles that vary from the normal direction of the antenna array in the positive direction of the azimuth angle and the positive direction of the elevation angle.

[0045] The scanning angle setting value gradually increases from the normal direction of the antenna array. For two adjacent angle settings, the initial compensation value of the later angle setting value is the ideal compensation value of the earlier angle setting value. Since the ideal compensation values ​​of two adjacent angle settings are relatively close, the above method can be used to establish a new initial population as the starting point for each genetic algorithm, which can further reduce the computational load of the algorithm.

[0046] It should also be noted that the change in active impedance of the antenna element is more drastic at large scanning angles compared to small scanning angles. The prediction of the feed compensation model at large scanning angles is more complex, and the ideal compensation value that needs to be adjusted is more precise. Therefore, the amount of data samples required at large scanning angles is also larger. As a result, the density of the distribution of angle setting values ​​increases accordingly with the increase of scanning angle.

[0047] The following explanation uses a one-dimensional phased array antenna as an example. When the scanning angle is 0°-45°, the interval of the angle setting value can be 3°-5°. However, when the scanning angle is 45°-60° (60° is the maximum scanning angle of the antenna array design in this embodiment), the interval of the angle setting value needs to be significantly increased, and can be set to 0.5°-1°.

[0048] S1019: After completing the construction of the training sample library for the current operating frequency, adjust the operating frequency to the next frequency setting value, and repeat S1012-S1018 to obtain the ideal compensation values ​​for each preset angle at each operating frequency, thereby constructing a training sample library for the entire domain. The samples are then divided into a training set and a validation set.

[0049] S102, construct neural networks with encoder-decoder LSTM architectures of different hyperparameters. Use the training set as input to the working parameters and the prediction results of the ideal compensation value as output to train each neural network and obtain different power supply compensation models.

[0050] Please see Figure 4 The specific method for constructing an encoder-decoder LSTM neural network to obtain a power supply compensation model includes the following steps.

[0051] S1021, Set the hyperparameters of the encoder-decoder LSTM neural network, including setting model structure parameters and training process parameters. Specifically, model structure parameters include hidden layer dimension, number of network layers, etc.; training process hyperparameters include learning rate, number of iterations, etc. These hyperparameters are used to control the training process of the encoder-decoder LSTM neural network to prevent underfitting or overfitting.

[0052] S1022, using the change in operating parameters as a time step, the training sample library is serialized according to the order of the change in operating parameters. Specifically, starting from the normal scan direction with the lowest operating frequency, the scan angle is increased step by step until the maximum scan angle is reached, then the operating frequency is increased and the scan direction is returned to the normal scan direction.

[0053] S1023 employs an encoder consisting of one or more LSTM layers to read the input sequence of working parameters, update the hidden state according to the input at each time step, and output the final hidden state encoded as a context vector.

[0054] Let's take a one-dimensional phased array antenna as an example for specific explanation. The input of each time step sample is (f, θ), where f is the operating frequency and θ is the scanning angle. The sample parameters are standardized using Z-score to form features of a uniform dimension. These features can then be projected into a higher-dimensional, more expressive vector space through an embedding layer or a fully connected layer. This step helps the model capture the deep nonlinear relationships between these parameters. Each time step forms a sequence of a preset length.

[0055] The processed sequence is input into the encoder's LSTM layer, and the hidden state is updated through the LSTM layer. When the LSTM layer has processed the last element of the input sequence, its final hidden state is the context vector output by the encoder.

[0056] S1024 employs a decoder consisting of one or more LSTM layers. It takes the context vector as the initial hidden state and obtains the output vector based on the input at each time step. The output vector is then passed through a fully connected layer to generate a sequence of predicted values ​​for the ideal compensation values ​​step by step.

[0057] Following the above embodiment, the decoder's input at the first time step is... <start>A marker is used to indicate the start of the sequence. Then, for each time step, the decoder's LSTM layer predicts the next prediction in the output sequence based on the acquired hidden state. The prediction is then mapped to the output dimension of the ideal compensation value through a fully connected layer.

[0058] S1025 compares the predicted value of the ideal compensation value with the ideal compensation value item by item using the loss function, calculates the total loss value of the ideal compensation value, and feeds the total loss value back to the decoder and encoder through the backpropagation algorithm for parameter update.

[0059] Following the above embodiment, after the encoder-decoder LSTM model completes forward propagation, it generates a sequence of predicted standard compensation values ​​for each antenna element. The standard compensation values ​​in the training set also have the same dimensional structure. The loss function compares these values ​​point-by-point at their corresponding positions, specifically comparing the amplitude error and the phase error. After obtaining the amplitude error and phase error, they are weighted and summed to obtain the total loss value: Total loss = α × amplitude error + β × phase error; In the formula, α and β are hyperparameters used to balance the contribution of the two types of errors to the total loss.

[0060] After training the entire training set, power supply compensation models with different hyperparameters were obtained.

