Radar and method for target detection using the radar

Abstract

The present invention relates to a radar and a method for using the radar for target detection, including an antenna, a receiving system, a transmitting system and a short track; the short track is provided with a receiving system capable of reciprocating thereon; or a transmitting system; or a receiving system and a transmitting system. The present invention has the following advantages: 1. The scanning speed is greatly improved, the track length is shortened, the mechanical complexity is reduced, the power consumption is reduced, and the scanning speed is improved; compared with array-type synthetic aperture radar, the detection distance is improved and the imaging resolution is improved. 2. The imaging speed of synthetic aperture radar is greatly improved, and the requirements of imaging algorithms for resources are reduced. The present invention adopts a method of quickly finding the angle of the antenna simulation array for imaging, which not only greatly speeds up the imaging speed, but also saves memory resources, and is convenient for real-time monitoring.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing information, in particular to a radar and a method for detecting a target using the radar. Background Art

[0002] Radar interferometry is an important precision measurement method. Microwave radar has the advantages of good penetration, long distance and high accuracy. It is widely used in disaster prevention and mitigation and deformation monitoring of structures. Since a single fixed real aperture radar does not have azimuth resolution, synthetic aperture radar is often used in practice to achieve azimuth resolution to obtain the position of the target. The amount of data of synthetic aperture radar is larger than that of a single fixed real aperture radar, and the processing method is more complicated. The current synthetic aperture methods are generally orbital, rotating, and MIMO. The most classic mode is orbital, which has extremely mature technology, stable operation and high accuracy. However, since the track length is directly related to the azimuth resolution, in order to achieve a certain azimuth resolution, the track length must be long enough. Due to mechanical reasons, the speed of the radar along the track is limited. Therefore, the longer the track, the longer the scanning time, that is, the longer the scanning cycle, the lower the monitoring frequency, so that the instantaneous displacement cannot be captured, and the monitoring effect of targets such as bridges and buildings will be reduced. The volume and weight of the long track are also large. The same problem also exists for rotating synthetic aperture radar. Although MIMO radar can achieve a very high monitoring frequency, there is a contradiction between antenna spacing and antenna size. Antenna size directly affects antenna gain, which in turn affects the observation distance, while antenna spacing affects the imaging quality. In addition, the currently commonly used synthetic aperture radar imaging algorithm is inefficient and not conducive to real-time observation. Summary of the invention

[0003] The above technical problems of the present invention are mainly solved by the following technical solutions:

[0004] A radar, comprising an antenna, a receiving system, and a transmitting system, characterized in that it also comprises a short track;

[0005] The short track is provided with a receiving system capable of reciprocating thereon;

[0006] Or a short track is provided with a launching system capable of reciprocating thereon;

[0007] In the above-mentioned radar, a receiving system capable of reciprocating thereon is arranged on the short track, and the length of the short track is greater than or equal to the spacing between antennas.

[0008] In the above-mentioned radar, a transmitting system capable of reciprocating thereon is arranged on the short track, and the length of the short track is greater than or equal to the spacing between antennas.

[0009] In the above-mentioned radar, a receiving system and a transmitting system capable of reciprocating thereon are arranged on the short track at the same time, and the length of the short track is greater than or equal to (the distance between the antennas) / 2.

[0010] A method for target detection using radar, characterized in that it includes

[0011] Perform data acquisition to obtain original radar echo data arranged in the acquisition space order;

[0012] The original radar echo data is subjected to focused imaging by antenna array phase shaping, that is, the original radar echo data is subjected to Fourier transform to obtain radar echo data consisting of an array converted into the frequency domain;

[0013] Use the set antenna phase delay parameters to perform delay correction on the array to obtain the corrected radar echo data;

[0014] Perform spatial Fourier transform on the corrected data to obtain converted radar echo data;

[0015] The actual angle of the azimuth represented by the converted radar echo data is determined to obtain an image.

[0016] In the above target detection method, during data collection, the moving interval of the track is defined as the antenna spacing a, the moving step number is k, and the track moving interval d is smaller than the wavelength λ, and d is set to be smaller than λ / 10;

[0017] When data collection starts, the track drives the transmitting antenna or the receiving antenna from the starting point, and moves d each time. The radar scans once and collects data once until the end point. At this time, the track moves a distance; the collected data is stored as:

[0018] Starting position 0*d: A 1,1 , A 2,1 , A 3,1 ……A n,1 ,

[0019] Orbital position 1*d: A 1,2 , A 2,2 , A 3,2 ……A n,2 ,

[0020] Orbital position 2*d:A 1,3 , A 2,3 , A 3,3 ……A n,3 ,

[0021] …

[0022] End position (k-1)*d=a:A 1,k , A 2,k , A3,k ……A n,k

[0023] Each unit A is an array, and the length of the array is related to the scanning time. If the scanning time is t and the sampling rate is sp, the array length is the product of t and sp. Each row is a radar scan when the track stays at a position, and each column corresponds to a receiving antenna. n is the number of receiving antennas, and k is the number of moving steps. That is, we get an array of k rows and n columns, and the length of each array is t*sp.

