Artificial neural network monopulse angle measurement method and system applied to feed source mismatched antenna

Abstract

The invention discloses an artificial neural network monopulse angle measurement method and an artificial neural network monopulse angle measurement system applied to a feed source mismatched antenna, and solves the problems that for a five-horn monopulse antenna with a feed source mismatched, an existing monopulse angle measurement method with independent orientation and pitching can introduce extra errors, and error factors between orientation and pitch can generate coupling, so that the angle measurement accuracy is poor. And the accuracy of an angle measurement result is seriously influenced. In order to solve the problem of angle measurement of a five-horn monopulse antenna with a mismatched feed source, a nonlinear mapping relation between a complex ratio of a pitch difference signal, an azimuth difference signal and a sum channel signal of the antenna and a pitch angle and an azimuth angle under the condition of feed source mismatch is established through an artificial neural network, and a high-precision angle measurement result is obtained.

Description

Technical Field

[0001] The invention relates to a single pulse angle measurement method and system, and in particular to an artificial neural network single pulse angle measurement method and system applied to a feed mismatch antenna. Background Art

[0002] Single pulse angle measurement has the advantages of high measurement accuracy, strong anti-interference ability and high data rate. It is widely used in the fields of precision tracking measurement of target range, target tracking and radio astronomy. Single pulse technology obtains target angle information by comparing the signals received simultaneously by two or more antennas.

[0003] The monopulse antenna composed of a reflector and a monopulse feed is a commonly used tracking antenna, and the monopulse feed is the core of the monopulse antenna. Commonly used monopulse feeds include: multi-horn feeds, single-aperture multi-mode feeds, and multi-horn multi-mode feeds. Among them, the sum and difference network of the five-horn feed is easier to implement in engineering and has become the mainstream form of monopulse antenna feed; however, the lack of accuracy in the installation position of the five-horn feed will lead to monopulse sum and difference beam distortion, the phenomenon of which is that the azimuth difference formed is not only related to the azimuth angle, but also has a certain coupling relationship with the elevation angle position; the elevation difference formed also has a similar situation, and the above phenomenon is called feed mismatch.

[0004] When the above phenomenon occurs, if the original independent single-pulse angle measurement method for azimuth and elevation is still used, such as the angle detection curve method, additional errors will be introduced, and the error factors between azimuth and elevation will also be coupled, seriously affecting the accuracy of the angle measurement results. Summary of the invention

[0005] The purpose of the present invention is to provide an artificial neural network single-pulse angle measurement method and system for feed mismatched antennas, so as to solve the technical problem that the existing independent single-pulse angle measurement methods for azimuth and elevation for five-horn single-pulse antennas with feed mismatch will introduce additional errors, and the error factors between azimuth and elevation will also produce coupling, seriously affecting the accuracy of its angle measurement results.

[0006] Aiming at the problem of angle measurement of a five-horn single-pulse antenna with feed mismatch, the present invention establishes a nonlinear mapping relationship between the complex ratio of the antenna elevation difference signal, azimuth difference signal and channel signal under feed mismatch condition and the elevation angle and azimuth angle through an artificial neural network, thereby obtaining a high-precision angle measurement result.

[0007] In order to achieve the above purpose and complete the above invention concept, the present invention adopts the following technical solutions:

[0008] An artificial neural network single pulse angle measurement method applied to a feed mismatched antenna is special in that it includes the following steps:

[0009] Step 1, setting a single pulse receiving antenna at the action end of the antenna servo mechanism, setting up a radio frequency signal source on a high tower, so that the relative position of the single pulse receiving antenna and the radio frequency signal source meets the far field condition; the single pulse receiving antenna can receive and channel beam, elevation difference beam, and azimuth difference beam;

[0010] Step 2, the antenna servo mechanism controls the single pulse receiving antenna to scan in two dimensions, elevation and azimuth, with the RF signal source as the center, to obtain the sum channel signal, the elevation difference signal and the azimuth difference signal, and the antenna servo mechanism outputs the corresponding elevation angle and azimuth angle;

[0011] Step 3, using the real part and imaginary part of the complex ratio of the pitch difference signal to the sum channel signal, and the real part and imaginary part of the complex ratio of the azimuth difference signal to the sum channel signal as inputs of the artificial neural network, using the corresponding pitch angle and azimuth angle as outputs of the artificial neural network, training the artificial neural network, and forming a neural network single pulse angle measurement model;

[0012] Step 4: Perform single pulse angle measurement using the neural network single pulse angle measurement model to complete the artificial neural network single pulse angle measurement.

