CN106597421A - Prediction model-based delay and delay rate fast search method applied to very long baseline antenna array - Google Patents

Abstract

The present invention discloses a prediction model-based delay and delay rate fast search method applied to a very long baseline antenna array. The method includes the steps of prediction value calculation, prediction model fitting and residual search. According to the delay and delay rate fast search method of the invention, the prediction model of delay and delay rates is introduced; the search range of the delay and delay rate is effectively limited; and therefore, the search efficiency of cross-correlation processing is greatly improved, and the delay and delay rates of each antenna in the very long baseline antenna array can be obtained fast, and support can be supported for subsequent antenna array data processing.

Description

Fast search method for delay and delay rate based on prediction model in remote antenna array

technical field

The invention relates to the technical field of deep space exploration, in particular to a fast search method for signal time delay and time delay rate based on a prediction model in a remote antenna array.

Background technique

With the continuous expansion of human deep space exploration activities, the flight distance of spacecraft is getting farther and farther, and the signals emitted by spacecraft are also getting weaker and weaker with the increase of distance. Due to the engineering limit of the caliber of the single-sided antenna and the technical indicators of the receiver system, it is necessary to use the antenna array technology to achieve the minimum signal-to-noise ratio required for receiving signals. The purpose of the antenna array is briefly summarized as follows: under the same transmission code rate, the antenna array technology can increase the receiving distance from the ground to the spacecraft; under the same receiving distance, the antenna array technology can improve the distance between the ground and the spacecraft. transmission code rate.

The off-site antenna array refers to the use of existing large-aperture antennas distributed in different areas to observe the spacecraft at the same time, and then record and synthesize the signals from each antenna through digital signal processing technology to achieve the purpose of improving the signal-to-noise ratio of the received signal. The advantage of this antenna array technology is that the existing antenna and receiving equipment can be used, and the antenna array can be realized only by adding back-end data processing equipment. The development period is short and the cost is low.

With the establishment of the Mars Exploration Project, my country's deep space exploration activities have also begun to expand from the moon to Mars. Since Mars is very far away from the earth, the signal attenuation is huge during space transmission. Compared with the moon, the distance of Mars increases by 1000 times, and the signal space attenuation increases by about 60dB. The detector signal received on the ground is very weak. If the existing single-antenna reception is used, it cannot meet the scientific data downlink requirements of the detector. If a single antenna with a larger caliber is newly built for data reception, due to the limitations of the single antenna engineering limit, development progress, and development funds, the entire project has great development risks.

Contents of the invention

(1) Technical problems to be solved

In view of this, the main purpose of the present invention is to provide a fast search method for time delay and time delay rate based on a prediction model in a remote antenna array, so as to quickly obtain the time delay and time delay rate values of each antenna in the remote antenna array.

(2) Technical solution

The present invention provides a fast search method for time delay and time delay rate based on a prediction model, which includes the following steps: Step 1: Preliminarily calculate the time delay prediction value between the observation target and the antenna according to the antenna position and the observation target position; step 2: Fit the least squares model to the calculated time delay prediction value to obtain a time-continuous time delay prediction model; Step 3: Use the obtained time delay prediction model to perform preliminary time delay and time delay on the observation data of the antenna array Rate compensation, and then use the cross-correlation method to carry out cross-correlation processing on the antenna data, and search for the residual between the real delay value and the delay prediction value.

In the above scheme, the step 1 includes: Step 101: read the observation target tracking file list (t, ra, dec) and the geographical coordinates ( _xi , y, _{zi )} of each antenna in the remote array, where (t, ra, dec) represent the right ascension and declination values of the observed target at time t; ( _xi , y _i , _zi ) represent the position of each antenna in the earth space Cartesian coordinate system, and the subscript i is used to distinguish different Antennas, if there are N antennas in the array, there are N groups ( _xi , y, _zi ), where _i is 0, 1...N-1; Step 102: Calculate the arrival time of the observed target according to the classical celestial delay calculation method Geometric delay predictions for antennas and reference points.