[0024] Starting from the upper left corner, from top to bottom, and then from left to right, take out the A elements in the array matrix one by one, reorganize the original data, and expand it into one dimension:

[0025] A 1,1 , A 1,2 , A 1,3 , ... A 1,k ,

[0026] A 2,1 , A 2,2 , A 2,3 , ... A 2,k ,

[0027] …………

[0028] A n,1 ,A n,2 ,A n,3 ……A n,k

[0029] So far, we have obtained n*k original radar echo data arranged in the order of acquisition space, and the length of each echo data is t*sp; because each A is actually an array of length t*sp:

[0030] A=B(1,2,3,4,5...t*sp)

[0031] Therefore, if each element of A is expanded into B, the above one-dimensional array is actually an array B with n*k in the first dimension and t*sp in the second dimension:

[0032] B 1,1 , B 1,2 , B 1,3 , B 1,4 , ... B 1，t*sp ,

[0033] B 2,1 , B 2,2 , B 2,3 , B 2,4 , ... B 2，t*sp ,

[0034] B3,1 , B 3,2 , B 3,3 , B 3,4 , ... B 3，t*sp ,

[0035] …

[0036] B n*k,1 , B n*k,2 , B n*k,3 , B n*k,4 , ... B n*k,t*sp

[0037] This data is equivalent to the data obtained by a single receiving antenna orbital radar with a track length of a*n, and the observation period is shortened to 1 / n of that of the traditional orbital radar.

[0038] In the above target detection method, the specific method of performing antenna array phase shaping focusing imaging on the original observation data is:

[0039] First, perform Fourier transform on each row of array B to convert each row from the time domain of the original information to the frequency domain:

[0040] F=fft(B)

[0041] Get the new array:

[0042] F 1,1 , F 1,2 , F 1,3 , F 1,4 , ... F 1，t*sp ,

[0043] F 2,1 , F 2,2 , F 2,3 , F 2,4 , ... F 2，t*sp ,

[0044] F 3,1 , F 3,2 , F 3,3 , F 3,4 , ... F 3，t*sp ,

[0045] …

[0046] F n*k,1 , F n*k,2 , F n*k,3 , F n*k,4 , ... F n*k,t*sp

[0047] Different elements in each row of the F array represent corresponding FFT bins, whose modulus means the reflection intensity of the target at the distance, and whose phase angle represents the distance phase of the target.

[0048] In the above target detection method, the specific method of using the stored antenna phase delay parameters to perform delay correction on the F array is:

[0049] The F array has n*k rows, where the data from row number 1 to k is obtained by antenna No. 1, the data from row k+1 to 2*k is obtained by antenna No. 2, and the data from row number (n-1)*k to n*k is obtained by antenna No. n;

[0050] The correction method is to divide the data of each row by the correction value of the corresponding antenna in the correction parameter array C, that is, the F data with row numbers (n-1)*k to n*k is divided by the nth element in the C array. Complex number division is geometrically equivalent to correcting the phase angle; each element of the F data with row numbers 1 to k is divided by the first element of the correction parameter array C;

[0051] The corrected data FC is obtained:

[0052] FC 1,1 , FC 1,2 , FC 1,3 , FC 1,4 , ……FC 1，t*sp ,

[0053] FC 2,1 , FC 2,2 , FC 2,3 , FC 2,4 , ……FC 2，t*sp ,

[0054] FC 3,1 , FC 3,2 , FC 3,3 , FC 3,4 , ……FC 3，t*sp ,

[0055] …

[0056] FC n*k,1 , FC n*k,2 , FC n*k,3 , FC n*k,4 , ……FC n*k,t*sp .

[0057] In the above target detection method, the same target is in the same position in each row of data, that is, the same FFT well, so the phase difference between two adjacent rows of sampled data is d*sinα / λ

[0058] Perform spatial Fourier transform on the array FC column by column, that is, 1,1 , FC 2,1 , FC 3,1 , ……FC n*k,1 ), (FC 1,2 , FC 2,2 , FC 3,2 , ……FC n*k,2 )... to (FC 1,t*sp , FC 2，t*sp , FC 3，t*sp , ……FC n*k,t*sp ) to perform Fourier transform, and then rotate it back to the original direction to get the FCF array:

[0059] FCF 1,1 , FCF 1,2 , FCF 1,3 , FCF 1,4 , ...FCF 1，t*sp ,

[0060] FCF 2,1 , FCF 2,2 , FCF 2,3 , FCF 2,4 , ...FCF 2，t*sp ,

[0061] FCF 3,1 , FCF 3,2 , FCF 3,3 , FCF 3,4 , ...FCF 3，t*sp ,

[0062] …

[0063] FCF n*k,1 , FCF n*k,2 , FCF n*k,3 , FCF n*k,4 , ……FC n*k,t*sp

[0064] The row where the target's azimuth is located is the peak of the FFT result; in the FCF array, the horizontal direction represents the upward target signal, that is, the distance information, while the vertical direction represents the azimuth information of the target signal, that is, the angle, that is, the azimuth information.

[0065] In the above target detection method, the actual angle of the azimuth represented by the longitudinal direction of the FCF array is determined to obtain an image. The specific method is:

[0066] Create an array ANG with a size of n*k and the contents are:

[0067] arcsin(x*λ / (2*d)), where x decreases from 1 to -1 with a step size of 2 / (n*k)

[0068] The ANG array is a series of angles, which correspond one-to-one to the elements of each column of the FCF array. The ANG array is used as the vertical axis scale, ranging from -90 degrees to 90 degrees, and the distance is used as the horizontal axis scale, ranging from 0 to the maximum distance of the radar. The FCF array values ​​are plotted to obtain the focused image.

[0069] The value range of the sin function is in [-1,1], and there will be outliers in the ANG array. If d<λ / 2 is satisfied, there will be some outliers at the beginning and end of the array. Delete the outliers and the corresponding rows in the FCF array at their corresponding positions.

[0070] At this time, the image is in polar coordinates; since the horizontal coordinate of the array corresponds to the distance, the amplitude of each element represents the reflection intensity of the target, the phase angle represents the distance phase of the target, and the value of the ANG array element corresponding to the row number represents the azimuth angle α of the target, if the data result is converted into Cartesian coordinates, the following method is used:

[0071] X=D*cosα

[0072] Y = -D*sinα.