[0013] Furthermore, step 3 is specifically as follows: using the real part and the imaginary part of the complex ratio of the pitch difference signal to the sum channel signal, and the real part and the imaginary part of the complex ratio of the azimuth difference signal to the sum channel signal as inputs of the artificial neural network, using the corresponding pitch angle and azimuth angle as outputs of the artificial neural network, and using the conjugate gradient descent method to train the artificial neural network to form a neural network single pulse angle measurement model.

[0014] Furthermore, in step 3, the input layer of the artificial neural network includes 4 units, the hidden layer includes 12 units, the output layer includes 2 units, and the nonlinear function used is the Sigmod function.

[0015] Furthermore, in step 1, the single-pulse receiving antenna uses five horn antennas as receiving feed sources, the middle horn antenna is used for transmitting and receiving and sharing the channel, the upper and lower horn antennas are used to form an elevation difference beam, and the left and right horn antennas form an azimuth difference beam.

[0016] Furthermore, in step 1, the single pulse receiving antenna operates in the X-band.

[0017] Furthermore, in step 1, the main reflection surface aperture of the single-pulse receiving antenna is 2.4 meters.

[0018] An artificial neural network monopulse angle measurement system applied to a feed mismatched antenna is used to implement the above-mentioned artificial neural network monopulse angle measurement method applied to a feed mismatched antenna, and its special features are:

[0019] It includes a high tower, a radio frequency signal source erected on the high tower, an antenna servo mechanism, and a single pulse receiving antenna arranged at the action end of the antenna servo mechanism;

[0020] The radio frequency signal source is used to radiate to the single pulse receiving antenna;

[0021] The antenna servo mechanism is used to control the single pulse receiving antenna to scan in two dimensions, elevation and azimuth, with the radio frequency signal source as the center;

[0022] The single pulse receiving antenna is used to continuously acquire the sum channel signal, the pitch difference signal and the azimuth difference signal during the scanning process; the relative position of the single pulse receiving antenna and the radio frequency signal source meets the far field condition.

[0023] Furthermore, the single-pulse receiving antenna uses five horn antennas as receiving feed sources. The horn antenna in the middle is used for transmitting and receiving and sharing the channel. The upper and lower horn antennas are used to form an elevation difference beam, and the left and right horn antennas form an azimuth difference beam.

[0024] Furthermore, the single-pulse receiving antenna operates in the X-band.

[0025] Furthermore, the main reflection surface aperture of the single-pulse receiving antenna is 2.4 meters.

[0026] Beneficial effects of the present invention:

[0027] 1. The present invention is applied to an artificial neural network single-pulse angle measurement method and system for a feed mismatched antenna. An artificial neural network is used to train a measured data set of two-dimensional sum and difference directions of the antenna to form a neural network single-pulse angle measurement model. The coupling relationship between the pitch difference and azimuth difference beams and the azimuth and pitch angles under the condition of feed mismatch is comprehensively considered. Compared with the traditional method of independently considering the pitch angle and azimuth angle, higher angle measurement accuracy can be obtained.

[0028] 2. The present invention is applied to an artificial neural network single-pulse angle measurement method and system for feed mismatched antennas. The neural network single-pulse angle measurement model uses the real part and imaginary part of the complex ratio of the elevation difference signal and azimuth difference signal received by the five-horn antenna feed and the channel signal as the input of the artificial neural network, and uses the azimuth angle and elevation angle as the network output. The trained neural network single-pulse angle measurement model can avoid phase compensation of the signal receiving channel, and can obtain angle measurement results with higher accuracy than the angle detection curve method.

