CN118914995A - SuperDARN radar elevation angle correction method based on virtual altitude model - Google Patents

Abstract

The invention discloses a SuperDARN radar elevation angle correction method based on a virtual altitude model, which comprises the following steps: step 1, acquiring radar observation data of a Zhongshan station SuperDARN; step 2, screening out ground scattering echo data of each day by using speed and spectrum width conditions, and carrying out visual analysis on the data; step 3, drawing a slope distance-elevation angle distribution characteristic diagram of the screened echo by using Matlab; step 4, drawing a virtual height model, and continuously adjusting the elevation angle to enable the original ground scattered echo elevation angle data received by the radar to better accord with the virtual height model; and step 5, introducing the RMSE of the elevation angle as a numerical analysis criterion for judging and screening the correction factors. The invention provides a numerical analysis algorithm for estimating radar correction factors by taking RMSE of elevation angle as a judgment criterion based on a virtual height model through the slope distance-elevation angle distribution characteristic of ground scattered echo, and improves the reliability of the elevation angle, thereby obtaining the accurate positioning information of a detection target.

Description

A SuperDARN radar elevation angle correction method based on virtual altitude model

Technical Field

The invention belongs to the technical field of radar measurement and positioning, and in particular relates to a SuperDARN radar elevation angle correction method based on a virtual altitude model.

Background Art

In radar positioning, the elevation angle refers to the pitch angle of the radar system or sensor relative to the horizon, and is one of the important parameters of the SuperDARN radar. The elevation angle can determine the range and position of the target that the radar system can detect. By adjusting the elevation angle, the radar system can expand its observation range and detect targets farther away. At the same time, the elevation angle can also locate the aerial height information of the target and achieve three-dimensional positioning capabilities. Choosing a reasonable elevation angle can maximize the measurement accuracy of the distance and avoid blind spots caused by close-range measurements.

Since it was put into use, the SuperDARN radar has achieved great success in the research of the magnetosphere, ionosphere, and middle and upper layers. In recent years, the technology of using the echo characteristics of the SuperDARN high-frequency radar to achieve elevation correction and ionospheric parameter inversion has been continuously improved and developed. The SuperDARN radar is mainly used to study the dynamics of the ionosphere and magnetosphere in the polar and mid-latitude regions, and to detect the backscattering of irregular plasma density bodies in the region and the ground backscattering through the ionospheric mirror reflection. Through international data sharing, these movement and distribution information has great application value in the fields of communication, navigation, remote sensing, etc. It also plays an important role in studying the coupling process of the solar wind-magnetosphere-ionosphere-middle and upper atmosphere.

Elevation angle is one of the important data in SuperDARN radar detection. SuperDARN radar is equipped with a main array and an interferometer array for elevation angle measurement. By calculating the phase difference between the two spatially separated arrays and combining the radar's own configuration information, such as radar operating frequency, beam direction, azimuth, etc., the elevation angle of the received echo is estimated. When the HF radio signal emitted by the radar passes through the ionosphere, a multi-hop echo will be generated. The echo is received by the radar again after backscattering. The received scattering types include ground scattering, ionospheric scattering, sea surface scattering, etc., and ionospheric scattering also needs to distinguish echoes from different areas, so the elevation angle is introduced to determine the type of received echo. Relying on the elevation angle information, the propagation mode of high-frequency wireless signals in the ionosphere can be determined, thereby obtaining accurate positioning information of the detection target. At the same time, a reliable elevation angle can determine the virtual height of the target, which is crucial for estimating ionospheric parameters.

Since the first SuperDARN radar was put into use in 1983, elevation angle measurements have been carried out. However, due to the different lengths of the electrical paths between the points where the return signals are received by the main and secondary arrays, the difference in propagation time between the signals propagating along the two different paths (also called the correction factor, Δt _cor ) is difficult to correct, and elevation angle data has been difficult to be effectively used. There are currently two solutions to the transmission delay problem. The first is to measure the delay of the test signal between the radar cable and the equipment, but it is restricted by the radar's geographical location, radar operating frequency, temperature and other conditions. These conditions cannot be completely controlled within a controllable range, so they are not taken into consideration. The other method abandons the influence of the radar hardware itself on the measurement, uses elevation angle data or phase difference data, and continuously adjusts the delay Δt _cor between the signal and the cable to make the calculated elevation angle meet the expectations of a specific propagation mode or a specific scattering position. For example, the elevation angle should decrease with the increase of the slant range, but the Δt _cor obtained by intuitive analysis is not accurate.