[0073] Therefore, the present invention has the following advantages: 1. It combines the advantages of mechanical synthetic aperture radar and array synthetic aperture radar (MIMO, SIMO, MISO). Compared with the original mechanical synthetic aperture radar, it greatly improves the scanning speed, shortens the track length, reduces the mechanical complexity, reduces the power consumption, and improves the scanning speed; compared with the array synthetic aperture radar, it improves the detection distance and the imaging resolution. 2. It greatly improves the imaging speed of synthetic aperture radar and reduces the requirements of imaging algorithms on resources. The traditional Omega-K or Back Project mode for imaging occupies a lot of resources and has low efficiency. The present invention adopts the method of quickly finding the antenna simulation array angle for imaging, which not only greatly speeds up the imaging speed, but also saves memory resources, which is convenient for real-time monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Attached Figure 1 It is a principle diagram of the present invention, which is an embodiment in which the transmitting antenna is installed on a slide rail.

[0075] Attached Figure 2 This is an embodiment in which both the transmitting and receiving antennas are placed on the slide rails.

[0076] Attached Figure 3 The invention is an embodiment in which the receiving antenna is installed on a slide rail.

[0077] Among them, 01 is the receiving antenna array, 02 is the transmitting antenna, and 03 is the spacing d between the track antennas. DETAILED DESCRIPTION

[0078] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0079] Example:

[0080] 1. First, the structure of the present invention is introduced. The present invention includes an antenna, a receiving system, a transmitting system, and a short track, on which a receiving system or a transmitting system, or a receiving system and a transmitting system, which can reciprocate thereon are arranged (it should be noted that in this embodiment, the receiving system and the transmitting system are also arranged on the track at the same time, but since this is not the most economical mode, it is not recommended. In order to reduce the power consumption of the track motor, generally only a small number of antennas are placed on the track). The present invention combines the characteristics of MIMO with the advantages of track-type synthetic aperture radar, adopts a single-transmit multiple-receive or multiple-transmit multiple-receive method in the antenna transceiver system, reasonably designs the size of the antenna, and ensures the gain. Since the size of the antenna is relatively large, the antenna spacing also increases accordingly. As mentioned above, the track length of the short track is slightly larger than the antenna spacing. The transceiver part of the antenna is placed on this short track for movement, or only the transmitting antenna is placed on the short track for movement, and the receiving antenna is fixed, or only the receiving antenna is placed on the short track for movement, and the transmitting antenna is fixed. No matter which of the three methods is used, the moving range of the track only needs to cover the spacing between the two antennas to fill the antenna gap, and then the collected data is reorganized into a sequential scanning sequence, thereby forming rich data for imaging. In this way, the imaging quality and observation distance can be guaranteed, and the observation speed can be improved; the receiving system and the transmitting system adopt a MIMO multi-antenna transceiver system or a SIMO multi-receive single-transmit, a MISO multi-transmit single-receive, and the like.

[0081] Second, the structural design principle of the present invention is introduced.

[0082] Assume that the central wavelength of the radar is λ, which is usually a few millimeters to a few centimeters. The sampling frequency is p, which is usually tens of hertz, and the antenna width is a, which is usually a few centimeters to more than ten centimeters. Generally speaking, in order to ensure the measurement distance, a high-gain antenna is used. The aperture of a high-gain antenna is also large, and it is usually larger than a wavelength, that is, a>λ. If a traditional orbital radar is used, assuming that the track length is l, which is generally 1-3 meters, and the track speed is v, which is generally a few centimeters per second, then the observation period is t rail=l / v, that is, the time for a one-way orbit, which is generally tens of seconds, and the observation spacing is v / p, about a few millimeters. If only single-column MIMO is used, n antennas are configured, and the antennas are closely arranged, so the antenna spacing is equal to a, n is usually 4, 8, 16, etc., and the n value must be close to l / a to obtain the effect of an orbital radar with a length of l. Then the antenna spacing will be greater than one wavelength. Although the observation period is equal to 1 / p, which is shortened to less than 1 second, the collected data is too sparse in space, which is not conducive to imaging.

[0083] The present invention first adopts a MIMO multi-antenna transceiver system, and then places it on a track with a travel distance a. When using the same movement and sampling rate as the above-mentioned track radar, the observation period is: a / v. Since a is close to l / n, the observation period t is close to t rail / n, so the present invention can effectively improve the observation efficiency. The observation spacing of the radar of the present invention is the same as that of the traditional orbital radar, so the imaging effect and the observation distance can also be guaranteed.

[0084] Since there are cables, low noise amplifiers and other components between the radar's receiving circuit and the antenna, the phase delay of each receiving antenna must be different. Before use, the phase delay needs to be measured and corrected to reduce the impact on imaging. The general measurement and correction method is to use a vector network analyzer or directly use the radar in a microwave darkroom to align the antenna array with a corner reflector. The corner reflector is located on the perpendicular bisector of the receiving antenna array. The distance Dcr from the corner reflector to the antenna array should be greater than:

[0085] (L*L / 4-λ*λ) / (2*λ), L is the length of the antenna array, and λ is the wavelength. This is to avoid phase ambiguity.

[0086] The phase delay measurement records of each antenna are stored in the correction constant array C in the form of complex numbers on the unit circle for later use. The length of C is n, where n is the number of antennas.