In the above scheme, the classic celestial time delay calculation method described in step 102 includes: Step A1: input the coordinates of the station and the position of the detector; step A2: correct the local coordinate error of the station caused by tides, plate movements, etc.; step A3: Convert the earth-centered earth coordinate system to the earth-centered celestial coordinate system; Step A4: Transform the earth-centered celestial coordinate system to the solar system barycentric celestial coordinate system through Lorentz transformation, calculate the geometric time delay in the solar system barycentric celestial coordinate system, and correct The bending error of the signal transmission path caused by the gravitational attraction of the solar system; Step A5: transform the barycentric celestial coordinate system of the solar system back to the earth-centered celestial coordinate system through the Lorentz transformation, and obtain the predicted value of the geometric delay.

In the above scheme, the step 2 includes: Step 201: read in the predicted value τ _i (t _j ) of the geometric delay of each antenna with respect to the geocentric reference point, and according to the motion characteristics of the observed target, the above predicted value The list is divided into M segments, and M is a natural number; step 202: use the least square method to perform polynomial fitting on the above-mentioned section of predicted values to obtain a polynomial model y; step 203: evaluate the fitting accuracy of the polynomial model y, and the fitting accuracy meets the requirements. Get the delay prediction model.

In the above solution, the least squares method described in step 202 uses a 5th order polynomial for least squares fitting.

In the above scheme, the step 203 includes: according to the formula A standard accuracy is calculated, f is the frequency of the observed signal, in Hz; the fitting error is calculated by the formula Obtained, where y _j represents the polynomial model y when X=j time value, τ ₀ (j), τ ₁ (j) correspond to the geometric time delay calculation value of Miyun station and Kunming station j time in the geometric time delay calculation result table , n represents the number of predicted values for each segment; the fitting error RMS needs to be less than the standard accuracy; if the accuracy meets the requirements, the group of polynomial coefficients can be used as the delay prediction model of the antenna, if the fitting accuracy does not meet the requirements, Return to step 201, adjust the segment or increase the polynomial order, and re-fit until the fitting accuracy meets the requirements.

In the above solution, the residual difference between the true delay value and the predicted delay value in step 3 includes residual delay and residual delay rate.

In the above scheme, the step 3 includes: step 301: select the antenna with the largest aperture in the antenna array as the reference antenna, select a delay prediction model value of the antenna array obtained in step 203 minus the prediction model value of the reference antenna, and obtain The delay difference prediction model y(i, t _j ) of other antennas relative to the reference antenna; Step 302: Read in the reference antenna observation data _OR (t _j ) and the other antenna observation data O _i (t _j ) in the array , remove the DC component of the observed data to obtain O ^a _R (t _j ), O ^a _i (t _j ); Step 303: Take 2 ⁿ data from the O ^a _R (t _j ) data, and use the time The delay difference prediction model y(i, t _j ) performs delay compensation and delay rate compensation on the observed data O ^a _i (t _j ), and obtains 2 ⁿ data after compensation [O ^a _i (t _j -y( i, t _j ))]; Step 304: Carry out fast Fourier transform to O ^a _R (t _j ), [O ^a _i (t _j -y (i, t _j ))] respectively, and perform cross-correlation to obtain cross-correlation Correlation spectrum fft(OR ^a (t _j ))× _fft ^* ([O ^a _i ( _{t j} -y(i, t _j ))]); find cross-correlation spectrum fft(OR ^a (t _j ))× fft ^* ([O ^a _i (t _j -y(i, t _j ))]), then unwrap the phase value, and then use the least squares method to fit a linear function on the phase curve to get The residual time delay Δτ ¹ of the 2 ⁿ data; step 305: remove the 2 ⁿ data, and perform the calculation in the above step 304 to obtain the residual time delay Δτ ² of the next 2 ⁿ data, and calculate the time delay rate Step 306: The residual delay Δτ ¹ and the delay rate obtained from the above search Added to the corresponding delay difference prediction model y(i, t _j ), a real time delay model that can be used for subsequent antenna array data processing can be obtained.

