A high-precision velocity and azimuth joint measurement method for MIMO radar

Abstract

This invention discloses a high-precision velocity and azimuth joint measurement method for MIMO radar, comprising: 1. Modeling of MIMO radar system echo signals; 2. Acquisition of one-dimensional range profiles across all channels for all periods; 3. Enhancement of one-dimensional range profiles for each period; 4. Range-oriented target detection based on CFAR; 5. Detection of the range cell containing the target center; 6. Linear detection of the range cell containing the target center; 7. Coarse estimation of the target's radial velocity; 8. Phase compensation across all channels for all periods; 9. Acquisition of signals from all range cells containing the target; 10. Precise joint estimation of target velocity and azimuth; 11. Final result of the target's radial velocity. This invention obtains the target's range by detecting changes in the range cells within the one-dimensional range profile and estimates the coarse radial velocity, ultimately obtaining a high-precision velocity estimation result for the target. For MIMO arrays, it fully utilizes the spatiotemporal characteristics of the target contained in the echo signals of spatially overlapping virtual array elements, improving the accuracy of joint velocity and angle measurement of moving targets.

Description

Technical Field

[0001] This invention belongs to the field of millimeter-wave radar signal processing technology, specifically relating to a high-precision velocity and azimuth joint measurement method for MIMO radar. Background Technology

[0002] In recent years, millimeter-wave radar technology has developed rapidly and plays an important role in smart cities, intelligent transportation, and environmental monitoring. Millimeter-wave radar operates at high frequencies, has short wavelengths, and small antenna sizes, offering significant advantages such as large bandwidth, high range resolution, no limitations imposed by lighting conditions, avoidance of privacy exposure, high system integration, and convenient application.

[0003] In existing methods, for static or low-speed moving targets, there is no target migration across radial range cells within a certain observation time that meets the requirements of high velocity resolution. Signal processing based on MIMO radar can achieve angle and velocity estimation by processing the signals of the same range cell where the target is located. Specifically, it can be achieved by using commonly used Fourier transform methods and spectral estimation methods such as MUSIC with super-resolution performance.

[0004] However, for dynamic targets, especially high-speed vehicles on roads, the range cell in high-resolution range-resolution observations will change over time within a short observation period. For MIMO radar, this change occurs not only during different velocity accumulation periods but also during the short time it takes for all the radar's transmitting antennas to sequentially transmit signals for measurement. This increases the difficulty of obtaining high-precision target parameter information from the observation data.

[0005] In order to achieve high resolution in azimuth, millimeter-wave radar systems usually use MIMO design to expand the number of channels in the azimuth direction, so as to realize range, velocity and angle measurement [1][2]. For millimeter-wave radar applied to moving target measurement scenarios, there are usually problems such as small maximum unambiguous velocity, inaccurate velocity measurement, and coupling between velocity, angle and distance in the measurement, which restricts the expansion of the application range of MIMO millimeter-wave radar systems. Due to the constraints of the working mode of MIMO array antenna and signal processing method, in the existing methods, the velocity measurement of moving targets usually needs to be achieved by the deambiguation method [3][4]. Moreover, since there is coupling between distance, velocity and angle in the echo signal of moving targets in the observation, the measurement accuracy of parameters such as target distance, velocity and angle is limited.

[0006] This invention proposes a method for simultaneous estimation of target azimuth and velocity based on the joint temporal and spatial characteristics of each channel of a MIMO radar system by modeling the echo signal. This method overcomes the limitation of the maximum unambiguous velocity in conventional MIMO radar processing and simultaneously employs a super-resolution method to improve angle measurement accuracy. Therefore, this method has the advantages of a large maximum unambiguous velocity and high measurement accuracy.

[0007] References

[0008] [1] Ma Yufang. Research on positioning and velocity measurement method of dual-base circular array MIMO radar based on parallel factor. Master's thesis of Xidian University, December 2015.

