Method for measuring millimeter wave antenna pointing error by using spreading source

Abstract

The invention provides a method for measuring a millimeter wave antenna pointing error by using a spread source, which comprises the following steps: scanning a selected calibration source by using a millimeter wave antenna of a telescope, and recording to obtain calibration data, the calibration source being a spread source; and processing the calibration data from the spread source through a deconvolution algorithm, and obtaining the pointing error angles of the current azimuth axis and pitch axis according to a deconvolution result. According to the method for measuring the pointing error of the millimeter wave antenna by using the spread source, through the deconvolution algorithm and the noise suppression technology, the problem of insufficient precision in pointing calibration of an existing millimeter wave radio telescope is effectively solved, the calibration precision of the radio telescope is remarkably improved, and the development of high-frequency-band radio astronomical observation is promoted.

Description

Technical Field

[0001] The invention belongs to the field of radio astronomy and relates to a method for measuring the pointing error of a millimeter wave antenna by utilizing a spread source. Background Art

[0002] Existing radio telescope pointing calibration technologies mainly include the cross calibration method and the five-point calibration method, which have been widely used in the field of astronomical observation to ensure the pointing accuracy and data quality of the telescope. They detect the deviation between the actual pointing of the telescope and the theoretical position by scanning known celestial bodies or signal sources, and correct it. However, with the continuous increase in the observation frequency of radio telescopes, especially in observations in high-frequency bands (such as the W band), the existing technologies have shown some limitations and shortcomings in their applications.

[0003] The cross calibration method determines the pointing deviation by scanning the radio telescope in both the horizontal and vertical directions and detecting the maximum signal intensity. Although this method is simple to operate, requires fewer steps, and is suitable for quickly estimating errors, its accuracy is limited in some scenarios. Especially when observing weaker celestial bodies, the signal intensity fluctuates less, resulting in insufficient accuracy in the calculated error. In addition, when the initial pointing error of the telescope is large, the accuracy of the cross calibration method may not be sufficient to effectively cope with observation tasks that require high precision.

[0004] The five-point calibration method provides a more accurate estimation of pointing error than the cross calibration method by measuring at five different points around the target celestial body. This method performs well in high-precision scenarios, but its operation is complex, the measurement time is long, and it requires analysis of multiple measurement data, which increases the time and resource costs in use. In addition, the five-point method is mainly suitable for point source calibration, and when extended sources (such as the sun, moon, etc.) are required for calibration, the existing five-point method shows obvious shortcomings. Due to the small change in signal strength between the center point and the edge point, it is difficult for the five-point calibration method to accurately identify the telescope pointing error.

[0005] As the observation frequency of radio telescopes increases, especially in the W-band and higher frequency bands, the design of radio telescopes faces new challenges. High-frequency observations require telescopes to have extremely high surface accuracy, because higher frequencies correspond to shorter wavelengths, and any tiny surface error will significantly affect the observation accuracy. However, in order to meet this surface accuracy requirement, the aperture of the telescope is often not designed to be very large. This limitation directly leads to a smaller effective aperture of the antenna, which reduces the system equivalent flux density (SEFD). In other words, the sensitivity of the telescope is limited, making it more difficult to effectively capture weak radio signals.

[0006] In addition, as the frequency increases, the radio source radiation of celestial bodies (such as blackbody radiation or thermal radiation) is greatly reduced. This makes the radio source weaker in the high-frequency band. Due to the small antenna aperture and low SEFD, traditional point source calibration targets (such as compact galaxies, active galactic nuclei or quasars) are difficult to meet the pointing calibration needs of high-frequency telescopes. Therefore, in these frequency bands, the calibration source often chooses extended sources with a larger viewing angle such as the sun and the moon. However, the existing cross calibration method and five-point calibration method are mainly designed for point sources, and the use effect of these extended sources is not good.

[0007] Specifically, when the cross calibration method uses a spread source for pointing calibration, the result is usually a convolution image of the celestial body outline and the antenna aperture field, rather than the actual beam pattern of the antenna. This convolution effect greatly reduces the calibration accuracy. The five-point calibration method is more problematic when facing a spread source. Since the signal strength in the center of the spread source varies little, the measurement error becomes blurred, and the angle of the center point may fluctuate by as much as 0.5 degrees, thus affecting the accuracy of the calibration.