(3) Beneficial effects

The time delay and time delay rate fast search method provided by the present invention have positive effects in that:

The prediction model of time delay and time delay rate is introduced, which effectively limits the search range of time delay and time delay rate, which greatly improves the search efficiency of cross-correlation processing, so that the time delay of each antenna in the remote antenna array can be quickly obtained and delay rate value to provide support for subsequent antenna array data processing.

Description of drawings

Fig. 1 is a schematic diagram of reception of antenna array signals in different places;

Fig. 2 is the flow chart of the rapid search method for time delay and time delay rate based on the prediction model in the remote antenna array of the present invention;

Figure 3 is a flow chart of classic celestial delay calculation;

Fig. 4 is the model calculation value of a kind of specific embodiment of the present invention and polynomial fitted value comparison;

Fig. 5 is the cross-correlation coefficient of the existence time delay of a kind of specific embodiment of the present invention;

Fig. 6 is a cross-correlation interference fringe with residual delay rate in a specific embodiment of the present invention;

Fig. 7 is a cross-correlation coefficient after residual time delay compensation of a specific embodiment of the present invention;

Fig. 8 is the interference fringe after fringe rotation of a specific embodiment of the present invention;

Fig. 9 is a cross-correlation phase spectrum before fractional delay adjustment according to a specific embodiment of the present invention;

Fig. 10 is a cross-correlation phase spectrum after fractional delay adjustment according to a specific embodiment of the present invention.

detailed description

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

In order to meet the requirements of the Mars exploration mission, in the Mars exploration mission, it is planned to adopt the method of receiving data in a remote array to meet the needs of data reception. Compared with the local array, due to the different paths for receiving the downlink signal of the detector, the signals received by the antennas at different locations will have a large delay and delay rate, and the correction of the delay and delay rate will affect the signal synthesis efficiency, The amplitude of the composite signal has a significant effect. In view of this, the present invention provides a fast search method for time delay and time delay rate based on a prediction model in a remote antenna array.

In the embodiment of the present invention, the remote antenna array is composed of N antennas, and finally can be simplified into a plurality of 2-antenna arrays for data processing. Here, a 2-antenna array is used as a schematic example to explain the present invention, but it does not serve as a reference to this invention. Limitations of the invention, one of the antennas is located in Beijing, and the other antenna is located in Kunming, as shown in Figure 1, which is a schematic diagram of signal reception by antenna arrays in different places.

In Figure 1, τ represents the geometric delay (time difference) between the arrival of the spacecraft and the two antennas, Represents the geometric delay rate (time difference change rate) between the spacecraft reaching the two antennas, g represents the geometric path difference between the spacecraft reaching the two antennas, and c represents the speed of light. It can be seen from Figure 1 that in the process of signal propagation in space, due to various factors such as the relative motion of the earth and the spacecraft, atmospheric refraction, ionospheric disturbance, and receiving channels, the delay value and delay rate of the signal reaching the two ground stations is constantly changing. In order to combine the data in the remote antenna array into signals, it is necessary to compensate the delay value and delay rate. The time delay is mainly composed of three parts: geometric time delay, time delay caused by the clock and time delay caused by the atmosphere. Geometric time delay is the largest part of the whole time delay, which is mainly caused by the geometric positions of the observation station and the observation target. In the case of knowing the geometric positions of the observation station and the observation target, the predicted value of the geometric time delay can be obtained through calculation. Due to the limitation of the predicted value, the efficiency of subsequent residual search of time delay and time delay rate will be greatly improved.

Fig. 2 is a flow chart of the fast search method for time delay and time delay rate based on the prediction model in the remote antenna array of the present invention, the method includes the following steps: prediction value calculation, prediction model fitting and residual search, wherein:

Step 1: Predicted value calculation is to preliminarily calculate the predicted value of the geometric delay between the observed target and the antenna based on the position of the antenna and the position of the observed target. This step specifically includes:

Step 101: Read the observation target tracking file list (t, ra, dec) and the geographical coordinates ( _xi , y _i , _zi ) of each antenna in the remote array, where (t, ra, dec) means that at time t, the observation The right ascension and declination values of the target; ( _xi , y _i , _zi ) represent the position of each antenna in the earth space Cartesian coordinate system, the subscript i is used to distinguish different antennas, and there are N antennas in the array, then There are N groups (x _i , y _i , z _i ), and i is 0, 1...N-1.