Based on the above status quo and the existing technology, it is difficult to determine the correction factor of the elevation angle. The present invention proposes a SuperDARN radar elevation angle correction method based on a virtual altitude model.

Summary of the invention

In order to improve the situation that the radar elevation correction factor is difficult to determine, the present invention proposes a SuperDARN radar elevation correction method based on a virtual altitude model. Based on the virtual altitude model, the present invention proposes a numerical analysis algorithm for estimating the radar correction factor by using the slant range-elevation distribution characteristics of the ground scattered echo and the RMSE of the elevation angle as the judgment criterion, thereby improving the reliability of the elevation angle and obtaining the precise positioning information of the detected target.

The present invention adopts the following technical scheme:

A SuperDARN radar elevation angle correction method based on a virtual altitude model, the steps comprising:

Step 1, obtain the elevation angle data detected by SuperDARN radar, pre-process it, and delete abnormal values and invalid values;

Step 2: Use Doppler velocity and Doppler width conditions to filter out the daily ground scatter echo data and perform a visual analysis of the data.

Step 3, extracting the slant range-elevation angle distribution characteristics of the ground scattered echo of Zhongshan Station;

Step 4: Use the proposed elevation correction algorithm to draw the elevation-slant range peak fitting curve, and continuously adjust the elevation angle to make the elevation angle data of the original ground scattering echo received by the radar better conform to the virtual height model;

Step 5: Introduce the RMSE of elevation angle as the numerical analysis criterion for judging and selecting correction factors.

Preferably, the preprocessing process in step 1 is as follows: data with elevation angles less than or equal to 0 in the extracted SuperDARN radar data are eliminated.

Preferably, the data screening process in step 2 is as follows:

Step 2.1: Filter the ground scattered echo according to the Doppler velocity (V) and Doppler spectrum width (W) received by the SuperDARN radar. The significant characteristics of the ground scattered echo are low Doppler velocity and low Doppler spectrum width. The echo conditions meet the following conditions:

|V|,|W|<50m/s

g(|V|,|W|)＝|V|-(V _max -(V _max /W _max )|W|)<0

Wherein, V _max =30 m/s, W _max =90 m/s are constants empirically derived from the SuperDARN data analysis software.

Step 2.2: After filtering out the daily ground scatter echo data using speed and spectrum width conditions, the data needs to be visualized for analysis. That is, the required data should conform to the theoretical correlation between the slant range of the ground scatter echo and the elevation angle. The expected pattern in the low-angle mode should be that the elevation angle gradually decreases with increasing distance until it approaches zero, and the virtual height should remain almost constant, so as to filter out a suitable data set.

Preferably, the extraction process of the slant range-elevation angle distribution characteristics in step 3 is as follows: use the screened elevation angle and slant range data to make an elevation angle-slant range density map. The virtual height density distribution of the ground scattered echo is close to the Gaussian distribution. The peak height in each range is estimated by fitting a Gaussian function to the distribution curve, and its peak corresponds to the maximum value of the elevation angle density under the slant range. Within the slant range where the ground scattered echo appears, each slant range corresponds to a Gaussian fitting model of virtual height density. According to the assumption of the expected pattern, the virtual height corresponding to the peak value of these Gaussian fits should remain constant with the change of slant range. Among them, the formula for Gaussian fitting is as follows:

Where α is the amplitude, b is the central abscissa, and c is the standard deviation, which is related to the peak width.

Preferably, the elevation correction algorithm in step 4 has the following steps: the maximum elevation angle at each slant distance is calculated in step 3, and a vector pair is formed with the slant distance in one-to-one correspondence, which is used as an input parameter for fitting the virtual height model, wherein the elevation angle expression of the virtual height formula is used as a custom formula of the fitting function:

In the formula, slant range r is used as the independent variable and h is used as the coefficient to be determined. According to different input vector pairs, h will be automatically determined when the best fit is achieved. Adjust the elevation angle and observe the matching between the elevation angle data of the original ground scatter echo received by the radar and the virtual height model.