[0008] In summary, with the increase of observation frequency, traditional pointing calibration methods, such as the cross calibration method and the five-point calibration method, face the following technical difficulties: First, the radiation intensity of commonly used point sources (such as extragalactic radio sources) is significantly reduced in high frequency bands, and cannot provide sufficient signal strength, resulting in the inability to be used as calibration sources for accurate pointing calibration. Therefore, observations usually rely on the calibration of extended sources (such as the sun, the moon, etc.). Secondly, as the frequency of radio telescopes increases, the panel accuracy requirements greatly increase. In radio telescope observations in millimeter wave and higher frequency bands (such as the W band), the accuracy requirements of the telescope panel are extremely high, which limits the size of the aperture and, in turn, the system equivalent flux density (SEFD). This makes it difficult to obtain high-precision results when using traditional methods for pointing calibration. Finally, the traditional cross calibration method and the five-point calibration method have great limitations when dealing with extended sources. The traditional method measures the calibration error through point sources, and when using extended sources, the calibration accuracy is not high due to the convolution effect between the contour of the extended source and the antenna aperture field. Due to the lack of spread source processing algorithms, the traditional cross calibration method cannot effectively distinguish the influence of the celestial body contour because the measured deviation angle is unable to effectively distinguish the influence of the spread source contour, so that the measured pointing error is affected by the spread source contour, resulting in the inability to obtain the true antenna beam pattern, thus affecting the accuracy of the pointing calibration.

[0009] Therefore, the existing cross calibration method and five-point calibration method have significant limitations in high-frequency band and extended source calibration, and cannot meet the current needs of high-precision observation. This urgently requires the development of high-precision calibration technology that can handle extended sources to address the deficiencies in existing technologies and improve the calibration capabilities of radio telescopes. Summary of the invention

[0010] The object of the present invention is to provide a method for measuring the pointing error of a millimeter wave antenna by using an extended source, so as to improve the accuracy of pointing calibration by using an extended source and avoid the pointing error being affected by the extended source profile.

[0011] In order to achieve the above object, the present invention provides a method for measuring the pointing error of a millimeter wave antenna by using a spread source, comprising:

[0012] S1: Scanning a selected calibration source using a millimeter wave antenna of a telescope and recording calibration data, wherein the calibration source is a spread source;

[0013] S2: Process the calibration data from the display source through the deconvolution algorithm, and obtain the current pointing error angles of the azimuth axis and the pitch axis according to the deconvolution result.

[0014] In the step S1, the scanning method used for scanning includes one of a cross scanning method, a butterfly scanning method, and an OTF imaging scanning method.

[0015] The scanning method used for scanning is a cross scanning method; the step S1 specifically includes:

[0016] S11: Input several control parameters;

[0017] S12: driving the millimeter wave antenna to track the spread source to ensure that the millimeter wave antenna points to the center of the spread source;

[0018] S13: Perform scanning, and record calibration data used to determine the pointing error of the antenna in real time during the scanning process.

[0019] When scanning, the scanning range includes a high-speed scanning area and a low-speed scanning area. The high-speed scanning area corresponds to a scanning area deviating from the spread source, and the low-speed scanning area corresponds to a scanning area pointing to the spread source or near the spread source.

[0020] The step S2 specifically includes:

[0021] S21: preprocessing the calibration data;

[0022] S22: Introduce the spread source model and discretize the spread source brightness value to distribute it in the azimuth angle offset δ AZ On the grid, it is recorded as the discrete value H(θ) of the source brightness;

[0023] S24: convert the preprocessed data from the time domain to the frequency domain using a fast Fourier transform to obtain a frequency domain scanning result Y(θ);

[0024] S25: performing a deconvolution operation based on the frequency domain scanning result Y(θ) and the discrete value H(θ) of the source brightness, and obtaining an ideal signal X(θ) after deconvolution processing;

[0025] S26: performing smoothing filtering on the ideal signal X(θ) after the deconvolution processing to obtain a filtered ideal signal X'(θ);

[0026] S27: Perform inverse Fourier transform on the filtered ideal signal X'(θ) to restore it to the time domain to obtain the deconvolution result X(t);

[0027] S28: Perform Gaussian model fitting on the deconvolution result X(t) to obtain the mean of the Gaussian model as the pointing error angles of the azimuth axis and the elevation axis.

[0028] The calibration data includes the azimuth angle θ of the antenna AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL , signal strength P and millisecond timestamp;

[0029] In step S21, the calibration data is preprocessed, specifically including: the azimuth angle θ of all scanning ranges AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL Grid processing is performed to obtain preprocessed data, and the signal strength P is normalized.