The observation target tracking file is a set of lists composed of observation target position data. If the observation target is a satellite, then the observation target tracking file is a set of satellite position data calculated from the satellite orbital elements. In order to facilitate the understanding of the data format of the observation target tracking file, the following takes the Chang'e-3 satellite tracking file as an example to give the data format of the observation target tracking file:

year MM UU HH MM SS Distance RA DEC 2015 5 10 0 0 0 374602 20.0849 -15.1805 2015 5 10 0 1 0 374600 20.0856 -15.188 2015 5 10 0 2 0 374597 20.0862 -15.1771 2015 5 10 0 3 0 374595 20.0869 -15.155 2015 5 10 0 4 0 374593 20.0875 -15.1738 2015 5 10 0 5 0 374591 20.0882 -15.1721 2015 5 10 0 6 0 374589 20.0888 -15.1704 2015 5 10 0 7 0 374587 20.0895 -15.1688 2015 5 10 0 8 0 374585 20.0901 -15.1671 2015 5 10 0 9 0 374583 20.0908 -15.1654

The meaning of each parameter in the table is: Year\MM\DD\HH\MM\SS indicates the time stamp, the format is year\month\day\hour\minute\second; Distance indicates the distance from the spacecraft to the center of the earth; RA( Right ascension) indicates the right ascension value of the observation target; DEC (Declination) indicates the declination value of the observation target.

Read the geographical coordinates of each antenna in the remote antenna array. At this time, N=2, there are 2 antennas, located in Kunming and Beijing respectively, as shown in the following table:

name X(m) Y(m) Z(m) Km -1281152.939 5640864.407 2682653.403 bj -2201304.82 4324789.045 4125367.718

In the table, Name is the code of each antenna in the remote antenna array, Km indicates the 40-meter antenna of Kunming Station, Bj indicates the 50-meter antenna of Beijing Miyun Station; X(m)\Y(m)\Z(m) respectively indicate that each antenna is in the center of the earth X-axis, Y-axis, Z-axis values in Cartesian coordinate system, the unit is meter.

Step 102: Calculate the predicted value of the geometric time delay from the observation target to the antenna and the reference point according to the classical celestial body time delay calculation method.

Within the category defined by the International Celestial Reference Frame (ICRF), with the center of the earth as the reference point, calculate the time delay prediction value list τ _i (t _j ) from the observation target to the antenna and the reference point, where t _j represents time, and the subscript i It is used to distinguish different antennas. If there are N antennas in the array, there are N columns of predicted geometric delay values τ _i (t _j ), where i is 0, 1...N-1.

In a celestial frame of reference, the celestial frame of reference is approximated as an inertial frame of reference. In this coordinate system, the default origin is the barycenter of the solar system, and the equatorial plane is defined as the plane determined by the epoch 2000 (J2000) equator and the vernal equinox. In the earth coordinate system, the main axis direction and the corresponding equatorial plane determined by the International Conventional Origin (CIO, Conventional International Origin) are adopted. Using the orbit data of the detector and the baseline vector between the antennas, the calculation process of the geometric time delay is shown in Figure 3. Figure 3 is a flow chart of classic celestial time delay calculation. First, input the coordinates of the station and the position of the detector, and then correct the local coordinate error of the station caused by tides, plate movements, etc., and convert the earth-centered earth coordinate system to the earth-centered celestial coordinate system. (correct the errors caused by precession, nutation, and perturbation), and then transform the geocentric celestial coordinate system to the solar system barycentric celestial coordinate system through the Lorentz transformation, calculate the geometric time delay in the solar system barycentric celestial coordinate system, and correct the solar system gravity The bending error of the signal transmission path caused by the attraction, and finally transform the barycentric celestial coordinate system of the solar system back to the earth-centered celestial coordinate system through the Lorentz transformation to obtain the required geometric time delay.