[0087] If a traditional orbital radar is used, the orbital travel (or orbital movement interval) l = 1 meter, and the orbital speed v = 0.05 meters per second, data is collected 20 times per second, and the antenna width is 0.0625 meters, then the observation period is 20 seconds, and the observation spacing is 0.0025 meters. If the existing MIMO radar technology is used, the antenna width is also 0.0625 meters. If you want to achieve the effect of a 1-meter track length, you need 16 receiving antennas with an antenna spacing of 0.0625 meters. If the observation frequency is 20 Hz, then the observation period is shortened to 0.05 seconds, but because the current monitoring radars mostly use frequencies above the C band, their wavelengths are generally less than 0.06 meters. Therefore, the antenna spacing of 0.0625 meters is too sparse and almost difficult to image. If the L band is used, the antenna gain of a width of 0.0625 meters is too small, the effective distance is too short, and it is also difficult to image. If the method of the present invention is used, the antenna width is 0.0625 meters, and the mode of 1 transmission and 16 reception is adopted. The receiving antenna array is fixed, and the transmitting antenna moves with the track. The track travel is only 0.0625 meters. Assuming that the track speed v is also 0.05 meters per second, the observation period only needs 1.25 seconds. If data is collected 20 times per second, the observation spacing is also 0.0025 meters. This is equivalent to forming a MIMO array of 16*0.0625 / 0.0025=16*25=400 antennas, but the complexity is much simpler than the MIMO array of 400 antennas, and only 17 antennas are needed. In effect, it can achieve the same imaging quality and observation distance as the traditional track radar, and has a much faster observation speed than the traditional track radar. In addition, the track length is shortened, the track load is greatly reduced, and the volume, weight, and power consumption are reduced accordingly, which combines the advantages of the traditional track radar and the traditional MIMO radar.

[0088] Traditional Omega-K or Back Project modes for imaging occupy large resources and have low efficiency. The present invention adopts a method of quickly finding the angle of the antenna simulation array for imaging, which not only greatly speeds up the imaging speed, but also saves memory resources and facilitates real-time monitoring.

[0089] 3. The target detection principle of the present invention is introduced below.

[0090] Step A: First, collect data as follows:

[0091] The moving interval of the track is the antenna spacing a, and the number of moving steps is k. To ensure data quality, the track moving interval d needs to be much smaller than the wavelength λ. Generally, d can be set less than λ / 10.

[0092] If the track drives both the transmitting and receiving antennas, the track movement interval d is less than λ / 2, and d is set to be less than λ / 20.

[0093] When data collection starts, the track drives the transmitting antenna or the receiving antenna from the starting point, and moves d each time. The radar scans once and collects data once until the end point. At this time, the track moves a distance. The collected data is stored as:

[0094] Starting position 0*d: A 1,1 , A 2,1 , A 3,1 ……A n,1 ,

[0095] Orbital position 1*d: A 1,2 , A 2,2 , A 3,2 ……A n,2 ,

[0096] Orbital position 2*d:A 1,3 , A 2,3 , A 3,3 ……A n,3 ,

[0097] …

[0098] End position (k-1)*d=a:A 1,k , A 2,k , A 3,k ……A n,k

[0099] Each unit A is an array, and the length of the array is related to the scanning time. If the scanning time is t and the sampling rate is sp, the array length is the product of t and sp. Each row is a radar scan when the track stays at a position, and each column corresponds to a receiving antenna. n is the number of receiving antennas, and k is the number of steps. That is, we get an array of k rows and n columns, and the length of each array is t*sp.

[0100] Starting from the upper left corner, from top to bottom, and then from left to right, take out the A elements in the array matrix one by one, reorganize the original data, and expand it into one dimension:

[0101] A 1,1 , A 1,2 , A 1,3 , ... A 1,k ,

[0102] A 2,1 , A 2,2 , A 2,3 , ... A 2,k ,

[0103] …………

[0104] A n,1 ,A n,2 ,An,3 ...A n,k

[0105] So far, we have obtained n*k original radar echo data arranged in the order of acquisition space, and the length of each echo data is t*sp. Because each A is actually an array of length t*sp:

[0106] A=B(1,2,3,4,5...t*sp)

[0107] Therefore, if each element of A is expanded into B, the above one-dimensional array is actually an array B with n*k in the first dimension and t*sp in the second dimension:

[0108] B 1,1 , B 1,2 , B 1,3 , B 1,4 , ... B 1，t*sp ,

[0109] B 2,1 , B 2,2 , B 2,3 , B 2,4 , ... B 2，t*sp ,

[0110] B 3,1 , B 3,2 , B 3,3 , B 3,4 , ... B 3，t*sp ,

[0111] …

[0112] B n*k,1 , B n*k,2 , B n*k,3 , B n*k,4 , ... B n*k,t*sp

[0113] This data is equivalent to the data obtained by a single receiving antenna orbital radar with a track length of a*n, but due to the use of this method, the observation period is shortened to 1 / n of the traditional orbital radar.

[0114] Step B: Next, the original observation data is subjected to focused imaging by antenna array phase shaping:

[0115] First, perform Fourier transform on each row of array B to convert each row from the time domain of the original information to the frequency domain:

[0116] F=fft(B)

[0117] Get the new array:

[0118] F 1,1 , F1,2 , F 1,3 , F 1,4 , ... F 1，t*sp ,

[0119] F 2,1 , F 2,2 , F 2,3 , F 2,4 , ... F 2，t*sp ,

[0120] F 3,1 , F 3,2 , F 3,3 , F 3,4 , ... F 3，t*sp ,

[0121] …

[0122] F n*k,1 , F n*k,2 , F n*k,3 , F n*k,4 , ... F n*k,t*sp

[0123] Different elements in each row of the F array represent corresponding FFT bins, whose modulus means the reflection intensity of the target at the distance, and whose phase angle represents the distance phase of the target.