[0030] In step S22, the brightness values ​​of the source are discretized and distributed in the azimuth angle offset δ AZ When the grid is on the same plane, it also includes the azimuth angle offset δ AZ The zero point is defined as the geometric center of the source.

[0031] After step S22 and before step S24, the method further includes step S23: estimating noise using pre-processed data of the scanning area deviating from the spread source, and deducting the noise to obtain pre-processed data of the low-speed scanning area after noise deduction;

[0032] The step S23 specifically includes:

[0033] S231: For the pre-processed data of the pitch angle, the signal strength data P(θ) of the scanning area deviating from the spread source during the scanning process is removed as noise data;

[0034] S232: Perform linear fitting on the signal intensity data P(θ) of the scanning area that deviates from the spread source, and the fitting result is the signal intensity y(θ) of the noise. Then, the signal intensity y(θ) of the noise is deducted from the signal intensity P(θ) of the low-speed scanning area to obtain the signal intensity P'(θ) of the low-speed scanning area after noise deduction.

[0035] In step S25, the ideal signal X(θ) after deconvolution is:

[0036]

[0037] Among them, X(θ) is the ideal signal after deconvolution processing, Y(θ) is the frequency domain scanning result, is the transfer function of the solar model, which is the FFT transformation result of the discrete value H(θ) of the source brightness.

[0038] The present invention uses a method for measuring the pointing error of a millimeter-wave antenna using an extended source to solve the technical problem in the pointing calibration of a millimeter-wave radio telescope, especially when using an extended source (such as the sun or the moon) for high-frequency band observations, where the traditional calibration method cannot provide sufficient accuracy. The present invention processes the extended source data using a deconvolution algorithm, accurately measures the azimuth and elevation errors of the antenna, and significantly improves the accuracy of the pointing calibration.

[0039] Improve calibration accuracy: Experiments show that the present invention can improve the measurement accuracy of the pointing error angle to the sub-radian level through the deconvolution algorithm, which is much higher than the traditional cross calibration method and five-point calibration method. This is crucial for the observation needs of millimeter wave and above frequency bands, especially for the pointing calibration of radio telescopes with small antenna aperture and low SEFD, providing higher accuracy and reliability.

[0040] Effectiveness of extended source calibration: Compared with the traditional method that can only rely on point sources for calibration, the present invention adopts a deconvolution algorithm, which can use extended sources for pointing error measurement, and can effectively use extended sources such as the sun and the moon for high-precision pointing calibration. This broadens the range of available calibration sources, especially when observing with high-frequency radio telescopes, when the radiation of extragalactic radio sources is insufficient, efficient calibration can still be achieved through strong extended sources such as the sun.

[0041] Economic and social benefits: In practical applications, the present invention can reduce observation errors caused by inaccurate calibration, thereby improving the observation efficiency of radio telescopes and reducing observation costs. At the same time, this technology can be applied to millimeter-wave radio telescopes around the world, promoting technological progress in astronomical observations, especially in the fields of high-frequency cosmic microwave background radiation and planetary exploration.

[0042] In addition, the method of measuring the pointing error of a millimeter-wave antenna using a spread source of the present invention designs an antenna acceleration scanning strategy, which reduces the time consumed for antenna scanning while ensuring the accuracy of the scanning data.

[0043] In summary, the method of measuring the pointing error of millimeter-wave antenna using a spread source of the present invention effectively solves the problem of insufficient accuracy in the pointing calibration of existing millimeter-wave radio telescopes through a deconvolution algorithm and noise suppression technology, significantly improves the calibration accuracy of radio telescopes, and promotes the progress of high-frequency radio astronomy observations. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is an overall flow chart of a method for measuring the pointing error of a millimeter wave antenna by using a spread source according to the present invention.

[0045] Figure 2 It is a scanning principle diagram that performs incremental scanning along the azimuth axis and the elevation axis respectively on the basis of maintaining the continuous tracking of the antenna to the sun.

[0046] Figure 3A This is the measured data diagram for azimuth axis scanning; Figure 3B Comparison diagram of the antenna radiation pattern (blue) restored after deconvolution and the Gaussian fitting curve (red).