According to the above calculation, with the center of the earth as the reference point, calculate the geometric delay prediction value list τ _i (t _j ) from the observation target to the antenna and the reference point, where t _j represents time, and the subscript i is used to distinguish different antennas. There are 2 antennas in , then there are 2 columns of geometric delay prediction values τ _i (t _j ), where i is 0, 1...N-1.

In order to facilitate the understanding of the calculation process of the time delay prediction value, a set of observation targets with the Chang'e-3 lander as the observation target (the time period is 00:02:00-59 seconds on May 10, 2015), Kunming Station 40 The meter antenna and the 50-meter antenna at Miyun Station are used as remote array antennas, and the calculation results of the geometric delay prediction value (geometric delay value with the center of the earth as the reference point). The calculation result of the geometric delay prediction value is segmented by 60 seconds, and this is one of the segments.

Geometric delay calculation result table

Step 2: Predictive model fitting is to use the least squares method to perform least squares model fitting on the time delay prediction value obtained from the above calculation to obtain a time-continuous geometric time delay prediction model; this step specifically includes:

Step 201: Read in the geometric time delay prediction value list τ _i (t _j ) of each antenna with respect to the geocentric reference point calculated in the above step 1, and divide the above prediction value list into M segments according to the motion characteristics of the observed target, M is a natural number. Corresponding to the above calculation example, read in the predicted delay values of Miyun station and Kunming station in the table, and segment them in 60 seconds.

Step 202: Using the least squares method, perform polynomial fitting on the above segment of predicted values to obtain a polynomial model y.

Corresponding to the above calculation example, combined with the motion characteristics of the detector and multiple fitting accuracy evaluations, the least squares fitting is performed using a 5th-order polynomial, which can meet the subsequent calculation requirements. The resulting polynomial model is:

Y=6212.21513+0.110956058X+(-1.236351374e- ^05X2 )+(-3.101422866e- ^10X3 )+(3.286324942e- ^12X4 )+(-1.306814583e- ^14X5 ).

Step 203: Evaluate the fitting accuracy of the polynomial model y, the fitting accuracy meets the requirements, and obtain the delay prediction model.

The fitting error RMS needs to be less than f is the observed signal frequency in Hz. If the accuracy meets the requirements, then the group of polynomial coefficients can be used as the prediction model of the antenna. If the fitting accuracy does not meet the requirements, return to step 201, adjust the segmentation or increase the polynomial order, and re-fit until the fitting accuracy meets the requirements. , the M groups of polynomial coefficients can be used as the prediction model of the antenna. Then, the above-mentioned model fitting work is performed on all the antennas in turn to obtain the prediction model of the entire remote antenna array.

Corresponding to the above calculation example, the polynomial model y=6212.21513+0.110956058X+(-1.236351374e-05X ² )+(-3.101422866e-10X ³ )+(3.286324942e-12X ⁴ )+(-1.306814583e-14X ⁵ ) Combined error RMS = 2.862e-07us, root mean square error. The fitting error is given by the formula Obtained, where y _j represents the polynomial model y when X=j time value, τ ₀ (j), τ ₁ (j) correspond to the geometric time delay calculation value of Miyun station and Kunming station j time in the geometric time delay calculation result table .

As shown in Figure 4, it is a comparison between the model calculation value and the polynomial fitting value of a specific embodiment of the present invention, the horizontal axis is time, and the vertical axis is time delay, and the error between the polynomial fitting value and the model calculation value is illustrated in conjunction with the embodiment is very small, the polynomial model obtained by fitting can be used as the delay model.

Step 3: Residual search is to use the above prediction model to perform preliminary delay and delay rate compensation on the observation data of the antenna array, and then use the cross-correlation method to perform cross-correlation processing on the antenna data, and calculate the difference between the real delay value and the predicted value. The residuals between are searched; this step specifically includes:

Step 301: Select the antenna with the largest caliber in the antenna array as the reference antenna, select a delay prediction model value of the antenna array obtained in step 203 minus the prediction model value of the reference antenna, and obtain the delay difference of other antennas relative to the reference antenna Value prediction model y(i, t _j ).

