Automatic calibration method for liquid crystal phased array antenna

Abstract

The present invention proposes an automatic calibration method for a liquid crystal phased array antenna, which relates to the field of communication technology and includes four steps of layout before calibration, data collection before calibration, calculation of calibration parameters and antenna calibration. The present invention performs environment layout before calibration, adopts a near-field probe array method, and uses multiple probes to collect data, thereby greatly improving the overall measurement efficiency. At the same time, the data acquisition method of multiple probes acquires more information than that of a single probe, thereby providing better data support for subsequent calculations. Subsequently, a convolutional neural network is introduced to construct an intelligent calibration algorithm model. The acquired data is calculated by the intelligent calibration algorithm model, and calibration is performed according to the calculation results to achieve automatic calibration of the liquid crystal phased array antenna.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to an automatic calibration method for a liquid crystal phased array antenna. Background Art

[0002] Liquid crystal phased array antenna is an antenna based on liquid crystal technology and phased array technology. It drives liquid crystal molecules through electronic devices to adjust the antenna radiation direction and beam shape. Compared with traditional fixed antennas, liquid crystal phased array antennas can achieve flexible electronic scanning, focusing and beam control, and have higher antenna gain and directivity. Liquid crystal phased array antennas are usually composed of one or more antenna elements and a liquid crystal layer array. Each antenna element is responsible for transmitting or receiving signals, and the direction and shape of the radiation beam can be adjusted by controlling the phase and amplitude of the liquid crystal layer. Due to the variability of the liquid crystal layer, the liquid crystal phased array antenna can be electronically scanned and focused in different directions, thereby achieving higher antenna gain and anti-interference capabilities.

[0003] For liquid crystal phased array antennas, in order to obtain more ideal antenna radiation characteristics during their use, they need to be calibrated. At present, the calibration method of phased array antennas generally adopts near-field measurement method for calibration. The near-field measurement method uses a single probe to sample the electric field around the closed surface of the array, and then obtains the far-field radiation pattern and the radial distribution of the unit channel through numerical calculation, and then further calibrates.

[0004] This method has the advantage of high measurement accuracy, but in actual application, it adopts a single-probe mechanical scanning form, and the test speed is slow, which affects the calibration efficiency of the entire liquid crystal phased array antenna. At the same time, the single-probe method still has some abnormal situations, such as a small amount of information obtained, which will further have an adverse effect on the subsequent calibration accuracy. Therefore, the present invention proposes a liquid crystal phased array antenna automatic calibration method to solve the problems existing in the prior art. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to propose an automatic calibration method for a liquid crystal phased array antenna, which has the advantages of improving calibration efficiency and accuracy and can solve the problems existing in the prior art.

[0006] To achieve the purpose of the present invention, the present invention is implemented by the following technical solution: a liquid crystal phased array antenna automatic calibration method, comprising the following steps:

[0007] Step 1: Layout before calibration

[0008] A measurement layout is constructed in a microwave darkroom, and a multi-probe near-field measurement layout adapted to the liquid crystal phased array antenna is arranged according to the size and frequency range of the liquid crystal phased array antenna. In the multi-probe near-field measurement layout, a near-field probe array is set, wherein the number of probes in the near-field probe array is N, and then an antenna turntable is arranged to place the liquid crystal phased array antenna to be tested;

[0009] Step 2: Data collection before calibration

[0010] In the layout constructed in step 1, a radio frequency unit and a near-field probe array controller are set, and the radio frequency unit transmits the frequency to the liquid crystal phased array antenna to be measured, and the near-field probe array receives the data and returns it to the near-field probe array controller to obtain the measurement data of the near-field probe array;

[0011] Step 3: Calculation of calibration parameters

[0012] An intelligent calibration algorithm model is constructed based on a neural network, and the measurement data obtained in step 2 is input into the intelligent calibration algorithm model, and the intelligent calibration algorithm model processes the measurement data to determine the phase adjustment amount required for each array element of the liquid crystal phased array antenna;

[0013] Step 4: Antenna Calibration

[0014] According to the result calculated in step three, the liquid crystal phased array antenna is adjusted to change the phase delay of the liquid crystal layer to achieve calibration.

[0015] A further improvement is that in step 1, the near-field measurement layout includes the moving path, step distance and arrangement density of the probes.

[0016] A further improvement is that in step 1, N≥2.

[0017] A further improvement is that in step 2, the step of measuring data is:

[0018] S1: The near-field probe array controller generates a modulation signal and an identification signal, and adds the modulation signal to each probe, while adding the identification signal to each probe in turn;

[0019] S2: The near-field probe array controller receives signal data without identification signals through a demodulator and samples it through a synchronous detector;

[0020] S3: The sampled data is transmitted to the converter.