Preferably, the calculation process of RMSE in step 5 is as follows: It is found from the observation result graph that the elevation angle data of the original ground scattering echo received by the radar cannot match the virtual height model well, and there is a deviation between the actual and fitted elevation angles. In order to describe this deviation value and use this deviation value as a numerical analysis criterion for judging and screening the correction factor, this paper introduces RMSE. The calculation formula of RMSE is:

In the formula, V _cor represents the RMSE value between the actual elevation angle after adding the correction factor each time and the elevation angle calculated by fitting the virtual height, n represents the distance threshold number with ground scattering echo characteristics, E _i represents the peak value of the actual elevation angle distribution at each slant range, and E _fit represents the elevation angle value calculated by fitting the virtual height formula. The smallest RMSE is selected from the obtained results, and the difference between the corresponding elevation angles is the correction factor of the radar estimated in this time period.

Compared with the prior art, the beneficial technical effects of the present invention are:

This paper proposes for the first time a new method for estimating the correction factor using the root mean square error of the elevation angle within different slant ranges, and combines the virtual height model with numerical analysis to improve the accuracy to a certain extent. Moreover, the algorithm has good scalability and for the first time gives the range change of the correction factor of different beams of the SuperDARN radar within three years, which provides a basis for subsequent research on the relationship between the beam and the correction factor.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of the overall steps of the present invention.

FIG. 2 is a schematic diagram of a specific process of a preferred embodiment of the present invention.

FIG. 3 is a schematic diagram of the results before and after correction of the present invention.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

This implementation provides a SuperDARN radar elevation correction method based on a virtual altitude model, as shown in Figure 1. This embodiment mainly includes six steps, namely data acquisition, data preprocessing, extraction of slant range-elevation distribution characteristics, virtual altitude model fitting, dynamic adjustment of elevation, and calculation of RMSE. Based on a virtual altitude model, the present invention proposes a numerical analysis algorithm that estimates the radar correction factor by using the slant range-elevation distribution characteristics of ground scattered echoes and the RMSE of the virtual altitude as a judgment criterion. The specific process of a preferred embodiment is shown in Figure 2, and the steps are as follows:

Step 1: Obtain SuperDARN radar observation data.

Since the data received by the radar itself will be affected by many factors, in this embodiment, data near the peak year of solar activity (2014) are selected. During this period, the data received by the radar will be more than usual. The three-year data can reflect the overall situation of the radar correction factor of Zhongshan Station to a certain extent.

Specifically, download the compressed rawacf data file of Zhongshan Station sampled with a 2-hour resolution from 2013 to 2015 from the server. On Linux, decompress, convert, and merge the data to convert the compressed rawacf data file into a fitacf file with daily resolution. Extract the required elevation angle, range threshold, beam, and other data;

Step 2: The Doppler velocity (V) and Doppler spectrum width (W) received by the SuperDARN radar can be used to screen the ground scattered echo.

The echo condition is met:

|V|,|W|<50m/s

g(|V|,|W|)＝|V|-(V _max -(V _max /Wx _ax )|W|)<0

Wherein, in this embodiment, V _max =30 m/s, W _max =90 m/s are constants empirically derived by SuperDARN data analysis software.

It should be noted that the high-frequency radar of Zhongshan Station can receive at least four different backscattered echo signals: ionospheric irregularities, meteor trail echoes, polar mid-layer summer echoes, and ground or sea surface echoes. Since the high-frequency radar of Zhongshan Station was built and put into use in 2010, it has been running stably and reliably so far, so the method proposed in this embodiment is still applicable.

Furthermore, after using the above conditions to filter out the daily ground scatter echo data, the data needs to be visualized and analyzed, that is, the required data should conform to the theoretical correlation between the slant range of the ground scatter echo and the elevation angle. Its expected pattern in the low-angle mode should be that the elevation angle gradually decreases with increasing distance until it approaches zero, and the virtual height should remain almost constant.

It is understandable that in this embodiment, the low angle mode refers to the situation where the observation or measurement angle is small. In scenes such as terrain, photography, satellite observation or radio signal transmission, the low angle mode means observing or measuring from a perspective close to the ground.

Step 3: Draw a virtual height probability curve based on the virtual height density distribution, and estimate the peak height in each range by fitting a Gaussian function to the distribution curve, so as to obtain the maximum value of the elevation density under the slant range corresponding to the peak, that is, the slant range-elevation distribution characteristics. Specifically, use the screened elevation and slant range data to make an elevation-slant range density map. The virtual height density distribution of the ground scattered echo of Zhongshan Station is close to the Gaussian distribution, and the Gaussian fitting formula is as follows:

Where α is the amplitude, b is the central abscissa, and c is the standard deviation, which is related to the peak width.