[0047] Figure 4A This is the measured data diagram for pitch axis scanning; Figure 4B Comparison diagram of the antenna radiation pattern (blue) restored after deconvolution and the Gaussian fitting curve (red). DETAILED DESCRIPTION

[0048] The present invention is further described below in conjunction with specific examples. It should be understood that the following examples are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0049] The method of measuring the pointing error of a millimeter wave antenna by using a spread source of the present invention is mainly used for the pointing calibration of a millimeter wave radio telescope before observation, and the goal is to measure the error angle of the current azimuth and pitch axis by using a spread source such as the sun and the moon. In this embodiment, the method of measuring the pointing error of a millimeter wave antenna by using a spread source is further improved on the basis of the existing cross calibration method, and is applicable to the measurement data of the cross calibration method.

[0050] The pointing calibration work done by the present invention mainly involves aligning the telescope itself (hardware, software and other structures) with a telescope coordinate system and another reference coordinate system ICRS (International Celestial Reference System) corresponding to the astronomical reference position to compensate for the error angle between the telescope coordinate system and the reference coordinate system. After using the error angle obtained by the present invention to compensate for the error angle between the telescope coordinate system and the reference coordinate system, the existing coordinate system of the telescope can be directly used for celestial body observation.

[0051] like Figure 1 As shown, a method for measuring the pointing error of a millimeter wave antenna by using a spread source of the present invention specifically includes:

[0052] Step S1: In the data collection stage, the millimeter wave antenna of the telescope is used to perform a cross scan on the selected calibration source and record the calibration data;

[0053] In other embodiments, the cross scanning method can be replaced by other types of continuous scanning methods, such as butterfly scanning method and OTF (on the fly) imaging scanning method. Although different scanning methods have different data processing methods, the idea of ​​deconvolution is the same. For continuous scanning, the influence of the spread source can be removed to make it the same as scanning point source, so that the original method can be used for processing.

[0054] Regardless of the scanning method used, the calibration data includes the antenna's azimuth angle, elevation angle, azimuth angle offset, elevation angle offset, signal strength, and millisecond timestamp.

[0055] The step S1 specifically includes:

[0056] Step S11: In order to ensure the accuracy of scanning, input several control parameters;

[0057] In this embodiment, the scanning method is a cross scanning method, wherein the scanning range includes a high-speed scanning area and a low-speed scanning area, such as Figure 2 As shown in the figure, the dotted part is the high-speed scanning area, which actually corresponds to the scanning area away from the spread source, and the solid part is the low-speed scanning area, which actually corresponds to the scanning area pointing to the spread source or near the spread source; the difference between the two is the speed of scanning. The input control parameters include: pitch scanning width B EL0 , instruction interval t0, low-speed scan width B EL , low speed scanning speed v s 、High and low magnification k v , the specific meaning of each control parameter is shown in Table 1.

[0058] That is to say, in this embodiment, the cross scanning method adopted is different from the existing cross scanning method. Conventional cross scanning has no speed constraints. In this embodiment, when performing cross scanning on the selected calibration source, the speed is constrained. Specifically, the low-speed scanning area inside and near the radio source is a slow scan to obtain more details; the high-speed scanning area outside the low-speed scanning width is a fast scan to quickly obtain statistical noise, as well as the noise change trend with the elevation angle during pitch axis scanning (since the noise change trend is linear, a fast scan is sufficient).

[0059] Table 1: Specific meaning of each control parameter

[0060] Step S12: driving the millimeter wave antenna to track the spread source to ensure that the millimeter wave antenna points to the center of the spread source;

[0061] Then, prepare to scan. Since this embodiment adopts cross scanning, the detailed trajectory description is shown in Figure 2 .

[0062] Step S13: perform scanning, and record calibration data used to determine the pointing error of the antenna in real time during the scanning process.

[0063] like Figure 2 As shown, at the beginning of the scan, the antenna is offset by a certain distance in the direction where the azimuth angle increases, and a positive azimuth scan result is obtained. Then, it is offset by a certain distance in the direction where the azimuth angle gradually decreases, and a negative azimuth scan result is obtained. At this point, the scan covers the positive and negative azimuth contours of the target, completing the horizontal scan of the azimuth. Next, the antenna is offset in the direction of the elevation angle, also covering the positive and negative elevation contours of the target, completing the vertical scan and forming a "cross" shaped scanning trajectory. The advantage of cross scanning is that it is fast. Cross scanning can be regarded as a way of controlling variables. The purpose is to take the error in a certain azimuth and elevation angle. After a quick scan in azimuth, take the maximum value; then quickly scan the elevation to take the maximum value. Figure 2 In the cross scanning method shown, AZ is the azimuth axis and EL is the elevation axis. Figure 2 The black solid line part is slow scanning, and the dotted line part is high-speed scanning. The specific speed value and slow speed range are determined by the input parameters.