[0021] A further improvement is that in step 3, the specific steps of the neural network constructing the intelligent calibration algorithm model are:

[0022] A1: Collect existing near-field measurement data sets, process the collected data sets, remove noise and outliers, and convert the formats;

[0023] A2: Select the main target features from the processed data;

[0024] A3: Divide the data set into training set, validation set and test set, build a convolutional neural network, and use the training set to train the convolutional neural network;

[0025] A4: Use the validation set to validate the trained convolutional neural network and adjust the parameters of the convolutional neural network based on the validation results;

[0026] A5: Use the test set to test the adjusted convolutional neural network model, and obtain the trained intelligent calibration algorithm model after testing.

[0027] A further improvement is that in step 4, the adjustment method is to apply corresponding voltage adjustment to each array element unit of the liquid crystal phased array antenna.

[0028] A further improvement is that in step 4, after calibration, step 2 is performed again to obtain a new measurement result, the measurement results before and after calibration are compared to confirm the accuracy of the calibration result, the obtained measurement result is used as a benchmark, and step 3 is performed again to obtain an iterative calculation result.

[0029] The beneficial effects of the present invention are as follows: the present invention provides an automatic calibration method for a liquid crystal phased array antenna, which greatly improves the overall measurement efficiency by performing environmental layout before calibration, adopting a near-field probe array, and using multiple probes to collect data. At the same time, the data acquisition method of multiple probes acquires more information than that of a single probe, providing better data support for subsequent calculations. Subsequently, a convolutional neural network is introduced to construct an intelligent calibration algorithm model, and the acquired data is calculated by the intelligent calibration algorithm model, so that calibration is performed according to the calculation results to achieve automatic calibration of the liquid crystal phased array antenna. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0031] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0032] Embodiment 1

[0033] according to Figure 1 As shown, this embodiment proposes a liquid crystal phased array antenna automatic calibration method, comprising the following steps:

[0034] Step 1: Layout before calibration

[0035] A measurement layout is constructed in a microwave darkroom, and a multi-probe near-field measurement layout adapted to the liquid crystal phased array antenna is arranged according to the size and frequency range of the liquid crystal phased array antenna, wherein a near-field probe array is set in the multi-probe near-field measurement layout, wherein the number of probes in the near-field probe array is N, wherein N=2, and then an antenna turntable is arranged for placing the liquid crystal phased array antenna to be measured, wherein the near-field measurement layout includes a moving path of the probe, a stepping distance, and a layout density of the probe;

[0036] Step 2: Data collection before calibration

[0037] In the layout constructed in step 1, a radio frequency unit and a near-field probe array controller are set, and the radio frequency unit transmits the frequency to the liquid crystal phased array antenna to be measured, and the near-field probe array receives the data and returns it to the near-field probe array controller to obtain the measurement data of the near-field probe array, wherein the steps of measuring the data are:

[0038] S1: The near-field probe array controller generates a modulation signal and an identification signal, and adds the modulation signal to each probe, while adding the identification signal to each probe in turn;

[0039] S2: The near-field probe array controller receives signal data without identification signals through a demodulator and samples it through a synchronous detector;

[0040] S3: The sampled data is transmitted to the converter;

[0041] Step 3: Calculation of calibration parameters

[0042] An intelligent calibration algorithm model is constructed based on a neural network, and the measurement data obtained in step 2 is input into the intelligent calibration algorithm model. The intelligent calibration algorithm model processes the measurement data to determine the phase adjustment amount required for each element of the liquid crystal phased array antenna. The specific steps of constructing the intelligent calibration algorithm model by the neural network are as follows:

[0043] A1: Collect existing near-field measurement data sets, process the collected data sets, remove noise and outliers, and convert the formats;

[0044] A2: Select the main target features from the processed data;

[0045] A3: Divide the data set into training set, validation set and test set, build a convolutional neural network, and use the training set to train the convolutional neural network;

[0046] A4: Use the validation set to validate the trained convolutional neural network and adjust the parameters of the convolutional neural network based on the validation results;

[0047] A5: Use the test set to test the adjusted convolutional neural network model, and obtain the trained intelligent calibration algorithm model after testing;

[0048] Step 4: Antenna Calibration

[0049] According to the result calculated in step three, the liquid crystal phased array antenna is adjusted to change the phase delay of the liquid crystal layer to achieve calibration. Specifically, the adjustment method is to apply corresponding voltage adjustment to each array element unit of the liquid crystal phased array antenna. After calibration, step two is performed again to obtain a new measurement result. The measurement results before and after calibration are compared to confirm the accuracy of the calibration result. The obtained measurement result can be used as a benchmark to perform step three again to obtain an iterative calculation result, and the liquid crystal phased array antenna is adjusted again according to the iterative calculation result.