[0061] S103, the power supply compensation model trained under different hyperparameters is validated using a validation set, and the power supply compensation model with the smallest deviation between the predicted result and the ideal compensation value is selected.

[0062] S104. Randomly set the working parameters as the test set, and input the working parameters into the power supply compensation model and the full-wave electromagnetic simulation software respectively. Compare the predicted result of the ideal compensation value output by the power supply compensation model with the ideal compensation value calculated by the full-wave electromagnetic simulation software through the genetic algorithm. If the deviation between the two is lower than the threshold, it is confirmed as the final power supply compensation model.

[0063] This completes the construction of the final power supply compensation model.

[0064] Please see Figure 5 The S2 power supply signal control steps specifically include the following steps: S201, deploying the final feed compensation model in the antenna array control system.

[0065] S202, after the antenna array control system acquires the operating parameters, it synchronously inputs the final feed compensation model to obtain the predicted value of the ideal compensation value.

[0066] S203, the antenna array control system, based on the predicted value of the ideal compensation value, superimposes the predicted value of the ideal compensation value with the antenna element feed signal to achieve adjustment of each antenna element.

[0067] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0068] Please see Figure 6 An embodiment of this application provides a power supply compensation system for a high cross-pole ratio circularly polarized antenna array, comprising an antenna array control module 1, a power supply compensation prediction module 2, a power supply adjustment module 3, and a phased array antenna 4.

[0069] Antenna array control module 1 acquires the operating parameters of the antenna array, outputs the operating parameters to feed compensation prediction module 2, acquires the predicted value of the ideal compensation value fed back by feed compensation prediction module 2, superimposes the predicted value of the ideal compensation value with the scanning feed signal, and outputs the superimposed signal to feed adjustment module 3.

[0070] The feed compensation prediction module 2 is equipped with the final feed compensation model. Based on the input working parameters, it obtains the predicted value of the ideal compensation value and returns it to the antenna array control module 1.

[0071] Antenna array 4 is a phased array antenna, which is composed of an array of antenna elements. Each antenna element has a feed signal inlet, which is connected to the feed adjustment module 3. The feed adjustment module 3 adjusts the feed signal transmitted to each antenna element based on the superimposed signal.

[0072] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the feed compensation system for a high cross-pole ratio circularly polarized antenna array described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0073] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0074] Please see Figure 7 This application provides an exemplary computer device. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores predicted values ​​of the compensation values ​​output by the final feed compensation model. The network interface connects to an antenna element feed signal modulation device. When executed by the processor, the computer program implements a feed compensation method for a high cross-pole ratio circularly polarized antenna array.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.< / start>

Claims

1. A feeding compensation method for a high cross-polarity circularly polarized antenna array, wherein the antenna array is a phased array antenna composed of an array of antenna elements, characterized in that... This includes the S1 power supply compensation model training steps and the S2 power supply signal control steps; The training steps for the S1 power supply compensation model include the following steps: S101. For the target antenna array, different operating parameters are preset, and simulation experiments are conducted using full-wave electromagnetic simulation software to obtain the ideal compensation value of each antenna element in the state of the highest cross-pole ratio of the antenna pattern when the operating parameters are constant. A training sample library is constructed based on the correspondence between the operating parameters and the ideal compensation value, and the samples are divided into training set and validation set. S102, construct neural networks with encoder-decoder LSTM architectures with different hyperparameters. Use the training set as input to the working parameters and the prediction results of the ideal compensation value as output to train each neural network and obtain different power supply compensation models. S103, the power supply compensation model trained under different hyperparameters is validated through the validation set, and the power supply compensation model with the smallest deviation between the predicted result of the ideal compensation value and the ideal compensation value is selected. S104. Randomly set the working parameters and input them into the power supply compensation model and the full-wave electromagnetic simulation software respectively. Compare the predicted results of the ideal compensation value with the ideal compensation value. If the deviation is lower than the threshold, it is confirmed as the final power supply compensation model. The S2 power supply signal control steps specifically include the following steps: S201, Deploy the final feed compensation model in the antenna array control system; S202, After the antenna array control system acquires the operating parameters, it synchronously inputs the final feed compensation model to obtain the predicted value of the ideal compensation value; S203, the antenna array control system adjusts each antenna element based on the predicted value of the ideal compensation value.