[0064] In this embodiment, the calibration data used to determine the pointing error of the antenna includes the azimuth angle θ of the antenna. AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL , signal strength P and millisecond timestamp, the specific meanings of these calibration data are shown in Table 2. These calibration data will be used in subsequent data processing to determine the pointing error of the antenna.

[0065] Among them, the azimuth angle offset δ AZ , pitch angle offset δ ELSpecifically, it is determined in the following way: Since the antenna is controlled by software, the present application knows the current angle and command angle (the command value from the computer) based on the received signal and recorded signal of the servo motor. In addition, the current target angle can be calculated based on the known time. Among them, the antenna is controlled to rotate by the servo system, and the command angle currently executed by the servo system and the antenna is received by the servo system is the command angle. In addition, the current angle of the antenna monitored by the code disk is also transmitted to the servo system, so the angle at which the antenna is currently pointing recorded by the servo system is the current angle. In other words, the command angle and the current angle are both based on the numerical value of the servo system. At the same time, the command angle currently executed and the angle of the current antenna provided by the servo system are the command angle and the current angle in this article. The target angle refers to the ideal azimuth position of the source, which is the geometric center of the extended source. Therefore, the target angle can be obtained by theoretical calculation based on astronomical ephemeris and time.

[0066] Scan source, that is, control the position of the antenna according to Figure 2 The source is moving in the direction shown, but because the source is also moving, in actual engineering applications, in addition to the command angle, there is also a deviation angle, which is used to indicate how many degrees the antenna deviates from the source with high precision due to the delay. When the computer sends the command angle in actual antenna control, the servo system will gradually rotate to point to the command angle, which will cause a certain delay. Therefore, in order to avoid the error caused by this delay, the computer usually discretizes the angle of the source and sends it periodically as a deviation angle. It only needs to ensure that the instructions are sent in order. Then, record the current angle of the servo system observing the entire scanning process, the command angle currently executed, and the millisecond timestamp. Then, according to the time, celestial orbit / coordinates, calculate the position where the antenna should point to. The difference between the position where the antenna should currently point and the current angle of the antenna (that is, the angle at which the antenna actually points) is the actual angle offset at this moment. In addition, all devices used in the telescope control process are synchronized by a time server, which is why delays can be avoided in this way.

[0067] Therefore, the angle offset satisfies: Angle offset = command angle + deviation angle - target angle. (The command angle + deviation angle can be regarded as the angle at which the antenna actually points, and the target angle refers to the position at which the antenna should currently point).

[0068] Table 2: Meaning of calibration data

[0069] It should be noted that the azimuth angle θ here is AZ and pitch angle θ EL are the current angles of the antennas mentioned above.

[0070] Step S2: Data processing stage, that is, processing the calibration data from the spread source through a deconvolution algorithm to eliminate the influence of the celestial body contour, and obtaining the current pointing error angles of the azimuth axis and the pitch axis according to the deconvolution result.

[0071] In this embodiment, the sun is taken as an example of a display source.

[0072] The step S2 specifically includes:

[0073] Step S21: pre-processing the calibration data to facilitate subsequent data processing;

[0074] In this embodiment, the calibration data is preprocessed, specifically including: AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL Grid processing is performed to obtain preprocessed data (including angle offset θ-signal intensity P(θ)), and the signal intensity P is normalized.

[0075] Among them, the angle accuracy of grid processing is t0×v s In this embodiment, the cross scanning data includes four parts: AZ+, AZ-, EL+, and EL-. Here, AZ+ is taken as an example. For all the azimuth angles θ in the scanning range, AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL Grid processing is performed, including:

[0076] Step S211: using the azimuth angle θ collected over time AZ , pitch angle θ EL and signal strength P as dependent variables, with azimuth angle offset δ AZ As the independent variable, the linear interpolation method is used to integrate and obtain the grid data;

[0077] During interpolation, only the azimuth angle is offset by δ AZ is the independent variable, and the others are offset by the azimuth angle δ AZ After interpolation, the azimuth angle between adjacent grid points is offset by δ AZ The interval is t0×v s .

[0078] Therefore, after gridding, multidimensional data with discrete data of uniform step length of offset angle as independent variable can be obtained, where each grid point represents an azimuth angle offset δ AZ And its corresponding signal strength, azimuth angle, and elevation angle.