The peak height in each range is estimated by fitting a Gaussian function to the distribution curve, and its peak corresponds to the maximum value of the elevation density at the slant range. Within the slant range where the ground scattered echo appears, each slant range corresponds to a Gaussian fitting model of virtual height density. According to the assumption of the expected pattern, the virtual height corresponding to the peak of these Gaussian fittings should remain constant with the change of slant range.

Step 4: Fit the elevation peak data, and draw a slant range-elevation density map based on the slant range-elevation data. In this embodiment, the maximum elevation angle at each slant range is calculated with an elevation resolution of 1°, and a vector pair is formed with the slant range in one-to-one correspondence, which is used as the input parameter for fitting the virtual height model, wherein the elevation expression of the virtual height formula is used as the custom formula of the fitting function:

In the formula, the slant range r is used as the independent variable and h is used as the coefficient to be determined. According to different input vector pairs, h will be automatically determined when the best fit is achieved. Adjust the elevation angle and observe the matching between the elevation angle data of the original ground scatter echo received by the radar and the virtual height model, so that the elevation angle distribution peak can better match the echo change trend determined by the virtual height model.

Step 5: Calculate RMSE. The elevation angle data of the original ground scatter echo received by the radar cannot match the virtual height model very well. There is a deviation between the actual and fitted elevation angles. In order to describe this deviation value and use this deviation value as a numerical analysis criterion for judging and screening the correction factor, this paper introduces RMSE. The calculation formula of RMSE is:

V _cor represents the RMSE value between the actual elevation angle after adding the correction factor each time and the elevation angle calculated by fitting the virtual height, n represents the distance threshold number with ground scattering echo characteristics, E _i represents the peak value of the actual elevation angle distribution at each slant range, and E _fit represents the elevation angle value calculated by fitting the virtual height formula. The smallest RMSE is selected from the obtained results, and the difference between the corresponding elevation angles is the correction factor of the radar estimated in this time period.

It is understandable that elevation angle is one of the important data in SuperDARN radar detection. Due to the different lengths of the electrical paths between the correlation points of the main and secondary arrays receiving the return signals, the difference in propagation time between the signals propagating along the two different paths (also called the correction factor, Δt _cor ) is difficult to correct, and elevation angle data has been difficult to be effectively used.

Therefore, based on the virtual height model, this paper adopts a visual analysis method that uses the ground backscatter echo to estimate the correction factor, and uses the RMSE of the elevation angle as the judgment criterion to estimate the numerical analysis algorithm of the radar correction factor, which feeds back a compensation to the elevation angle data.

Furthermore, this embodiment uses the virtual height obtained after elevation correction under different annual accumulation days as the verification premise, and its value changes in real time, and uses the original phase difference data and the virtual height corrected based on the phase deviation model to verify the estimated correction factor. The echo slant range-phase deviation change diagram should be close to the theoretical model curve under the virtual height, and the correction factor value obtained using the phase deviation model is also close to the value obtained in the previous article.

The present invention proposes an elevation correction algorithm for SuperDARN radar based on a virtual altitude model, including: downloading SuperDARN radar source data files from 2013 to 2015 from a server to construct a data set; screening the data set according to the theoretical correlation between the characteristics of the surface ground echo, the echo slant range and the elevation angle; using Matlab to make a slant range-elevation density distribution diagram of the ground scattered echo of Zhongshan Station; fitting the elevation peak data, calculating the maximum elevation value under each slant range, and forming a vector pair with the slant range in one-to-one correspondence, using it as an input parameter for fitting the virtual altitude model, and automatically determining the virtual height at the best fitting effect according to different input directions; dynamically adjusting the elevation value, calculating the RMSE after each adjustment, selecting the minimum RMSE, and the difference between the corresponding elevation angles is the correction factor of the radar estimated in the time period. The present invention uses a virtual altitude model, through the slant range-elevation distribution characteristics of the ground scattered echo, based on the SuperDARN radar elevation data, to achieve elevation correction.