[0050] Embodiment 2

[0051] according to Figure 1 As shown, this embodiment proposes a liquid crystal phased array antenna automatic calibration method, comprising the following steps:

[0052] Step 1: Layout before calibration

[0053] In this embodiment, the present invention adopts a near-field measurement method to automatically calibrate the liquid crystal phased array antenna. However, the difference from the traditional near-field measurement method is that the traditional single-probe near-field measurement requires point-by-point scanning, which is inefficient. Therefore, in this application, multiple near-field probes are used for simultaneous measurement, which significantly improves the speed of data acquisition and improves efficiency. Specifically, a measurement layout is constructed in a microwave darkroom, and a multi-probe near-field measurement layout adapted to the liquid crystal phased array antenna is arranged according to the size and frequency range of the liquid crystal phased array antenna. In the multi-probe near-field measurement layout, a near-field probe array is set, wherein the near-field probe The number of probes in the head array is N, and an antenna turntable is arranged to place the liquid crystal phased array antenna to be tested. The position of the antenna turntable corresponds to the position of the near-field probe. Specifically, N>2. In this embodiment, two groups of near-field probes are set as an example, wherein the near-field measurement layout includes the moving path of the probe, the stepping distance and the arrangement density of the probe. It is based on the rectangular scanning method to ensure that the moving path of the near-field probe array can cover the near-field area of ​​the antenna. Therefore, when collecting data, the near-field probe array will move on the external mobile device according to the preset moving path to collect data;

[0054] Step 2: Data collection before calibration

[0055] After the layout is built, relevant data is collected for subsequent calculations and adjustments. Specifically, a radio frequency unit and a near-field probe array controller are set up. The radio frequency unit transmits the frequency of the liquid crystal phased array antenna to be tested. The radio frequency unit uses the main signal source and the auxiliary signal source as signal sources. The main signal source sends a frequency of F0, and the auxiliary signal source works at a frequency of F0+F if The RF unit actually plays the role of frequency conversion and transmits the intermediate frequency to the near-field probe array controller. Correspondingly, the near-field probe array controller generates a modulation signal F M and identification signal F I , and add the modulation signal to each probe, and add the identification signal to each probe in turn, so that at a certain moment, only one probe has a frequency of F if +F M , the frequency of other probes is F if +F M +F I Then, the near-field probe array controller receives the signal data without the identification signal through the demodulator (the filter receiving circuit of the demodulator), so that the real part and the imaginary part of the electromagnetic field of each probe can be sampled in turn through the synchronous detector and transmitted to the converter (analog / digital converter) to obtain the measurement data of the near-field probe array;

[0056] Step 3: Calculation of calibration parameters

[0057] After measuring the near-field data of the liquid crystal phased array antenna, the far-field radiation pattern and the radiation and phase distribution of each unit channel can be obtained through numerical calculation. However, this is only the basis for antenna calibration. The purpose of its calibration is to ensure that the antenna system can maintain optimal performance under various working conditions. Therefore, in the process of automatic calibration, a convolutional neural network is introduced, which is one of the deep learning algorithms. Therefore, an intelligent calibration algorithm model is constructed based on the neural network, and then the measurement data obtained in step 2 is input into the intelligent calibration algorithm model. The intelligent calibration algorithm model processes the measurement data to determine the phase adjustment required for each array element of the liquid crystal phased array antenna, thereby achieving an automatic calibration effect. The measured data is input into the model to obtain the calculation results, and then the antenna parameters are automatically adjusted according to the calculation results.

[0058] Specifically, the specific steps of the neural network to build an intelligent calibration algorithm model are:

[0059] A1: Collect existing near-field measurement data sets, process the collected data sets, remove noise and outliers, fill in missing data, and convert the format. The collected data includes the electric field amplitude and phase distribution of the liquid crystal phased array antenna, as well as related environmental variables;

[0060] A2: From the processed data, select the main target features that can fully reflect the radiation characteristics of the antenna;

[0061] A3: Divide the data set into training set, validation set and test set, build a convolutional neural network, and use the training set to train the convolutional neural network. During the training process, the convolutional neural network will continuously adjust its weights and biases through back propagation and gradient descent to minimize the prediction error.

[0062] A4: Use the validation set to validate the trained convolutional neural network, and adjust the parameters of the convolutional neural network, such as the number of iterations and learning rate, based on the validation results. The performance of the network model is mainly optimized through validation.