[0029] 3. In the traditional single-pulse angle measurement system, the RF signal derived from the RF summator needs to pass through different transmission lines and different receiver channels to reach the digital baseband. When using the single-pulse angle measurement method, the difference channel needs to be phase calibrated. The artificial neural network single-pulse angle measurement method proposed in the present invention leaves the phase calibration work to the neural network itself, and there is no need to compensate the channel phase.

[0030] 4. The artificial neural network single pulse angle measurement model in the present invention uses the real part and imaginary part of the complex ratio of the pitch difference signal, the azimuth difference signal and the channel signal as neural network input. This input method normalizes the signal strength using the channel, so that the trained neural network single pulse angle measurement model meets the requirements for use under different signal strengths. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a structural schematic diagram of an embodiment of an artificial neural network single pulse angle measurement system applied to a feed mismatched antenna of the present invention;

[0032] Figure 2 1 is a two-dimensional directional pattern test result diagram of the single pulse receiving antenna in the embodiment of the present invention; wherein (a) is a two-dimensional distribution diagram of the sum beam, (b) is a two-dimensional distribution diagram of the elevation difference beam amplitude, (c) is a two-dimensional distribution diagram of the azimuth difference beam amplitude, (d) is a two-dimensional distribution diagram of the elevation difference beam phase, and (e) is a two-dimensional distribution diagram of the azimuth difference beam phase;

[0033] Figure 3 is a schematic diagram of the structure of an artificial neural network single pulse angle measurement model in an embodiment of the present invention;

[0034] Figure 4 : is a probability distribution diagram of angle measurement error of an artificial neural network in an embodiment of the present invention; wherein (a) is a probability distribution diagram of pitch error, and (b) is a probability distribution diagram of azimuth error;

[0035] Figure 5 It is the existing single pulse angle measurement error probability distribution diagram; among them, (a) is the pitch error probability distribution diagram, and (b) is the azimuth error probability distribution diagram.

[0036] Figure Number:

[0037] 1- Single pulse receiving antenna, 2- Antenna servo mechanism, 3- High tower, 4- RF signal source. DETAILED DESCRIPTION

[0038] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] An embodiment of the present invention provides an artificial neural network single pulse angle measurement system applied to a feed mismatched antenna, such as Figure 1 As shown, the single pulse angle measurement system includes a high tower 3, a radio frequency signal source 4 mounted on the high tower 3, an antenna servo mechanism 2, and a single pulse receiving antenna 1 arranged at the action end of the antenna servo mechanism 2.

[0040] In this embodiment, the relative position of the single pulse receiving antenna 1 and the RF signal source 4 meets the far field condition in the test field, that is, for the single pulse receiving antenna 1 with an aperture of D, the relative distance R between the single pulse receiving antenna 1 and the RF signal source 4 satisfies R>2D 2 / λ; the RF signal source 4 is used to radiate to the single-pulse receiving antenna; the antenna servo mechanism 2 is used to control the single-pulse receiving antenna 1 to scan in two dimensions, elevation and azimuth, with the RF signal source 4 as the center; the single-pulse receiving antenna 1 used includes three coherent independent receiving channels, which are used to continuously obtain the sum channel signal, the elevation difference signal and the azimuth difference signal during the scanning process. The single-pulse receiving antenna 1 works in the X-band, and the main reflector has an aperture of 2.4 meters. It uses a 5-horn antenna as a receiving feed source, in which the middle horn is shared by the transmitting and receiving sum channels, the upper and lower feed sources form a pitch difference beam, and the left and right feed sources form an azimuth difference beam.

[0041] The above-mentioned artificial neural network single pulse angle measurement system applied to the feed mismatch antenna is used to perform single pulse angle measurement, which specifically includes the following steps:

[0042] Step 1: Set the single pulse receiving antenna 1 at the action end of the antenna servo mechanism 2, and set up a radio frequency signal source 4 on a high tower 3; the radio frequency signal source 4 radiates toward the single pulse receiving antenna 1, so that the relative position of the single pulse receiving antenna 1 and the radio frequency signal source 4 meets the far field condition.