[0009] [2] Ge You, Liu Jingping, Zhao Huichang, Chen Si, Liu Dingye. MIMO radar velocity and ranging algorithm based on orthogonal phase coded signal. Journal of Detection and Control, 2016, 38(6), pp:80-83.

[0010] [3] Ran Yanwei, Jin Sheng, Liang Xiaohu, et al. A method for high-speed target velocity measurement and positioning across range gates in bistatic MIMO radar, Chinese Invention Patent, Patent No. ZL201610709772.1, Application Date 2016.08.23, Authorization Announcement Date 2019.03.29.

[0011] [4] Huang Yulin, Zhang Yin, Zhu Peixi, et al. A method for extending the multi-target velocity measurement of MIMO radar, Chinese Invention Patent, Application No. 202111507596.0, Application Date 2021.12.10, Publication Date 2022.3.18. Summary of the Invention

[0012] The technical problem to be solved by this invention is: Based on the spatial, temporal, and frequency characteristics of the target echo signal of a sparse MIMO radar system, this invention designs a method that utilizes all virtual antenna elements of a MIMO radar system, including spatially overlapping virtual antenna array elements, to integrate the temporal, spatial, and frequency signal characteristics of all channels, thereby achieving joint high-precision measurement of target distance, velocity, and angle by MIMO radar.

[0013] The MIMO radar high-precision velocity and azimuth joint measurement method proposed in this invention has significant technical advantages over existing technologies in improving the accuracy of target velocity and azimuth joint measurement. These advantages mainly include: (1) By estimating the coarse radial velocity of the target, the mismatch problem that may be caused by de-ambiguity in conventional methods is avoided. This method obtains the change in the radial migration range cell of the target by enhancing the original one-dimensional range image of all channels, target detection, discrimination of the range cell where the target center is located, and line detection. Combined with the radar observation time, a coarse estimate of the target's radial velocity is calculated. (2) It can realize the joint accurate estimation of the target's radial velocity and azimuth, making full use of the target's spatiotemporal characteristics contained in the echo signal of the spatially overlapping virtual array elements, thereby improving the accuracy of joint velocity and angle measurement of moving targets.

[0014] The technical solution adopted in this invention is as follows:

[0015] Step 1: Modeling the echo signal of the MIMO radar system

[0016] Based on the antenna structure of the MIMO radar system and the time-division multiplexing of signals transmitted by each transmitting antenna, and the simultaneous reception of target echo signals by each receiving antenna, a space-time observation echo signal model of the MIMO radar system is established using a MIMO radar system model based on the frequency-modulated continuous wave (FMCW) transmission waveform. The specific method is as follows:

[0017] The transmitting antenna Tx transmits linear frequency modulated signals sequentially, while all receiving antennas Rx are used simultaneously to receive the target echo signals.

[0018] Suppose a target in a given observation scenario is located at an angle θ away from directly in front of the radar, with an initial distance of R0, and its radial velocity relative to the radar is v. and Transmitting antenna and Time-division transmission of signals, located in and Receiving antenna and Simultaneously, it receives echo signals from the target.

[0019] For amplitude A t The relationship between the transmitted signal st(t) of a linear frequency modulated continuous wave radar with an initial frequency of f0 and a modulation slope of γ and time t can be expressed as follows:

[0020]

[0021] Among them, the imaginary unit For a target at a distance R from the radar, the echo signal s r (t) is

[0022]

[0023] Among them, A r The echo amplitude is given by τ = 2(R0 - vt) / c, the signal delay time is given by c, and the speed of light is given by c.