Corresponding to the above calculation example, the 50-meter aperture antenna of Miyun Station is selected as the reference antenna,

Step 302: Read in the reference antenna observation data O _R (t _j ) and another antenna observation data O _i (t _j ) in the array, remove the DC component of the observation data, O ^a _R (t _j ), O ^a _i (t _j ) get.

Corresponding to the above calculation example, the antenna has been taken as an example, read a section of recorded data _OR (t _j ) of the Miyun station antenna, and calculate the average value (DC component) of this section of data Then subtract this record data _OR (t _j ) O ^a _R (t _j ) is obtained, and the processing of other antennas is deduced by analogy.

Step 303: Take 2 ⁿ data from the O ^a _R (t _j ) data, and use the time delay difference prediction model y(i, t _j ) obtained in step 301 to time delay the observed data O ^a _i (t _j ) Compensation and delay rate compensation to obtain 2 ⁿ data [O ^a _i (t _j -y(i, t _j ))] after compensation.

Corresponding to the above calculation example, assume that 256 data are taken from O ^a _R (t _j ) first, and then the prediction model tj=0 is taken to obtain the time delay prediction value τ=6.21221513e-03s at this moment, and the time delay rate prediction value Since the sampling rate of antenna recording data is 1e+07, the sampling time of each data is Ts=1e-07s, the number of delay compensation data bits=τ/1e-7=62122, and the delay compensation is performed on another antenna data , it is necessary to skip the number of delay compensation data bits at the same recording time and read 256 data bits. This data is aligned with the 256 data integer bits of the reference antenna, and the alignment error is less than one sample Ts.

Then, Doppler frequency shift compensation is performed on this piece of data.

Step 304: perform fast Fourier transform on O ^a _R (t _j ), [O ^a _i (t _j -y(i, t _j ))] respectively, and perform cross-correlation to obtain cross-correlation spectrum fft(O _R ^a ( t _j ))×fft ^* ([O ^a _i (t _j -y(i, t _j ))]); find cross-correlation spectrum fft(OR ^a (t _j ))× _fft ^* ([O ^a _i ( t _j -y(i, t _j ))]), then unwrap the phase value, and then use the least squares method to perform a linear function fitting on the phase curve to obtain the residual time of the 2 ⁿ data Extend Δτ ¹ .

Step 305: Take down ²ⁿ data, and perform the calculation in the above step 304 to obtain the residual time delay Δτ ² of the next 2n data, and calculate the time delay rate

Step 306: The residual delay Δτ ¹ and the delay rate obtained from the above search Added to the corresponding geometric time delay prediction y(i, t _j ), the real time delay model that can be used for subsequent antenna array data processing can be obtained.

Corresponding to Δτ ¹ =5e-7s in this embodiment, Then the real delay model

By repeating the operations from step 201 to step 306 for the remaining M-1 sets of data, the real time delay model corresponding to all the observation data of the antenna can be obtained.

By performing all the above operations from step 101 to step 306 on the remaining antennas in the antenna array, the real time delay models corresponding to the observation data of all antennas can be obtained.

Corresponding to the above calculation example, the preliminary delay and delay rate corrections are performed on the data of the two stations. After the correction is completed, the cross-correlation processing is performed on the data of the two stations, and the cross-correlation coefficients and correlation stripes are obtained as shown in Figure 5 and Figure 6.

As shown in Figure 5, it is a cross-correlation coefficient diagram of the existence of a specific embodiment of the present invention, the horizontal axis is time, and the vertical axis is the magnitude of the cross-correlation coefficient, and the embodiment illustrates that due to the existence of residual time delay, the cross-correlation The maximum value of the number does not appear at the zero point of time, and there is a certain offset.

As shown in Figure 6, there is a cross-correlation interference fringe diagram of a specific embodiment of the present invention with a residual delay rate, the horizontal axis is time, and the vertical axis is the phase of the cross-correlation interference fringe. The rate exists, and the phase slope of the cross-correlation interference fringes is not zero.