[0124] Step C: Use the stored antenna phase delay parameters to perform delay correction on the F array:

[0125] The F array has n*k rows, where the data from row numbers 1 to k are obtained by antenna No. 1, the data from row k+1 to 2*k are obtained by antenna No. 2, and so on. It can be concluded that the data from row numbers (n-1)*k to n*k are obtained by antenna No. n.

[0126] The correction method is to divide the data of each row by the correction value of the corresponding antenna in the correction parameter array C, that is, the F data with row numbers (n-1)*k to n*k is divided by the nth element in the C array (complex number division is geometrically equivalent to correcting the phase angle). For example, each element of the F data with row numbers 1 to k is divided by the first element of the correction parameter array C.

[0127] The corrected data FC is obtained:

[0128] FC 1,1 , FC 1,2 , FC 1,3 , FC 1,4 , ……FC 1，t*sp ,

[0129] FC 2,1 , FC 2,2 , FC2,3 , FC 2,4 , ……FC 2，t*sp ,

[0130] FC 3,1 , FC 3,2 , FC 3,3 , FC 3,4 , ……FC 3，t*sp ,

[0131] …

[0132] FC n*k,1 , FC n*k,2 , FC n*k,3 , FC n*k,4 , ……FC n*k,t*sp

[0133] Step D: Since there is a distance d between the antennas in each row of data, there is a slight difference in the distance phase of the target for the same target. This slight difference is related to the distance D of the target and the azimuth α of the target. The distance D of the target is much larger than d, and the same target is generally in the same position (i.e., the same FFT well) in each row of data, so the phase difference between two adjacent samples (i.e., two adjacent rows) is d*sinα / λ; if the track drives the transmitting and receiving antennas at the same time, the phase difference is 2*d*sinα / λ.

[0134] Therefore, the spatial Fourier transform of the array FC is performed column by column in the column direction, that is, (FC 1,1 , FC 2,1 , FC 3,1 , ……FC n*k,1 ), (FC 1,2 , FC 2,2 , FC 3,2 , ……FC n*k,2 )... to (FC 1,t*sp , FC 2，t*sp , FC 3，t*sp , ……FC n*k,t*sp ) to perform Fourier transform, and then rotate it back to the original direction to get the FCF array:

[0135] FCF 1,1 , FCF 1,2 , FCF 1,3 , FCF 1,4 , ...FCF 1，t*sp ,

[0136] FCF 2,1 , FCF 2,2 , FCF 2,3 , FCF 2,4, ...FCF 2，t*sp ,

[0137] FCF 3,1 , FCF 3,2 , FCF 3,3 , FCF 3,4 , ...FCF 3，t*sp ,

[0138] …

[0139] FCF n*k,1 , FCF n*k,2 , FCF n*k,3 , FCF n*k,4 , ……FC n*k,t*sp

[0140] This operation makes the row where the target's azimuth is located the peak of the FFT result. Therefore, in the FCF array, the horizontal direction represents the target signal in the upward direction (i.e., distance information), while the vertical direction represents the azimuth information of the target signal (i.e., angle, i.e., azimuth information).

[0141] Step E: Next, determine the actual angle of the azimuth represented by the longitudinal direction of the FCF array and obtain the image:

[0142] Create an array ANG with a size of n*k and the contents are:

[0143] arcsin(x*λ / (2*d)), where x decreases from 1 to -1 with a step size of 2 / (n*k)

[0144] If the track drives both the transmitting and receiving antennas, the formula arcsin(x*λ / (4*d)) should be used. It can be seen that the ANG array is a series of angles, which correspond one-to-one to the elements of each column of the FCF array. Just use the ANG array as the vertical axis scale (range -90 degrees to 90 degrees) and the distance as the horizontal axis scale (range 0 to the maximum distance of the radar), and plot the FCF array values ​​to get the focused image.

[0145] Since the value range of the sin function is in [-1,1], there may be outliers in the ANG array (as long as d<λ / 2, there will be some outliers at the beginning and end of the array). These outliers should be deleted, and the corresponding rows in the FCF array at their corresponding positions should be deleted at the same time.

[0146] At this time, the image is in polar coordinates. Since the horizontal coordinate of the array corresponds to the distance, the amplitude of each element represents the reflection intensity of the target, the phase angle represents the distance phase of the target, and the value of the ANG array element corresponding to the row number represents the azimuth angle α of the target, if you want to convert the data result into Cartesian coordinates, you can use the following method:

[0147] X=D*cosα

[0148] Y=-D*sinα

[0149] In the whole process of focusing and imaging, the Fourier transform occupies the most resources, and the fast Fourier transform is very mature. It is much simpler and more effective than other algorithms (other focusing algorithms require interpolation and other processing in addition to Fourier transform, which is more resource-intensive), and it is easy to implement GPU acceleration and multi-threaded computing. Actual tests show that with the same CPU and memory resources, this method takes only about 1% of the time of the Back Project algorithm, and occupies only about 1 / 10 of the memory.

[0150] 4. The following is a specific case application of the structural design principle adopted in the second point.

[0151] Take C-band radar (assuming the center frequency is 5450MHz) as an example. At this time, the wavelength is 0.055cm, the design sampling interval is 0.0125m, the track moving speed is 0.05m / s, the radar sampling rate is 1,000,000 times / s, the scanning time is 0.001s, each scan has 1,000 data, and the effective range is 500m. The mode adopted is that the receiving antennas form an array and only one transmitting antenna is used, and the transmitting antenna moves along a short track.

[0152] 1. Design and manufacture antennas and antenna arrays.