[0024] The intermediate frequency (IF) output signal is obtained after mixing the echo signal and the transmitted signal. For MIMO radar, taking into account the target echo direction, the path difference of each channel, and the target's velocity, and assuming that the amplitude of the IF echo signal of each target is 1, then the IF signal... It can be represented as

[0025]

[0026] Where, x m and x n These represent the positions of the transmitting and receiving antennas, with the electromagnetic wave wavelength λ as a reference. This indicates that the displacement of the m-th transmitting antenna relative to the reference transmitting antenna is half a wavelength. Times, N Tx The number of transmitting antennas, This indicates that the displacement of the nth receiving antenna relative to the reference receiving antenna is half a wavelength. times, t p Let t be the start time of the p-th frame during observation. p = (p-1)·T(p=1,2,…,N) Loop ), T = T LFM ·N Tx t represents the duration of each observation period. m For each frame located at x m The transmit antenna relative to the frame start time t p Signal transmission start time, t m =m·T LFM (m=1,2,…,N Tx -1), T LFM For the period of a linear frequency modulated signal, Then it is located at x n The fast time of signal reception by the receiving antenna. Considering the radial motion of the target, assuming that the target motion is negligible within a single linear frequency modulation cycle, then the echo delay τ in (3) mnp It can be represented as

[0027]

[0028] Where, x m0 and x n0 R1 and R2 are the reference positions of the transmitting and receiving antennas when calculating the path difference of each channel, respectively, and R0 is the initial distance to the target.

[0029] The phase of equation (3) is expressed as

[0030]

[0031] Meanwhile, the phase in equation (5) that does not change with fast time is denoted as

[0032]

[0033] Under normal circumstances, Therefore,

[0034]

[0035] The frequency f of the intermediate frequency echo signal can be obtained from equation (5). IF (x m ,x n ,t p ,t m )for

[0036]

[0037] Substituting equation (4) into equation (8) yields

[0038]

[0039] The target echo intermediate frequency f of the MIMO reference channel at the initial observation time. IF,ini,ref

[0040]

[0041] Using this as a reference, the frequency difference Δf of the intermediate frequency signal IF (x m ,x n ,t p ,t m )for

[0042]

[0043] Therefore, in the calculation of one-dimensional high-resolution range profiles based on Fourier transform, the phase difference of the range cell where the target is located is caused by the mid-frequency difference. It can be represented as

[0044]

[0045] Among them, T LFM This represents the period of a single linear frequency modulated signal.

[0046] Therefore, for a target located at an angle θ away from directly in front of the radar, with a relative radial velocity v, the constant phase component represented by R0 is removed, and the positions of the transmitting and receiving antennas x are set... m and x n relative displacement and It can be shown from equations (7) and (12) that the total phase difference ΔΦ(x) of the distance cell in which it is located is... m ,x n ,t p ,t m ;θ,v) can be represented as

[0047]

[0048] Therefore, it can be seen from the phase of each target range unit observed at different times, as expressed by equation (13), that the target's azimuth and radial velocity relative to the radar can be estimated simultaneously based on this model.

[0049] Step 2: Acquisition of one-dimensional distance images across all channels for all periods

[0050] Pulse compression is performed on all channel echo signals obtained from multiple transmissions of signals from all transmitting antennas to obtain a one-dimensional range image Y(x). m ,x n ,t p ,t m ,i), where i represents the i-th distance unit;

[0051] Step 3: Enhancement of one-dimensional range images in each period

[0052] To improve the signal-to-noise ratio and target detection accuracy, the amplitudes of the one-dimensional range images obtained from all receiving channels after transmitting a signal from all transmitting antennas are summed and superimposed according to the range cell to obtain an enhanced one-dimensional range image.

[0053]

[0054] Step 4: Range-based target detection based on CFAR

[0055] For all the obtained augmented one-dimensional distance images Target detection is performed along the range direction using a constant false alarm rate (CFAR) method, and the range cell I(t) occupied by the target is marked. p ,i), is represented as

[0056]

[0057] Step 5: Detect the distance between the target center and the cell.

[0058] The target detection result I(t)p i) with enhanced one-dimensional distance image Combined, the peak distance unit of the enhanced one-dimensional range image in the target detection result is obtained. Represented as

[0059]

[0060] Step 6: Linearity detection of the target center at distance from the cell

[0061] For the obtained enhanced one-dimensional distance image Enhanced one-dimensional range image target range cell line detection based on Hough transform, for the detected first... For each target, the line containing its distance cell can be represented as a function of the observation period number p and the distance cell number i, i.e.