[0079] The normalization formula of the signal strength P adopts a conventional formula for normalization using a maximum value and a minimum value.

[0080] Where the angle offset θ is the azimuth angle θ AZ Or pitch angle θ EL .

[0081] Step S22: Introduce the spread source model and distribute the spread source brightness values ​​discretized in the azimuth angle offset δ AZ On the grid, it is recorded as the discrete value H(θ) of the source brightness;

[0082] Therefore, subsequent processing will be performed based on the center position of the source to eliminate errors caused by position deviation.

[0083] The source spreading model specifically includes the intensity distribution of the brightness of the source in the reference coordinate system, and the brightness value of the source is discretized and distributed in the azimuth angle offset δ AZ When the grid is on the same plane, it also includes the azimuth angle offset δ AZ The zero point is defined as the geometric center of the spread source. Since the time is known, the position of the geometric center of the spread source (i.e., the target angle) is known, and the brightness change from the geometric center position is the spread source model of the present invention. The specific physical unit of the spread source brightness value can be temperature (Kelvin) or power (mW), as long as it is a linear unit, not a dB type unit, because the calibration does not care about the absolute intensity.

[0084] In this embodiment, the display source is the sun, so the display source model is a sun model, and the input sun model is the intensity distribution of the sun's brightness.

[0085] The solar model can use private observation data (such as the observed discrete value of the solar brightness - the offset angle model data), or directly use a square wave instead. In this embodiment, the azimuth angle can be directly offset by δ AZ Substitute it into the solar model as an input parameter to obtain the corresponding solar brightness value. It can also be interpolated with the azimuth angle offset δ based on the observation data of the solar model. AZ The grid is aligned with the data granularity to obtain the discretized distribution of solar brightness values.

[0086] Next, we use mean filtering to reduce the effect of noise on deconvolution.

[0087] Step S23: using the pre-processed data of the scanning area deviating from the spread source to estimate the noise, and deducting the noise to obtain the pre-processed data of the low-speed scanning area after noise deduction;

[0088] In this embodiment, since the scanning speed is differentiated, the scanning area deviating from the spreading source actually refers to the high-speed scanning area.

[0089] Step S23 specifically includes:

[0090] Step S231: for the pre-processed data of the pitch angle, remove the signal strength data P(θ) of the high-speed scanning area during the scanning process as noise data;

[0091] Step S232: Linear fitting is performed on the signal strength data P(θ) of the high-speed scanning area (i.e., fitting is performed using the formula y(θ)=kθ+b), and the fitting result is the signal strength y(θ) of the noise. Then, the signal strength y(θ) of the noise is deducted from the signal strength P(θ) of the low-speed scanning area to obtain the signal strength P'(θ) of the low-speed scanning area after noise deduction. Thus, the influence of environmental noise during pitch scanning is deducted.

[0092] During pitch scanning, the change of elevation angle has a greater impact on the environmental noise, while during azimuth scanning, the change of pitch angle is not significant. The change of pitch axis leads to the change of noise, and the change of pitch axis in azimuth scanning can be ignored. Therefore, it is necessary to specifically deduct the environmental noise in the preprocessed data during pitch scanning. For the preprocessed data of azimuth angle, skip this step.

[0093] The fitting parameters obtained according to the formula y(θ)=kθ+b include k, b, and then the angle of the low-speed part is taken as θ. Then the pre-processed data of the low-speed scanning area after noise subtraction includes:

[0094] The signal intensity P'(θ) of the low-speed scanning area after noise subtraction = the signal intensity P(θ) of the low-speed scanning area - the signal intensity y(θ) of the noise.

[0095] Step S24: using fast Fourier transform (FFT) to convert the pre-processed data (angle offset θ-signal strength P(θ)) from the time domain (or space domain) to the frequency domain to obtain the frequency domain scanning result Y(θ);

[0096] Among them, for all scanning areas, the data processing method of fast Fourier transform (FFT) is the same. The pre-processed data (angle offset θ-signal strength P(θ)) includes signal strength, offset angle, corresponding current pitch angle, current azimuth angle and time, which is a sequence strung together by time slices, so the pre-processed data can be transformed into the frequency domain.

[0097] The frequency domain scan result Y(θ) is the frequency domain representation of the observed data after noise subtraction. The independent variable of the frequency domain scan result Y(θ) is the angle offset θ, and the dependent variable is the frequency domain representation of the signal strength. Therefore, in the frequency domain, the convolution operation is converted into multiplication, making the subsequent deconvolution operation more convenient.