[0063] A5: Use the test set to test the adjusted convolutional neural network model, that is, use the test set to evaluate the trained convolutional neural network model and check its prediction ability. Correspondingly, the trained intelligent calibration algorithm model is obtained after the test.

[0064] Step 4: Antenna Calibration

[0065] According to the result calculated in step three, the corresponding adjustment parameters are obtained, so the liquid crystal phased array antenna is adjusted. The adjustment method is to apply corresponding voltages to each array element unit of the liquid crystal phased array antenna to adjust the phase delay of the liquid crystal layer to achieve calibration. The working principle of the liquid crystal phased array antenna is to apply deflection voltages above and below the liquid crystal layer, and use the dielectric anisotropy of the liquid crystal to control the deflection direction of the liquid crystal, thereby changing the phase shift size of the phase shifter. This change will affect the electrical length of the transmission line feeding each antenna unit, thereby achieving the regulation of the phase of the incident signal of each antenna unit. In this way, the beam pointing of the antenna can be accurately controlled to achieve the beam scanning function. Correspondingly, after calibration, step two is performed again to obtain new measurement results, and the measurement results before and after calibration are compared to confirm the accuracy of the calibration results.

[0066] In combination with Example 1 and Example 2, the present invention is based on a near-field measurement method and then adopts a multi-probe measurement method to improve the measurement efficiency, and introduces a deep learning method to construct an intelligent calibration algorithm model based on a convolutional neural network. Combined with the multi-probe measurement method, the measurement efficiency is guaranteed while improving the accuracy of subsequent antenna automatic calibration.

[0067] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the framework and scope of application of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A liquid crystal phased array antenna automatic calibration method, characterized in that: The following steps are involved: Step 1: Layout before calibration A measurement layout is constructed in a microwave darkroom, and a multi-probe near-field measurement layout adapted to the liquid crystal phased array antenna is arranged according to the size and frequency range of the liquid crystal phased array antenna. In the multi-probe near-field measurement layout, a near-field probe array is set, wherein the number of probes in the near-field probe array is N, and then an antenna turntable is arranged to place the liquid crystal phased array antenna to be tested; Step 2: Data collection before calibration In the layout constructed in step 1, a radio frequency unit and a near-field probe array controller are set, and a radio frequency signal is transmitted to the liquid crystal phased array antenna to be measured by the radio frequency unit, and the near-field probe array receives the data and returns it to the near-field probe array controller to obtain the measurement data of the near-field probe array. In step 2, the steps of measuring the data are as follows: S1: The near-field probe array controller generates a modulation signal and an identification signal, and adds the modulation signal to each probe, while adding the identification signal to each probe in turn; S2: The near-field probe array controller receives signal data without identification signals through a demodulator and samples it through a synchronous detector; S3: The sampled data is transmitted to the converter; Step 3: Calculation of calibration parameters An intelligent calibration algorithm model is constructed based on a neural network, and the measurement data obtained in step 2 is input into the intelligent calibration algorithm model. The intelligent calibration algorithm model processes the measurement data to determine the phase adjustment amount required for each element of the liquid crystal phased array antenna. In step 3, the specific steps of the neural network to construct the intelligent calibration algorithm model are as follows: A1: Collect existing near-field measurement data sets, process the collected data sets, remove noise and outliers, and convert the formats; A2: Select the target features that reflect the antenna radiation characteristics from the processed data; A3: Divide the data set into training set, validation set and test set, build a convolutional neural network, and use the training set to train the convolutional neural network; A4: Use the validation set to validate the trained convolutional neural network and adjust the parameters of the convolutional neural network based on the validation results; A5: Use the test set to test the adjusted convolutional neural network model, and obtain the trained intelligent calibration algorithm model after testing; Step 4: Antenna Calibration According to the result calculated in step three, the liquid crystal phased array antenna is adjusted to change the phase delay of the liquid crystal layer to achieve calibration.

2. The automatic calibration method of a liquid crystal phased array antenna according to claim 1, characterized in that: In the step 1, the near-field measurement layout includes the moving path, step distance and arrangement density of the probes.

3. The automatic calibration method of a liquid crystal phased array antenna according to claim 1, characterized in that: In the step 1, N≥2.

4. The automatic calibration method for a liquid crystal phased array antenna according to claim 1, characterized in that: In the step 4, the adjustment method is to apply corresponding voltage adjustment to each array element of the liquid crystal phased array antenna.

5. The automatic calibration method of a liquid crystal phased array antenna according to claim 1, characterized in that: In step 4, after calibration, step 2 is performed again to obtain a new measurement result, the measurement results before and after calibration are compared to confirm the accuracy of the calibration result, the obtained measurement result is used as a benchmark, and step 3 is performed again to obtain an iterative calculation result.

Images