[0062]

[0063] in, For the detected first The one-dimensional distance from the line containing the target is shown in the image. point distance, For a straight line relative to a one-dimensional distance image The angle of rotation in the first row.

[0064] Step 7: Coarse estimation of the target's radial velocity

[0065] Based on the rotation angle of the straight line where the distance cell corresponding to each target changes trajectory. Obtain a coarse estimate of the velocity for each target. Represented as

[0066]

[0067] Where Δr is the total number of observation periods P of the target. total Change in inner radial distance.

[0068] Step 8: Phase compensation for all channels in all cycles

[0069] Based on the obtained rough estimates of the velocity of each target According to the observation time (t) p ,t m Frequency correction is performed on the echoes of all channels for each target to achieve range alignment. Compensation frequency for each target Represented as;

[0070]

[0071] Compensated signal for

[0072]

[0073] Step 9: Acquisition of distance cell signals for all targets

[0074] Echo signal after phase alignment Obtain the target location Complex signals on a distance unit That is, the first The signal of each target in each channel

[0075] Represented as

[0076]

[0077] Step 10: Precise joint estimation of target velocity and azimuth

[0078] Based on the spatiotemporal model constructed in step one, represented by equations (3) and (4), and the complex signals on each target channel obtained in step nine. The target's velocity and high-resolution angle are obtained using spectral estimation. The specific method is as follows.

[0079] For an array observation system whose observation sensor consists of M antenna channels, the observation data can be represented as:

[0080] x(t)=As(t)+n(t) (22)

[0081] Where x(t) is the array sensor snapshot observation data vector with dimension M×1, s(t) is the target signal vector with dimension K×1, and n(t) is the noise vector with dimension M×1, respectively represented as follows:

[0082] x(t) = [x1(t), ..., x m (t),…,x M (t)] T (twenty three)

[0083] s(t)=[s1(t),…,s k (t),…,s K (t)] T (twenty four)

[0084] n(t) = [n1(t), ..., n m (t),…,n M (t)] T (25)

[0085] Where A is the guidance matrix. For an observation scenario with K targets, the dimension of A is M×K, which can be expressed as:

[0086] A = [a1, ..., a k ,…,a K (26)

[0087] Wherein, the guidance vector a of the k-th target k , can be represented as

[0088] a k =[a 1,k ,…,a m,k ,…,a M,k ] T (27)

[0089] When the direction of arrival of the k-th target echo signal is θ k When the first element of the sensor array is taken as the reference element, the steering vector a k It can be represented as

[0090]

[0091] Where λ is the wavelength of the electromagnetic wave and d is the spacing between the antenna array elements.

[0092] For the observation scenario represented by equation (22), when the azimuth angle of the kth target is θ k The radial velocity is v k When, the guiding vector a represented by equation (28) k Then equation (13) is updated to

[0093]

[0094] It can be seen that at this time, a k It is of dimension N Tx N Rx N Loop A vector of size 1, where N Tx and N Rx N represents the number of transmitting antennas and the number of receiving antennas, respectively. Loop It is the number of consecutive observation cycles, that is, the number of times each transmitting antenna transmits a signal.

[0095] With R x =E{x(t)x H Let (t)} represent the covariance matrix of the observed data. The Capon method can be used to obtain the result as a function of echo angle. and speed Changing spatial spectrum Represented as

[0096]

[0097] Among them, equation (29)a(θ) k ,v k ) in θ k and v k Replace with and You can get for The conjugate transpose of . For R x The inverse matrix. The angle range is... The accurate speed estimation range is

[0098] Therefore, the first From the perspective of a goal and speed The precise estimation result is

[0099]

[0100] Step 11: Final Result of Target Radial Motion Velocity

[0101] The rough estimate of the velocity of each target With the results of the precise velocity estimation The summation yields an accurate estimate of the target's actual radial velocity, i.e., the target's actual radial motion velocity. for

[0102]

[0103] The beneficial effects of the technical solution of this invention are as follows:

[0104] The high-precision velocity-azimuth joint measurement method for MIMO radar proposed in this invention, compared with existing MIMO radar velocity-azimuth measurement methods, utilizes all antenna channels, including spatially overlapping virtual channels, thus improving the accuracy of joint velocity and azimuth measurement. Its main technical advantages include:

[0105] (1) Without increasing the complexity of the transmitted signal, the target distance is obtained by detecting the change of the target's distance cell in the one-dimensional range image, and the coarse radial velocity is estimated. Then, the target's high-resolution azimuth estimate is obtained through phase compensation, while the target's radial velocity is precisely estimated, and finally the target's high-precision velocity estimation result is obtained.

[0106] (2) For MIMO arrays, the conventional method of processing spatially overlapping virtual array elements is broken through. The target spatiotemporal characteristics contained in the echo signal of spatially overlapping virtual array elements are fully utilized, which improves the accuracy of joint velocity and angle measurement of moving targets. Attached Figure Description

[0107] Figure 1 This is a schematic diagram of an antenna array.

[0108] Figure 2 This is a schematic diagram of the target echo direction.

[0109] Figure 3 This is a flowchart of signal processing.

[0110] Figure 4 This is a MIMO radar observation scenario.

[0111] Figure 5 It is a multi-period, full-channel, one-dimensional range image.

[0112] Figure 6 Enhance the one-dimensional range image for each period.

[0113] Figure 7 This is the result of the distance-oriented CFAR detection.

[0114] Figure 8 This represents the peak detection result based on CFAR in the distance direction.

[0115] Figure 9 This is the result of the line detection for the target's location within the distance cell.

[0116] Figure 10 The target measurement result. Detailed Implementation

[0117] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0118] The specific steps of the high-precision velocity-azimuth joint measurement method for MIMO radar of the present invention are as follows:

[0119] Step 1: Modeling the echo signal of the MIMO radar system

[0120] Based on the antenna structure of the MIMO radar system and the time-division multiplexing of signals transmitted by each transmitting antenna, and the simultaneous reception of target echo signals by each receiving antenna, a space-time observation echo signal model of the MIMO radar system is established using a MIMO radar system model based on the FMCW transmitted waveform. In this embodiment, the following is used... Figure 1 The MIMO radar antenna structure shown contains 9 transmitting antennas and 16 receiving antennas, which can be combined to form 144 antenna channels. However, spatially, this results in only 86 non-overlapping virtual antenna element positions. Conventional methods only process these 86 virtual antenna channels to obtain the target's azimuth estimate. In this invention, all 144 channels are used for joint estimation of target velocity and azimuth through a space-time combined method, thereby improving the measurement accuracy of velocity and azimuth. The specific method is described below.

[0121] Assuming the observation scenario is as follows Figure 2 As shown, the target is located at an angle θ away from directly in front of the radar, with an initial distance of R0 and a radial velocity v relative to the radar. Nine transmitting antennas Tx sequentially transmit linear frequency modulated signals, while sixteen receiving antennas Rx simultaneously receive the target echo signal. The radar system sampling rate f... s =8MHz, number of effective sampling points N s =256.

[0122] For amplitude A t =1, the starting frequency is f0 = 77GHz, and the frequency modulation slope is γ = 39 × 10 12 The linear frequency modulated continuous wave radar transmission signal of Hz / s can be expressed as:

[0123]

[0124] For a target with an initial distance of R0 = 50m to the radar, a radial velocity of v = 25m / s, and an echo signal direction of θ = 30°, the echo signal is:

[0125]

[0126] Where τ=2(R0-vt) / c is the signal delay time, and c=3×10 8 m / s is the speed of light.