[0098] Step S25: performing a deconvolution operation based on the frequency domain scanning result Y(θ) and the discrete value H(θ) of the source brightness to obtain an ideal signal X(θ) after deconvolution processing;

[0099] According to the deconvolution theorem, the ideal signal X(θ) after deconvolution is:

[0100]

[0101] Among them, X(θ) is the ideal signal after deconvolution processing, Y(θ) is the frequency domain scanning result, is the transfer function of the solar model, which is the FFT transformation result of the discrete value H(θ) of the source brightness.

[0102] Deconvolution is the core step, which is used to remove the signal spreading effect caused by the extended source (the sun) from the measured data. By using the previously acquired extended source model as the convolution kernel, the convolution effect can be reversed and the point spread function of the radio telescope can be restored.

[0103] Step S26: performing smoothing filtering on the ideal signal X(θ) after the deconvolution processing to obtain a filtered ideal signal X'(θ);

[0104] As the deconvolution formula shows, when H(θ) has a small value (noise), the noise will be amplified. Therefore, we need to perform smoothing filtering on the real and imaginary parts of the ideal signal X(θ) after deconvolution in the frequency domain.

[0105] Since the beam width of the actual point spread function should be close to the theoretical beam width of the antenna, in step S26, when performing the smoothing filter process, the filter window is set to a common factor smaller than the theoretical beam width of the antenna. The theoretical beam width of the antenna usually refers to HPBW (the main lobe beam width of the antenna, i.e., the width of the half power point in the antenna power pattern (HPBW, half power beam width), calculated as: HPBW = 1.02 * observation wavelength / aperture).

[0106] In addition, since the sampling rate of the data item signal strength is limited, the antenna offset angle corresponding to the sampling interval is the lower limit of the filter window. If the filter window exceeds the antenna offset angle corresponding to the sampling interval by too much, the data itself will have too large an interpolation characteristic. Therefore, in step S26, when performing smoothing filtering, the filter window is also set to be larger than the antenna offset angle corresponding to the actual sampling interval. In other words, it is recommended that the size of the filter window satisfy the theoretical beam width of the antenna> window> measured data sampling interval.

[0107] Step S27: Perform inverse Fourier transform on the filtered ideal signal X'(θ) to restore it to the time domain to obtain a deconvolution result X(t).

[0108] The deconvolution result X(t) includes the azimuth angle offset δ AZ and the pitch angle offset δ EL The corresponding signal strength data has the angle offset as the independent variable and the signal strength as the dependent variable. At this time, the signal strength data corresponding to the angle offset has eliminated the convolution effect of the spread source profile and can more accurately reflect the actual pointing error of the antenna.

[0109] Step S28: performing Gaussian model fitting on the deconvolution result X(t) to obtain the mean of the Gaussian model as the pointing error angle of the azimuth axis and the elevation axis.

[0110] The Gaussian model formula is as follows:

[0111]

[0112] Among them, A is the amplitude (peak value) of the function, μ is the mean (center position), σ is the standard deviation (determines the width of the distribution), x is the independent variable, and f(x) is the dependent variable. Therefore, the fitting can obtain three parameters A, μ and σ, and the obtained mean μ is the precise error angle.

[0113] In summary, the present invention can obtain the positive and negative error angles of the azimuth and elevation angles through the above deconvolution operation. These error data will be used to adjust the actual pointing direction of the antenna to ensure that the calibration accuracy of the telescope meets the observation requirements.

[0114] Experimental results:

[0115] In order to study and verify the relevant technologies of millimeter-wave antennas, a 5-meter-diameter millimeter-wave radio telescope was built in the Tianma Park of Shanghai Astronomical Observatory. Among them, millimeter-wave pointing calibration is one of the technologies that need to be studied most. The antenna system of the 5-meter millimeter-wave radio telescope adopts the Cassegrain antenna form, a high-rigidity azimuth-elevation full turntable mount, and a fully digital antenna controller. It is used to receive high-quality radio signals from natural and artificial celestial bodies, and the highest frequency covers the W band. The antenna subsystem is mainly composed of the antenna feed subsystem, the mechanical structure subsystem, the antenna control subsystem, etc.

[0116] We implemented this method on the antenna and the measured results are shown in the following example: Figure 3A and Figure 3B (azimuth axis), Figure 4A and Figure 4B Finally, at the position where the azimuth angle is 204.673 degrees and the pitch angle is 63.338 degrees, the azimuth axis angle deviation is -1.225 degrees and the pitch axis deviation is -0.156 degrees.