From the cross-correlation results, it can be seen that due to certain errors in the estimated value of the model, there is a certain deviation between the correlation coefficient and the phase of the correlation spectrum and the ideal autocorrelation spectrum. This deviation is the difference between the model delay value and the real delay value. residual. Through further delay and delay rate estimation, the residual delay is 5e-7s, and the delay rate is 4.11e-11s/s.

As shown in Figure 7, it is a cross-correlation coefficient diagram after residual time delay compensation of a specific embodiment of the present invention, the horizontal axis is time, and the vertical axis is the cross-correlation coefficient magnitude, and it is explained in conjunction with the embodiment that due to the residual time delay For compensation, the maximum value of the cross-correlation coefficient occurs at time zero.

As shown in Figure 8, the cross-correlation interference fringe diagram of residual delay rate compensation, the horizontal axis is the time, and the vertical axis is the phase of the cross-correlation interference fringe. The embodiment illustrates that due to the compensation of the residual delay rate, the cross-correlation The phase slope of the interference fringes is basically zero.

As shown in Figure 9, it is the cross-correlation phase spectrogram before the fractional time delay adjustment, the horizontal axis is the frequency, and the vertical axis is the phase. In combination with the embodiment, it is illustrated that due to the existence of the fractional bit time delay, the phase slope in the cross-correlation phase spectrogram is not zero.

As shown in Figure 10, it is the cross-correlation phase spectrogram after the fractional delay adjustment, the horizontal axis is the frequency, and the vertical axis is the phase. In combination with the embodiment, it is explained that the phase slope in the cross-correlation phase spectrogram is due to the compensation of the fractional time delay. Basically zero.

From the cross-correlation results, it can be seen that the cross-correlation coefficient, fringe and phase spectrum are consistent with the ideal autocorrelation spectrum.

The delay and delay rate fast search method of the present invention introduces the prediction model of delay and delay rate, effectively limits the search range of delay and delay rate, and greatly improves the search efficiency of cross-correlation processing , so that the delay and delay rate values of each antenna in the remote antenna array can be quickly obtained, providing support for subsequent antenna array data processing.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention, and are not intended to limit the present invention. Within the spirit and principles of the present invention, any modifications, equivalent replacements, improvements, etc., shall be included in the protection scope of the present invention.

Claims (8)

1. a kind of time delay and time delay rate method for fast searching based on forecast model, it is characterised in that comprise the steps：

Step 1：According to aerial position and Observed Position, primary Calculation goes out observed object to the latency prediction between antenna Value；

Step 2：Least square models fitting is carried out to calculated latency prediction value, the latency prediction mould of Time Continuous is obtained Type；

Step 3：Observing antenna array data using the latency prediction model for obtaining carries out preliminary time delay and the compensation of time delay rate, so Afterwards cross correlation process is carried out to antenna data using cross-correlation method, the residual error between time delay true value and latency prediction value is carried out Search.

2. the time delay and time delay rate method for fast searching based on forecast model according to claim 1, it is characterised in that institute Stating step 1 includes：

Step 101：Read each antenna geographical coordinate (x in observed object tracking listed files (t, ra, dec) and strange land battle array_i, y_i, z_i), wherein (t, ra, dec) is represented in t, the right ascension and declination value of observed object；(x_i, y_i, z_i) represent terrestrial space Each aerial position in rectangular coordinate system, subscript i is used for distinguishing different antennas, there is N number of antenna in battle array, then have N group (x_i, y_i, z_i), i takes 0,1...N-1；

Step 102：It is pre- to the geometric delays of antenna and reference point according to classical celestial body time-delay calculation method calculating observation target Measured value.

3. the time delay and time delay rate method for fast searching based on forecast model according to claim 2, it is characterised in that step Classical celestial body time-delay calculation method includes described in rapid 102：

Step A1：Input survey station coordinate and detector position；

Step A2：The survey station local coordinate system error that amendment is caused by tide, plate motion etc.；

Step A3：The earth's core terrestrial coordinate system is transformed into into the earth's core celestial coordinate system；

Step A4：The earth's core celestial coordinate system is transformed to by solar system barycenter celestial coordinate system by Lorentz transformation, in the solar system Geometric delays are calculated in barycenter celestial coordinate system, and corrects the signal transmission path bending error that solar system gravitating is caused；

Step A5：Solar system barycenter celestial coordinate system is switched back to by the earth's core celestial coordinate system by Lorentz transformation, obtains required Geometric delays predictive value.