[0127] The intermediate frequency (IF) output signal is obtained after mixing the echo signal and the transmitted signal. For MIMO radar, taking into account the target echo direction, the path difference of each channel, and the target's velocity, and assuming that the amplitude of the IF echo signal of each target is 1, the IF signal can be expressed as:

[0128]

[0129] Where, x m and x n These are the locations of the transmitting and receiving antennas, respectively. This indicates that the displacement of the m-th transmitting antenna relative to the reference transmitting antenna is m times half the wavelength, N. Tx The number of transmitting antennas, This indicates that the displacement of the nth receiving antenna relative to the reference receiving antenna is half a wavelength. times, t p N is the start time of each frame during observation. When each frame accumulates 64 observation cycles, N... Loop =64,t p =p·T (p=0,1,2,…,63),t m For each frame located at x m The transmit antenna relative to the frame start time tp The signal transmission start time,

[0130] t m =m·T LFM (m=0,1,2,…,8),T LFM =45μs is the period of the linear frequency modulated signal. Then it is located at x n The fast time of signal reception by the receiving antenna. Considering the radial motion of the target, assuming that the target motion is negligible within a single linear frequency modulation cycle, the echo delay τ in (3) can be expressed as

[0131]

[0132] Where, x m0 and x n0 These are the reference positions for the transmitting and receiving antennas when calculating the path difference for each channel.

[0133] The phase of equation (3) is expressed as

[0134]

[0135] Meanwhile, the phase in equation (5) that does not change with fast time is denoted as

[0136]

[0137] Under normal circumstances, Therefore,

[0138]

[0139] From equation (5), the frequency of the intermediate frequency echo signal can be obtained as follows:

[0140]

[0141] Substituting equation (4) into equation (8) yields

[0142]

[0143] The target echo intermediate frequency of the MIMO reference channel at the initial observation time.

[0144]

[0145] Using this as a reference, the frequency difference Δf of the intermediate frequency signal IF (x m ,x n ,t p ,t m )for

[0146]

[0147] Therefore, in the calculation of one-dimensional high-resolution range images based on Fourier transform, the mid-frequency difference Δf IF (x m ,x n ,t p ,t m The phase difference of the target's range cell caused by ) can be expressed as

[0148] Φ ΔfIF (x m ,x n ,t p ,t m )=2πΔf IF (x m ,x n ,t p ,t m )·T LFM (12)

[0149] Among them, T LFM This represents the period of a single linear frequency modulated signal.

[0150] Therefore, for a target with θ = 30° and v = 25 m / s, the constant phase component represented by R0 = 50 m is removed, and the positions of the transmitting and receiving antennas x are... m and x n relative displacement and

[0151] It can be shown from equations (7) and (12) that the total phase difference of the distance cell can be expressed as follows:

[0152]

[0153] Therefore, it can be seen from the phase of each target range unit observed at different times, as expressed by equation (13), that the target's azimuth and radial velocity relative to the radar can be estimated simultaneously based on this model.

[0154] Step 2: Acquisition of one-dimensional distance images across all channels for all periods

[0155] for Figure 4 The MIMO radar observation scenario shown is based on Figure 3 The signal flow diagram shown compresses all channel echo signals obtained from multiple transmissions by all transmitting antennas to acquire a one-dimensional range profile. Where i represents the i-th distance unit, such as Figure 5 As shown.

[0156] Step 3: Enhancement of one-dimensional range images in each period

[0157] according to Figure 3 The signal flow diagram shown illustrates how all receiving channels receive signals after each transmission from all transmitting antennas. Figure 5 The one-dimensional range image shown is summed and superimposed according to the range cell to obtain an enhanced one-dimensional range image. The results are as follows Figure 6 As shown,

[0158]

[0159] Step 4: Range-based target detection based on CFAR

[0160] according to Figure 3 The signal flowchart shown represents all the obtained enhanced one-dimensional range images. Target detection is performed along the range direction using a constant false alarm rate (CFAR) method, and the range cell I(t) occupied by the target is marked. p ,i), is represented as

[0161]

[0162] Test results as follows Figure 7 As shown.

[0163] Step 5: Detect the distance between the target center and the cell.

[0164] The target detection result I(t) p i) with enhanced one-dimensional distance image Combined, the peak distance unit of the enhanced one-dimensional range image in the target detection result is obtained. Represented as

[0165]

[0166] The center distance cell corresponding to the target location is as follows Figure 8 As shown.