[0043] Step 2, control the single-pulse receiving antenna 1 through the antenna servo mechanism 2 to scan in two dimensions, namely, elevation and azimuth, with the RF signal source 4 as the center, to obtain the sum channel signal, the pitch difference signal and the azimuth difference signal. The sum channel signal, the pitch difference signal and the azimuth difference signal obtained are all complex numbers containing real and imaginary parts. At the same time, the antenna servo mechanism 2 outputs the corresponding elevation angle and azimuth angle.

[0044] like Figure 2 As shown, Figure 2(a) is the two-dimensional distribution diagram of the sum beam, Figure 2 (b) and Figure 2 (d) The amplitude and phase of the elevation difference beam are well symmetrical about the elevation plane, and the elevation plane is slightly tilted. Figure 2 (c) and Figure 2 (e) The symmetry of the amplitude and phase of the azimuth difference with respect to the azimuth plane is destroyed, and the azimuth plane has a more obvious tilt.

[0045] In engineering, the elevation angle and azimuth angle of a single pulse receiving antenna are considered independently. Figure 2 (c) and Figure 2 It is obvious from (e) that the amplitude of the difference ratio of the azimuth is different at different elevation angles. In this case, directly using the azimuth angle detection curve will lead to a larger measurement error of the azimuth angle.

[0046] Step 3: Use the real part and imaginary part of the complex ratio of the pitch difference signal and the sum channel signal, and the real part and imaginary part of the complex ratio of the azimuth difference signal and the sum channel signal as the input of the artificial neural network, use the corresponding pitch angle and azimuth angle as the output of the artificial neural network, use the conjugate gradient descent method to train the neural network, and form a neural network single pulse angle measurement model.

[0047] The complex ratio is the ratio of two complex numbers. For example, when the pitch difference signal and the sum channel signal are a+bi and c+di respectively, when the complex ratio of the pitch difference signal and the sum channel signal is (a+bi) / (c+di), the ratio of the two complex numbers is still a complex number.

[0048] Similar to single pulse angle measurement, the real and imaginary parts of the complex ratio of the elevation difference and the sum signal channel, and the real and imaginary parts of the complex ratio of the azimuth difference and the sum signal channel are used as the input of the artificial neural network; this input method normalizes the signal strength using the sum channel, so that the training results of the artificial neural network meet the requirements of use under different signal strengths. In addition, for the actual single pulse receiving antenna, the RF signal derived from the RF sum and difference device needs to pass through different transmission lines and different receiver channels to reach the digital baseband. When using the single pulse angle measurement method, the difference channel also needs to be phase calibrated. However, when using artificial neural network angle measurement, the phase calibration work can be directly handed over to the network itself. In other words, the artificial neural network is trained directly using the complex ratio data of the tested directional pattern, and there is no need to compensate the channel phase when using the artificial neural network.

[0049] like Figure 3 As shown, the input layer of the artificial neural network used in this embodiment has 4 units, the hidden layer has 12 units, and the output layer has 2 units. The nonlinear function used is the Sigmod function. The network training method used is the conjugate gradient descent method.

[0050] The artificial neural network is trained using two-dimensional sum and difference directional pattern data, and the data is used to build a neural network single pulse angle measurement model to analyze the angle measurement error.

[0051] Step 4: Perform single pulse angle measurement through the neural network single pulse angle measurement model to complete the artificial neural network single pulse angle measurement. Figure 4 and Figure 5 ,It can be seen that the error distribution of the angle measurement of the artificial neural network is obviously more concentrated and the angle measurement accuracy is higher.