4. the time delay and time delay rate method for fast searching based on forecast model according to claim 1, it is characterised in that institute Stating step 2 includes：

Step 201：By each antenna for obtaining with regard to the earth's core reference point geometric delays predictive value τ_i(t_j) read in, according to observation The kinetic characteristic of target, M segmentations are divided into above-mentioned prediction value list, and M is natural number；

Step 202：Using method of least square, fitting of a polynomial is carried out to above-mentioned one section of predictive value, obtain multinomial model y；

Step 203：Assessment multinomial model y fitting precision, fitting precision meets requirement, obtains latency prediction model.

5. the time delay and time delay rate method for fast searching based on forecast model according to claim 4, it is characterised in that step Method of least square carries out least square fitting using 5 rank multinomials described in rapid 202.

6. the time delay and time delay rate method for fast searching based on forecast model according to claim 4, it is characterised in that institute Stating step 203 includes：

According to formulaAn algnment accuracy is calculated, f is observation signal frequency, and unit is Hz；Error of fitting is logical Cross formulaTry to achieve, y in formula_jMultinomial model y is represented as X=j moment values, τ₀(j), τ₁J () correspondence is several When the geometric delays value of calculation in result of calculation table Miyun station and Station in Kunming j moment is prolonged, and n represents the individual of each piecewise prediction value Number；Error of fitting RMS is needed less than algnment accuracy；If precision meets required, this group of multinomial coefficient can be used as the day The latency prediction model of line, if fitting precision is unsatisfactory for requiring, return to step 201, adjustment is segmented or increases order of a polynomial Number, re-starts fitting, requires until fitting precision meets.

7. the time delay and time delay rate method for fast searching based on forecast model according to claim 1, it is characterised in that step Residual error described in rapid 3 between time delay true value and latency prediction value includes remaining time delay and remaining time delay rate.

8. the time delay and time delay rate method for fast searching based on forecast model according to claim 1, it is characterised in that institute Stating step 3 includes：

Step 301：The maximum antenna of antenna array medium caliber is chosen as reference antenna, the 1 of antenna array is obtained in selecting step 203 Individual latency prediction model value deducts the forecast model value of reference antenna, obtains time delay difference of other antennas relative to reference antenna Forecast model y (i, t_j)；

Step 302：Read in reference antenna observation data O_R(t_j) and battle array in other antenna observe data O_i(t_j), remove observation The DC component of data, obtains O^a _R(t_j), O^a _i(t_j)；

Step 303：From O^a _R(t_j) 2 are taken in dataⁿIndividual data, delay inequality value prediction model y (i, the t obtained using step 301_j) To observing data O^a _i(t_j) carry out delay compensation and the compensation of time delay rate, 2 after being compensatedⁿIndividual data [O^a _i(t_j- y (i, t_j))]；

Step 304：Respectively to O^a _R(t_j), [O^a _i(t_j- y (i, t_j))] ' fast fourier transform is carried out, and cross-correlation is carried out, obtain Cross-correlation frequency spectrum fft (O_R ^a(t_j))×fft^*([O^a _i(t_j- y (i, t_j))]′)；Seek cross-correlation frequency spectrum fft (O_R ^a(t_j))×fft^* ([O^a _i(t_j- y (i, t_j))] ') phase value, then to phase value solution wind, then using least square method to phase curve Carry out once linear Function Fitting, obtain this 2ⁿThe remaining time delay △ τ of individual data¹；

Step 305：Remove 2ⁿIndividual data, equally carry out above-mentioned 304 step and calculate, and obtain down 2ⁿThe remaining time delay △ τ of individual data², It is calculated time delay rate

Step 306：The remaining time delay △ τ that above-mentioned search is obtained¹With time delay rateIt is added to corresponding delay inequality value prediction model y (i, t_j) in, you can obtain the true Time Delay Model for being available for subsequent antenna group battle array data processing to use.