[0167] Step 6: Linearity detection of the target center at distance from the cell

[0168] For the obtained enhanced one-dimensional distance image Enhanced one-dimensional range image target range cell line detection based on Hough transform, for the detected first... For each target, the straight line containing the distance cell can be represented as:

[0169]

[0170] in, For the detected first The one-dimensional distance from the line containing the target is shown in the image. point distance, For a straight line relative to a one-dimensional distance image The rotation angle of the first row. The line detection results corresponding to the distance cell of the target are as follows. Figure 9 As shown.

[0171] Step 7: Coarse estimation of the target's radial velocity

[0172] Based on the rotation angle of the straight line where the distance cell corresponding to each target changes trajectory. Obtain a coarse estimate of the velocity for each target. Represented as

[0173]

[0174] For the assumed target in the example, data was acquired for a total of P = 64 periods, at observation time t. p＝P -t p＝1 Within (64-1)×9×45μs=0.025515s, the theoretical value of the change in the target's radial distance is 25×0.025515=0.637875m. For a frequency modulation slope γ=39×10 12 Hz / s, sampling rate f s =8MHz, number of effective sampling points N s =256, Fourier transform points. From the one-dimensional range image, the distance change of the target cell can be obtained as Δr = 0.637875m.

[0175] Step 8: Phase compensation for all channels in all cycles

[0176] Based on the obtained rough estimates of the velocity of each target According to the observation time (t) p ,t m Phase compensation is performed on all channel echoes of each target to achieve range alignment. The compensated phase is represented as follows:

[0177]

[0178] Compensated signal for

[0179]

[0180] Step 9: Acquisition of distance cell signals for all targets

[0181] Echo signal after phase alignment Obtain the target location Complex signals on a distance unit That is, the first The signal of each target in each channel Represented as

[0182]

[0183] Step 10: Precise joint estimation of target velocity and azimuth

[0184] Based on the spatiotemporal model constructed in step one and the complex signals on each channel of the target obtained in step nine. The target's velocity and high-resolution angle are obtained using spectral estimation. The specific method is as follows.

[0185] For an array observation system whose observation sensor consists of M antenna channels, the observation data can be represented as:

[0186] x(t)=As(t)+n(t) (22)

[0187] Where t is the observation time, x(t) is the array sensor snapshot observation data vector with dimension M×1, s(t) is the target signal vector with dimension K×1, and n(t) is the noise vector with dimension M×1, respectively represented as follows:

[0188] x(t) = [x1(t), ..., x m (t),…,x M (t)] T (twenty three)

[0189] s(t)=[s1(t),…,s k (t),…,s K (t)] T (twenty four)

[0190] n(t) = [n1(t), ..., n m (t),…,n M (t)] T (25)

[0191] A is the guidance matrix. For an observation scenario with K targets, A has dimensions M×K and can be represented as follows:

[0192] A = [a1, ..., a k ,…,a K (26)

[0193] Wherein, the guidance vector a of the k-th target k , can be represented as

[0194] a k =[a 1,k ,…,a m,k ,…,a M,k ] T (27)

[0195] When the target's azimuth angle is θ k When the first element of the sensor array is taken as the reference element, the steering vector a k It can be represented as

[0196]

[0197] Where λ is the wavelength of the electromagnetic wave, and θ k Let d be the direction of arrival of the echo signal from the k-th target, and d be the spacing between the antenna array elements.

[0198] For the observation scenario represented by equation (22), when the azimuth angle of the kth target is θ k The radial velocity is v k When, the guiding vector a represented by equation (28) k Then equation (13) is updated to

[0199]

[0200] It can be seen that at this time, a k It is of dimension N Tx N Rx N Loop A vector of size 1, where N Tx and N Rx N represents the number of transmitting antennas and the number of receiving antennas, respectively. Loop It is the number of consecutive observation cycles, that is, the number of times each transmitting antenna transmits a signal.