2. The feeding compensation method for a high cross-ratio circularly polarized antenna array according to claim 1, characterized in that, In S101, the ideal compensation values ​​of each antenna element in the state of maximum cross-pole ratio of the antenna pattern when the operating parameters are constant are obtained using a genetic algorithm, specifically including: S1011, Construct an antenna array model composed of antenna elements in full-wave electromagnetic simulation software; S1012, Set the operating frequency of the antenna array in the current simulation test; S1013, Set the normal direction of the antenna array as the initial scanning angle for the simulation experiment; S1014, with the compensation value of the antenna element being 0 as the initial compensation value, an initial population of individuals with randomly distributed compensation values ​​for each antenna element is constructed based on the initial compensation value. S1015, using the current population individuals to conduct simulation experiments at the working frequency and scanning angle, to obtain the antenna array cross-pole ratio corresponding to the current population individuals; S1016, select the compensation values ​​of each antenna element in several groups with the highest cross-pole ratio as parent individuals; cross-recombine the compensation values ​​of antenna elements corresponding to different parent individuals to form cross individuals; randomly select and mutate the compensation values ​​of each antenna element in the parent individuals to form mutated individuals. S1017: Generate a new population through parent individuals, crossover individuals and mutated individuals, and repeat S1015-S1016 until the optimal solution for an individual is obtained. The optimal solution for an individual is the ideal compensation value for the current working frequency and scanning angle. S1018, adjust the scanning angle to the next angle setting value, repeat S1013-S1017 to obtain the ideal compensation value of each preset angle at the current working frequency; S1019, adjust the working frequency to the next frequency setting value, repeat S1012-S1018 to obtain the ideal compensation value for each preset angle at each working frequency, thereby constructing a training sample library.

3. The feeding compensation method for a high cross-ratio circularly polarized antenna array according to claim 2, characterized in that, The antenna array is a one-dimensional phased array antenna. In S1018, the scanning angle is the azimuth angle, and the angle setting value is a number of azimuth angles offset to one side from the normal direction of the antenna array.

4. The feeding compensation method for a high cross-ratio circularly polarized antenna array according to claim 2, characterized in that, The antenna array is a two-dimensional phased array antenna. In S1018, the scanning angle is a combination of azimuth and elevation angles. The angle setting value is a combination of several preset azimuth and elevation angles from the normal direction of the antenna array to a change direction of the azimuth angle and a change direction of the elevation angle.

5. A feeding compensation method for a high cross-polarity circularly polarized antenna array according to claim 3 or 4, characterized in that, The scanning angle setting value gradually increases from the normal direction of the antenna array. For two adjacent angle settings, the initial compensation value of the later angle setting value is the ideal compensation value of the earlier angle setting value.

6. A feeding compensation method for a high cross-pole ratio circularly polarized antenna array according to claim 3 or 4, characterized in that, The density of the angle setting values ​​increases accordingly as the scanning angle increases.

7. The feeding compensation method for a high cross-ratio circularly polarized antenna array according to claim 1, characterized in that, The ideal compensation value is the compensation value for the real-time amplitude and phase of the feed signal of each antenna element.

8. The feeding compensation method for a high cross-ratio circularly polarized antenna array according to claim 1, characterized in that, S102, The specific method for constructing a neural network with an encoder-decoder LSTM architecture to obtain the power supply compensation model is as follows; S1021, Set the hyperparameters of the encoder-decoder LSTM architecture neural network, including setting model structure parameters and training process parameters; S1022, using the change of the working parameters as a time step, the training sample library is serialized according to the order of the change of the working parameters; S1023 employs an encoder consisting of one or more LSTM layers to read the input sequence of working parameters, update the hidden state according to the input at each time step, and output the final hidden state encoded as a context vector. S1024 employs a decoder consisting of one or more LSTM layers. The input context vector is used as the initial hidden state. Based on the input at each time step, the output vector is obtained. The output vector is then passed through a fully connected layer to generate a sequence of predicted values ​​for the ideal compensation values ​​step by step. S1025 compares the predicted value of the ideal compensation value with the ideal compensation value item by item using the loss function, calculates the total loss value of the ideal compensation value, and feeds the total loss value back to the decoder and encoder through the backpropagation algorithm for parameter update.

9. A feed compensation system for a high cross-pole ratio circularly polarized antenna array, employing the feed compensation method described in claim 1, characterized in that, Includes an antenna array control module, a feed compensation prediction module, a feed adjustment module, and a phased array antenna; The antenna array control module acquires the operating parameters of the antenna array, outputs the operating parameters to the feed compensation prediction module, acquires the predicted value of the ideal compensation value fed back by the feed compensation prediction module, superimposes the predicted value of the ideal compensation value with the scanning feed signal, and outputs the superimposed signal to the feed adjustment module. The feed compensation prediction module is equipped with a final feed compensation model. Based on the input operating parameters, it obtains the predicted value of the ideal compensation value and returns it to the antenna array control module. The antenna array is a phased array antenna, which is composed of an array of antenna elements. Each antenna element has a feed signal inlet, which is connected to the feed adjustment module. The feed adjustment module adjusts the feed signal transmitted to each antenna element based on the superimposed signal.

10. A computer program product, characterized in that, The computer program product includes a computer program or instructions that enable the computer program or instructions to implement the feeding compensation method for the high cross-pole ratio circularly polarized antenna array as described in claim 1.

Images