The preferred embodiments and principles of the present invention are described in detail above. For those skilled in the art, according to the ideas provided by the present invention, there may be changes in the specific implementation methods, and these changes should also be regarded as the protection scope of the present invention.

Claims (10)

1. A SuperDARN radar elevation angle correction method based on a virtual altitude model, characterized in that it comprises the following steps: Step 1: Obtain SuperDARN radar observation data; Step 2: After filtering out the daily ground scatter echo data using velocity and spectrum width conditions, secondary screening is performed by visually analyzing the data; Step 3, extracting the slant range-elevation angle distribution characteristics of the ground scattered echo; Step 4: Use the elevation correction algorithm to draw the elevation-slant range peak fitting curve, dynamically analyze the elevation angle, and make the original ground scattering echo elevation angle data received by the radar suitable for the input requirements of the virtual height model; Step 5: Introduce the RMSE of elevation angle as the numerical analysis criterion for judging and selecting correction factors. 2. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 1, characterized in that the SuperDARN radar observation data includes required elevation angle, Doppler velocity and Doppler bandwidth. 3. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 1 or 2, characterized in that in step 2, the screening method is: The Doppler velocity V and Doppler spectrum width W received by the SuperDARN radar can be used to screen ground scattered echoes. The echo conditions meet the following conditions: |V|,|W|<50m/s g(|V|,|W|)＝|V|-(V _max -(W _max /W _max )|W|)<0 Where V _max is the maximum Doppler velocity and W _max is the maximum Doppler spectrum width. 4. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 3, characterized in that in the step 2, the secondary screening method is: the expected mode of the screened data in the low angle mode is that the elevation angle gradually decreases with increasing distance until it tends to zero, and the virtual altitude should be kept almost constant. 5. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 4, characterized in that in the step 3, the slant range-elevation angle distribution feature extraction method is as follows: After screening the data, the elevation angle-slant range characteristic diagram was made using Matlab simulation software, and the virtual height density distribution of the echo was obtained according to the Gaussian distribution. The Gaussian fitting formula is as follows: Where α is the amplitude, b is the central abscissa, and c is the standard deviation, which is related to the peak width. 6. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 5, characterized in that, within the slant range in which the ground scattered echo appears, each slant range corresponds to a Gaussian fitting model of a virtual altitude density. 7. A SuperDARN radar elevation angle correction method based on a virtual altitude model as claimed in claim 5, characterized in that the range of the slant range is: Slope distance r = 180 + (n-1) * 45 Where n is the range gate of the SuperDARN radar. 8. A SuperDARN radar elevation angle correction method based on a virtual altitude model as claimed in claim 7, characterized in that in step 4, the elevation angle correction algorithm process steps are as follows: The slant range-elevation angle density map is drawn based on the slant range-elevation angle data. The maximum elevation angle at each slant range is calculated with the elevation angle resolution, and a vector pair is formed with the slant range in one-to-one correspondence. It is used as the input parameter of the virtual height model fitting. The elevation angle expression of the virtual height formula is used as the custom formula of the fitting function as follows: The slope distance r is taken as the independent variable and h is taken as the coefficient to be determined; According to different input vector pairs, the unknown coefficient h is determined according to the fitting effect, and by adjusting the elevation angle, the matching situation between the elevation angle data of the original ground scattering echo received by the radar and the virtual height model is observed, so that the elevation angle distribution peak can better match the echo change trend determined by the virtual height model. 9. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 8, characterized in that in the step 5, the calculation process of RMSE is as follows: Comparing the fitting distribution before and after correction, the elevation angle data of the original ground scattering echo received by the radar cannot match the virtual height model very well. There is a deviation between the actual and fitted elevation angles. This deviation value is used as the numerical analysis criterion for judging and selecting the correction factor. The calculation formula of RMSE is: Wherein, V _cor represents the RMSE value between the actual elevation angle after adding the correction factor each time and the elevation angle calculated by fitting the virtual height, n represents the distance threshold number with ground scattering echo characteristics, E _i represents the peak value of the actual elevation angle distribution at each slant range, and E _fit represents the elevation angle value calculated by fitting the virtual height formula. 10. A SuperDARN radar elevation correction method based on a virtual altitude model as claimed in claim 9, characterized in that in the step 5, the minimum RMSE is obtained by calculation, and the difference between the corresponding elevation angles is the correction factor of the radar estimated in the time period.