[0153] First, a single receiving antenna is designed. A commonly used horn antenna is used. With the antenna width of 0.125 meters as a constraint, a horn antenna with a gain of about 15dBi can be obtained. The design uses 8 receiving antennas, so the total length of the receiving antenna array is 1.0 meters, and the antenna spacing is 0.125 meters.

[0154] 2. Calibrate the antenna array.

[0155] In a microwave darkroom, connect the eight receiving antennas of the antenna array to the eight receiving channels of the radar, and place a corner reflector at a certain distance Dcr from the antenna (when the corner reflector is located on the perpendicular bisector of the antenna array, the formula Dcr>(L*L / 4-λ*λ) / (2*λ) can be used to determine the minimum Dcr, L is the length of the antenna array, and λ is the wavelength). Here, Dcr should be greater than 2.25 meters. Turn on the radar, transmit the signal, record the phase value of the corner reflector received by each receiving antenna, and record it as the C array for backup. Due to the large span and high frequency of the receiving antenna array, in addition to the cable, there are also connectors, RF switches and other devices on the link, and the loss can often reach 3dB. In order to compensate for these losses, the design size of the transmitting antenna can be appropriately larger to obtain greater gain. In this example, the transmitting antenna is designed as a horn antenna with a width of 0.18 meters and a gain of about 18dBi. It is just used to compensate for the loss of the receiving channel.

[0156] 3. Design and manufacture antenna assembly.

[0157] This embodiment adopts a mode in which the receiving antenna is stationary and the transmitting antenna moves along the track. The transmitting antenna moving platform (track) uses a stepper motor to drive the lead screw. The stepper motor is synchronized with the radar transmitting and receiving part and is controlled by the main controller. The radar transmits and receives data once for each step. The track travel is greater than 0.125 meters, and a safe distance needs to be left on both sides, so the total travel is 0.2 meters. The transmitting antenna is fixed on the slider of the track so that it can move along the track.

[0158] The antenna direction of the receiving antenna array is the same as that of the transmitting antenna, and there is a certain distance between them to ensure the isolation between the receiving and transmitting. The specific spacing is related to the shape of the antenna's directional pattern lobe. The receiving and transmitting antennas should not fall into each other's strong side lobes.

[0159] 5. The following introduces a specific case application of the target detection principle adopted in the third point.

[0160] 1. Collect data for preprocessing.

[0161] Control the orbital movement, control the radar to transmit signals when the orbital position is i*0.0125 meters (i=0,1,2,3,4...9), and receive data through 8 antennas in 8 channels:

[0162] Starting position 0*0.0125m: A 1,1 , A 2,1 , A 3,1 ...A 8,1 ,

[0163] Track position 1*0.0125m: A 1,2 , A 2,2 , A 3,2 ...A8,2 ,

[0164] Track position 2*0.0125m:A 1,3 , A 2,3 , A 3,3 ……A 8,3 ,

[0165] …

[0166] End position 9*0.0125m: A 1,10 , A 2,10 , A 3,10 ……A 8,10

[0167] Reorganize the above data from top to bottom and from left to right and stretch it into a one-dimensional array A:

[0168] A 1,1 , A 1,2 , A 1,3 , ... A 1,10 ,

[0169] A 2,1 , A 2,2 , ... A 2,10 ,

[0170] A 3,1 , A 3,2 , ... A 3,10 ,

[0171] …

[0172] A 8,1 ,A 8,2 ,……A 8,10

[0173] There are 80 lines in total.

[0174] Each cell A is an array of length 1000. Expand array A into array B:

[0175] B 1,1 , B 1,2 , B 1,3 , B 1,4 , ... B 1，1000 ,

[0176] B 2,1 , B 2,2 , B 2,3 , B 2,4 , ... B 2，1000 ,

[0177] B 3,1 , B 3,2 , B 3,3, B 3,4 , ... B 3，1000 ,

[0178] …

[0179] B 80,1 , B 80,2 , B 80,3 , B 80,4 , ... B 80，1000

[0180] Perform fast Fourier transform F=fft(B) on the B array row by row

[0181] Get the new array F:

[0182] F 1,1 , F 1,2 , F 1,3 , F 1,4 , ... F 1，1000 ,

[0183] F 2,1 , F 2,2 , F 2,3 , F 2,4 , ... F 2，1000 ,

[0184] F 3,1 , F 3,2 , F 3,3 , F 3,4 , ... F 3，1000 ,

[0185] …

[0186] F 80,1 , F 80,2 , F 80,3 , F 80,4 , ... F 80，1000

[0187] Divide the data of each row by the correction value of the corresponding antenna in the correction parameter array C, that is, divide the F data with row numbers (n-1)*k to n*k by the nth element in the C array (complex number division is geometrically equivalent to correcting the phase angle). For example, each element of the F data with row numbers 1 to k is divided by the first element of the correction parameter array C.

[0188] The corrected data FC is obtained:

[0189] FC 1,1 , FC 1,2 , FC 1,3 , FC 1,4 , ……FC 1，1000 ,

[0190] FC2,1 , FC 2,2 , FC 2,3 , FC 2,4 , ……FC 2，1000 ,

[0191] FC 3,1 , FC 3,2 , FC 3,3 , FC 3,4 , ……FC 3，1000 ,

[0192] …

[0193] FC 80,1 , FC 80,2 , FC 80,3 , FC 80,4 , ……FC 80，1000

[0194] 2. Focus imaging.