[0117] The above is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. The above embodiments of the present invention can also be modified in various ways. All simple, equivalent changes and modifications made according to the claims and the description of the present invention fall within the scope of protection of the claims of the present invention. The contents not described in detail in the present invention are all conventional technical contents.

Claims

1. A method for measuring the pointing error of a millimeter wave antenna using a spread source, characterized in that: include: Step S1: Scanning a selected calibration source using a millimeter wave antenna of a telescope to record calibration data, wherein the calibration source is a spread source; Step S2: Process the calibration data from the spread source by a deconvolution algorithm, and obtain the current pointing error angles of the azimuth axis and the elevation axis according to the deconvolution result.

2. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 1, characterized in that: In the step S1, the scanning method used for scanning includes one of a cross scanning method, a butterfly scanning method, and an OTF imaging scanning method.

3. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 2, characterized in that: The scanning method used for scanning is a cross scanning method; the step S1 specifically includes: Step S11: input several control parameters; Step S12: driving the millimeter wave antenna to track the spread source to ensure that the millimeter wave antenna points to the center of the spread source; Step S13: perform scanning, and record calibration data used to determine the pointing error of the antenna in real time during the scanning process.

4. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 1, characterized in that: When scanning, the scanning range includes a high-speed scanning area and a low-speed scanning area. The high-speed scanning area corresponds to a scanning area deviating from the spread source, and the low-speed scanning area corresponds to a scanning area pointing to the spread source or near the spread source.

5. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 1, characterized in that: The step S2 specifically includes: Step S21: pre-processing the calibration data; Step S22: Introduce the spread source model and distribute the spread source brightness values ​​discretized in the azimuth angle offset δ AZ On the grid, it is recorded as the discrete value H(θ) of the source brightness; Step S24: using fast Fourier transform to convert the pre-processed data from the time domain to the frequency domain to obtain a frequency domain scanning result Y(θ); Step S25: performing a deconvolution operation based on the frequency domain scanning result Y(θ) and the discrete value H(θ) of the source brightness to obtain an ideal signal X(θ) after deconvolution processing; Step S26: performing smoothing filtering on the ideal signal X(θ) after deconvolution processing to obtain a filtered ideal signal X'(θ); Step S27: Perform inverse Fourier transform on the filtered ideal signal X'(θ) to restore it to the time domain to obtain a deconvolution result X(t); Step S28: performing Gaussian model fitting on the deconvolution result X(t) to obtain the mean of the Gaussian model as the pointing error angle of the azimuth axis and the elevation axis.

6. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 5, characterized in that: The calibration data includes the azimuth angle θ of the antenna AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL , signal strength P and millisecond timestamp; In step S21, the calibration data is preprocessed, specifically including: the azimuth angle θ of all scanning ranges AZ , pitch angle θ EL , azimuth angle deviation δ AZ , pitch angle offset δ EL Grid processing is performed to obtain preprocessed data, and the signal strength P is normalized.

7. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 5, characterized in that: In step S22, the brightness values ​​of the source are discretized and distributed in the azimuth angle offset δ AZ When the grid is on the same plane, it also includes the azimuth angle offset δ AZ The zero point is defined as the geometric center of the source.

8. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 5, characterized in that: After step S22 and before step S24, the method further includes step S23: estimating the noise using the preprocessed data of the scanning area deviating from the spread source, and deducting the noise to obtain the preprocessed data of the low-speed scanning area after noise deduction.

9. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 8, characterized in that: The step S23 specifically includes: Step S231: for the pre-processed data of the pitch angle, the signal strength data P(θ) of the scanning area deviating from the spread source during the scanning process is removed as noise data; Step S232: Perform a linear fit on the signal strength data P(θ) of the scanning area that deviates from the spread source. The fitting result is the signal strength y(θ) of the noise. Then, the signal strength y(θ) of the noise is deducted from the signal strength P(θ) of the low-speed scanning area to obtain the signal strength P'(θ) of the low-speed scanning area after noise subtraction.

10. The method for measuring the pointing error of a millimeter wave antenna by using a spread source according to claim 5, characterized in that: In step S25, the ideal signal X(θ) after deconvolution is: Among them, X(θ) is the ideal signal after deconvolution processing, Y(θ) is the frequency domain scanning result, is the transfer function of the solar model, which is the FFT transformation result of the discrete value H(θ) of the source brightness.

Images