[0052] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An artificial neural network single pulse angle measurement method applied to feed mismatched antenna, characterized in that: The following steps are involved: Step 1, a single pulse receiving antenna (1) is arranged at the action end of an antenna servo mechanism (2), and a radio frequency signal source (4) is installed on a high tower (3), so that the relative position of the single pulse receiving antenna (1) and the radio frequency signal source (4) meets the far field condition; the single pulse receiving antenna (1) can receive and channel beam, elevation difference beam, and azimuth difference beam; Step 2, the antenna servo mechanism (2) controls the single pulse receiving antenna (1) to scan in two dimensions, elevation and azimuth, with the RF signal source (4) as the center, to obtain a sum channel signal, an elevation difference signal and an azimuth difference signal, and the antenna servo mechanism (2) outputs the corresponding elevation angle and azimuth angle; Step 3, using the real part and imaginary part of the complex ratio of the pitch difference signal to the sum channel signal, and the real part and imaginary part of the complex ratio of the azimuth difference signal to the sum channel signal as inputs of the artificial neural network, using the corresponding pitch angle and azimuth angle as outputs of the artificial neural network, training the artificial neural network, and forming a neural network single pulse angle measurement model; Step 4: Perform single pulse angle measurement using the neural network single pulse angle measurement model to complete the artificial neural network single pulse angle measurement.

2. The artificial neural network single pulse angle measurement method applied to feed mismatched antenna according to claim 1, characterized in that: Specifically, step 3 uses the real part and imaginary part of the complex ratio of the pitch difference signal to the sum channel signal, and the real part and imaginary part of the complex ratio of the azimuth difference signal to the sum channel signal as the input of the artificial neural network, uses the corresponding pitch angle and azimuth angle as the output of the artificial neural network, and uses the conjugate gradient descent method to train the artificial neural network to form a neural network single pulse angle measurement model.

3. The artificial neural network single pulse angle measurement method applied to feed mismatched antenna according to claim 2, characterized in that: In step 3, the input layer of the artificial neural network includes 4 units, the hidden layer includes 12 units, the output layer includes 2 units, and the nonlinear function used is the Sigmod function.

4. The artificial neural network single pulse angle measurement method applied to feed mismatched antenna according to any one of claims 1 to 3, characterized in that: In step 1, the single pulse receiving antenna (1) uses five horn antennas as receiving feed sources, the horn antenna in the middle is used for transmitting and receiving and sharing the channel, the upper and lower horn antennas are used to form an elevation difference beam, and the left and right horn antennas form an azimuth difference beam.

5. The artificial neural network single pulse angle measurement method applied to feed mismatched antenna according to claim 4, characterized in that: In step 1, the monopulse receiving antenna (1) operates in the X-band.

6. The artificial neural network single pulse angle measurement method applied to feed mismatched antenna according to claim 5, characterized in that: In step 1, the main reflection surface aperture of the single pulse receiving antenna (1) is 2.4 meters.

7. An artificial neural network single pulse angle measurement system applied to a feed mismatched antenna, used to implement the artificial neural network single pulse angle measurement method applied to a feed mismatched antenna as described in any one of claims 1 to 6, characterized in that: It comprises a high tower (3), a radio frequency signal source (4) erected on the high tower (3), an antenna servo mechanism (2), and a single pulse receiving antenna (1) arranged at the action end of the antenna servo mechanism (2); The radio frequency signal source (4) is used to radiate to the single pulse receiving antenna; The antenna servo mechanism (2) is used to control the single pulse receiving antenna (1) to scan in two dimensions, elevation and azimuth, with the radio frequency signal source (4) as the center; The single pulse receiving antenna (1) is used to continuously acquire the sum channel signal, the pitch difference signal and the azimuth difference signal during the scanning process; the relative position of the single pulse receiving antenna (1) and the radio frequency signal source (4) meets the far field condition.

8. The artificial neural network monopulse angle measurement system applied to feed mismatched antenna according to claim 7, characterized in that: The single pulse receiving antenna (1) uses five horn antennas as receiving feed sources, the horn antenna in the middle is used for transmitting and receiving and sharing a channel, the upper and lower horn antennas are used to form an elevation difference beam, and the left and right horn antennas form an azimuth difference beam.

9. The artificial neural network monopulse angle measurement system for feed mismatched antenna according to claim 8, characterized in that: The single pulse receiving antenna (1) operates in the X-band.

10. The artificial neural network monopulse angle measurement system applied to feed mismatched antenna according to claim 9, characterized in that: The main reflection surface diameter of the single pulse receiving antenna (1) is 2.4 meters.

Images