[0195] First, perform Fourier transform on the array FC column by column, that is,

[0196] (FC 1,1 , FC 2,1 , FC 3,1 , ……FC 80,1 ),

[0197] (FC 1,2 , FC 2,2 , FC 3,2 , ……FC 80,2 )……

[0198] arrive

[0199] (FC 1,1000 , FC 2，1000 , FC 3，1000 , ……FC 801000 )

[0200] Do the Fourier transform and then rotate back to the original direction to get the FCF array:

[0201] FCF 1,1 , FCF 1,2 , FCF 1,3 , FCF 1,4 , ...FCF 1，1000 ,

[0202] FCF 2,1 , FCF 2,2 , FCF 2,3 , FCF 2,4 , ...FCF 2，1000 ,

[0203] FCF 3,1 , FCF 3,2 , FCF 3,3 , FCF 3,4 , ...FCF 3，1000 ,

[0204] …

[0205] FCF 80,1 , FCF 80,2 , FCF 80,3 , FCF 80,4 , ...FCF 80,1000

[0206] This operation makes the row where the target's azimuth is located the peak of the FFT result. Therefore, in the FCF array, the horizontal direction represents the target signal (range information) in the upward direction, while the vertical direction represents the azimuth information (angle information) of the target signal.

[0207] Create an array ANG whose elements are:

[0208] arcsin(x*0.055 / (2*0.0125)), where x decreases from 1 to -1, with a step length of 2 / 80, or 0.025;

[0209] The size of the ANG array is 80 elements, corresponding one to one to the column elements of the FCF array.

[0210] Since the range of the sine function is limited to [-1,1], the elements of the ANG array are valid only when x is between 0.025 / 0.055 and -0.025 / 0.055. The actual ANG array is:

[0211] ANG=[n / a,n / a,……1.3453,1.1655,1.04175……-1.04175,-1.1655,-1.3453……n / a,n / a]

[0212] There are 22 invalid elements in the front and back.

[0213] According to the validity of the ANG array elements, rows 1 to 22 and rows 59 to 80 in the FCF array are deleted, leaving only rows 23 to 58, and then only elements 23 to 58 of the ANG array are retained.

[0214] At this time, the size of the FCF array becomes 36 rows and 1000 columns. Each row corresponds to an element of the ANG array. The corresponding value in the ANG array is the azimuth of the point. Each column corresponds to a distance from 0 to 500 meters with a step size of 0.5 meters, that is, a complex image in polar coordinates.

[0215] Since the horizontal axis of the array corresponds to the distance, the amplitude of each element represents the reflection intensity of the target, the phase angle represents the distance phase of the target, and the row number represents the azimuth angle α of the target, if you want to convert the data result into Cartesian coordinates, you can use the following method:

[0216] X=D*cosα

[0217] Y=-D*sinα

[0218] In this embodiment, before the present invention is used, the time for one scan is 1 / 0.05=20 seconds, while after the present invention is used, the time for one scan is only 0.125 / 0.05=2.5 seconds. The speed is increased to 8 times of the original speed, and the track length is only 1 / 8 of the original speed. The overall volume and weight of the radar are greatly reduced, the reliability is improved, and the maintenance cost is also significantly reduced.

[0219] This embodiment is only for explaining the implementation method, and the parameters are not carefully optimized.

[0220] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A method for target detection using radar, characterized in that: The radar comprises an antenna, a receiving system, a transmitting system and a short track; the short track is provided with a receiving system capable of reciprocating thereon; Or a short track is provided with a launching system capable of reciprocating thereon; The method for target detection using radar comprises: Perform data acquisition to obtain original radar echo data arranged in the acquisition space order; The original radar echo data is subjected to focused imaging by antenna array phase shaping, that is, the original radar echo data is subjected to Fourier transform to obtain radar echo data consisting of an array converted into the frequency domain; Use the set antenna phase delay parameters to perform delay correction on the array to obtain the corrected radar echo data; Perform spatial Fourier transform on the corrected data to obtain converted radar echo data; Determine the actual angle of the converted radar echo data indicating the azimuth and obtain an image; During data collection, the moving interval of the track is defined as the antenna spacing a, the moving step number is k, and the track moving interval d is smaller than the wavelength λ, and d is set to be smaller than λ / 10; When data collection starts, the track drives the transmitting antenna or the receiving antenna from the starting point, and moves d each time. The radar scans once and collects data once until the end point. At this time, the track moves a distance; the collected data is stored as: Starting position 0×d:A 1,1 , A 2,1 , A 3,1 ……A n,1 , Orbital position 1×d: A 1,2 , A 2,2 , A 3,2 ……A n,2 , Orbital position 2×d:A 1,3 , A 2,3 , A 3,3 ……A n,3 , …… End position (k-1)×d=a:A 1,k , A 2,k , A 3,k ……A n,k Each unit A is an array, and the length of the array is related to the scanning time. If the scanning time is t and the sampling rate is sp, the array length is the product of t and sp; each row is a radar scan when the track stays at a position, and each column corresponds to a receiving antenna. n is the number of receiving antennas, and k is the number of moving steps, that is, an array of k rows and n columns is obtained, and the length of each array is t×sp; Starting from the upper left corner, from top to bottom, and then from left to right, take out the A elements in the array matrix one by one, reorganize the original data, and expand it into one dimension: A 1,1 ,A 1,2 ,A 1,3 ,…… A 1,k , A 2,1 ,A 2,2 ,A 2,3 ,……A 2,k , ………… A n,1 , A n,2 , A n,3 ……A n,k So far, n×k original radar echo data arranged in the order of acquisition space are obtained, and the length of each echo data is t×sp; because each A is actually an array of length t×sp: A= B(1, 2, 3, 4, 5... t×sp) Therefore, if each element of A is expanded into B, the above one-dimensional array is actually an array B with a first dimension of n×k and a second dimension of t×sp: B 1,1 ,B 1,2 ,B 1,3 ,B 1,4 ,……B 1，t×sp , B 2,1 ,B 2,2 ,B 2,3 ,B 2,4 ,……B 2，t×sp , B 3,1 ,B 3,2 ,B 3,3 ,B 3,4 ,……B 3，t×sp , …… B n×k,1 ,B n×k,2 ,B n×k,3 ,B n×k,4 ,……B n×k, t×sp This data is equivalent to the data obtained by a single receiving antenna orbital radar with a track length of a×n, and the observation period is shortened to 1 / n of that of the traditional orbital radar.

2. The method according to claim 1, characterized in that The specific method of focusing imaging the original observation data by antenna array phase shaping is First, perform Fourier transform on each row of array B to convert each row from the time domain of the original information to the frequency domain: F=fft(B) Get the new array: F 1,1 ,F 1,2 ,F 1,3 ,F 1,4 ,……F 1，t×sp , F 2,1 ,F 2,2 ,F 2,3 ,F 2,4 ,……F 2，t×sp , F 3,1 ,F 3,2 ,F 3,3 ,F 3,4 ,……F 3，t×sp , …… F n×k,1 ,F n×k,2 ,F n×k,3 ,F n×k,4 ,……F n×k, t×sp Different elements in each row of the F array represent corresponding FFT bins, whose modulus means the reflection intensity of the target at the distance, and whose phase angle represents the distance phase of the target.

3. The method according to claim 1, characterized in that The specific method of using the stored antenna phase delay parameters to perform delay correction on the F array is: The F array has n×k rows, where the data from row number 1 to k is obtained by antenna No. 1, the data from row k+1 to 2×k is obtained by antenna No. 2, and the data from row number (n-1)×k to n×k is obtained by antenna No. n; The correction method is to divide the data of each row by the correction value of the corresponding antenna in the correction parameter array C, that is, the F data with row numbers (n-1)×k to n×k is divided by the nth element in the C array. Complex number division is geometrically equivalent to correcting the phase angle; each element of the F data with row numbers 1 to k is divided by the first element of the correction parameter array C; The corrected data FC is obtained: FC 1,1 ,FC 1,2 ,FC 1,3 ,FC 1,4 ,……FC 1，t×sp , FC 2,1 ,FC 2,2 ,FC 2,3 ,FC 2,4 ,……FC 2，t×sp , FC 3,1 ,FC 3,2 ,FC 3,3 ,FC 3,4 ,……FC 3，t×sp , …… FC n×k,1 ,FC n×k,2 ,FC n×k,3 ,FC n×k,4 ,……FC n×k, t×sp 。 4. The method according to claim 1, characterized in that The same target is in the same position in each row of data, that is, the same FFT well. Therefore, the phase difference between two adjacent rows of sampled data is d×sinα / λ. Perform spatial Fourier transform on the array FC column by column, that is, (FC 1,1 , FC 2,1 , FC 3,1 , ……FC n×k,1 ),FC 1,2 , FC 2,2 , FC 3,2 , ……FC n×k,2 ) ... to (FC 1,t×sp , FC 2，t×sp , FC 3，t×sp , ……FC n×k ,t×sp ) to perform Fourier transform, and then rotate back to the original direction to obtain the FCF array: FCF 1,1 ,FCF 1,2 ,FCF 1,3 ,FCF 1,4 ,……FCF 1，t×sp , FCF 2,1 ,FCF 2,2 ,FCF 2,3 ,FCF 2,4 ,……FCF 2，t×sp , FCF 3,1 ,FCF 3,2 ,FCF 3,3 ,FCF 3,4 ,……FCF 3，t×sp , …… FCF n×k,1 ,FCF n×k,2 ,FCF n×k,3 ,FCF n×k,4 ,……FC n×k, t×sp The row where the target's azimuth is located is the peak of the FFT result. In the FCF array, the horizontal direction represents the upward target signal, that is, the distance information, while the vertical direction represents the azimuth information of the target signal, that is, the angle, that is, the azimuth information.

5. The method according to claim 1, characterized in that Determine the actual angle of the azimuth represented by the longitudinal direction of the FCF array and obtain the image. The specific method is: Create an array ANG of size n×k, with the following contents: arcsin(x×λ / (2×d)), where x decreases from 1 to -1 with a step size of 2 / (n×k) The ANG array is a series of angles, which correspond one-to-one to the elements of each column of the FCF array. The ANG array is used as the vertical axis scale, ranging from -90 degrees to 90 degrees, and the distance is used as the horizontal axis scale, ranging from 0 to the maximum distance of the radar. The FCF array values ​​are plotted to obtain the focused image. The value range of the sin function is in [-1,1], and there will be outliers in the ANG array. If d<λ / 2 is satisfied, there will be some outliers at the beginning and end of the array. Delete the outliers and the corresponding rows in the FCF array at their corresponding positions. At this time, the image is in polar coordinates; since the horizontal coordinate of the array corresponds to the distance, the amplitude of each element represents the reflection intensity of the target, the phase angle represents the distance phase of the target, and the value of the ANG array element corresponding to the row number represents the azimuth angle α of the target, if the data result is converted into Cartesian coordinates, the following method is used: X = D × cos α Y = -D × sin α.

6. The method according to claim 1, characterized in that The short track is provided with a receiving system capable of reciprocating thereon, and the length of the short track is greater than or equal to the spacing between the antennas.

7. The method according to claim 1, characterized in that The short track is provided with a transmitting system capable of reciprocating thereon, and the length of the short track is greater than or equal to the spacing between antennas.

8. The method according to claim 1, characterized in that The short track is provided with a receiving system and a transmitting system capable of reciprocating thereon, and the length of the short track is greater than or equal to the antenna spacing / 2.

Images