[0201] With R x =E{x(t)x H Let (t)} represent the covariance matrix of the observed data. The spatial spectrum of the Capon method can be expressed as:

[0202]

[0203] The angle range is as follows: The accurate speed estimation range is

[0204] Therefore, the first From the perspective of a goal and speed The precise estimation result is

[0205]

[0206] Step 11: Final Result of Target Radial Motion Velocity

[0207] The rough estimate of the velocity of each target With the results of the precise velocity estimation The summation yields an accurate estimate of the target's actual radial velocity, i.e., the target's actual radial motion velocity. for

[0208]

[0209] For example Figure 4 The observation scene shown Figure 5 This is a multi-period, full-channel, one-dimensional range image, where the horizontal axis ChannelIndex represents the index of all channels arranged in spatiotemporal order throughout the entire observation process, and the vertical axis Distance represents the radial distance. Figure 6 To enhance the one-dimensional range image for each period, the horizontal axis Chirp Index represents the period index of all transmitted signals from all transmitting antennas during the entire observation process, and the vertical axis Distance represents the radial distance. Figures 7-9 These represent the range-direction CFAR detection results, the range-direction CFAR-based peak detection results, and the target location range cell line detection results, respectively. The horizontal axis Chirp Index represents the period index of the transmitted signals of all transmitting antennas throughout the entire observation process, and the vertical axis Distance represents the radial distance. Figure 10 The target measurement results are shown, where the horizontal axis represents the target's lateral distance, the vertical axis represents the target's radial distance, and the target's speed and direction of motion are as follows: Figure 10 As shown. Figure 10 In the diagram, the position of the dot indicates the location of the target at the current measurement moment, and the arrowed line segment connected to the dot indicates the magnitude of the velocity, with the direction of motion consistent with the direction pointed to by the arrow. Figure 10 The thick vertical line in the middle represents the dividing line between the two lanes of traffic in the middle of the road. It should be noted that... Figures 4 to 10 The implementation results given are merely one example of a specific implementation of the present invention.

Claims

1. A high-precision velocity-azimuth joint measurement method for MIMO radar, characterized in that: The specific steps are as follows: Step 1: Modeling the echo signal of the MIMO radar system: Based on the antenna structure of the MIMO radar system and the working mode of each transmitting antenna transmitting signals in a time-division manner and each receiving antenna simultaneously receiving the target echo signal, a space-time observation echo signal model of the MIMO radar system is established by adopting a MIMO radar system model based on the frequency modulated continuous wave (FMCW) transmission waveform. Step 2: Acquisition of one-dimensional distance images across all channels for all periods: Pulse compression is performed on all channel echo signals obtained from multiple transmissions of signals from all transmitting antennas to obtain a one-dimensional range image Y(x). m ,x n ,t p ,t m ,i), where i represents the i-th distance unit; Step 3: Enhancement of one-dimensional range images in each period: By transmitting a signal once from all transmitting antennas and summing the amplitudes of the one-dimensional range images obtained from all receiving channels according to the range cells, an enhanced one-dimensional range image is obtained. Step 4: Range-based target detection based on CFAR: For all the obtained augmented one-dimensional distance images Target detection is performed along the range direction using a constant false alarm rate (CFAR) method, and the range cell I(t) occupied by the target is marked. p ,i); Step 5: Detect the distance between the target center and the cell: The target detection result I(t) p i) with enhanced one-dimensional distance image Combined, the peak distance unit of the enhanced one-dimensional range image in the target detection result is obtained. Step Six: Linearity Detection of the Target Center at Distance to the Cell: For the obtained enhanced one-dimensional distance image Enhanced one-dimensional range image target range cell line detection based on Hough transform, for the detected first... For each target, the line containing the distance cell is represented as a function of the observation period number p and the distance cell number i. Step 7: Coarse estimation of the target's radial velocity: Based on the rotation angle of the straight line where the distance cell corresponding to each target changes trajectory. Obtain a coarse estimate of the velocity for each target. Step 8: Phase compensation for all channels across all cycles: Based on the obtained rough estimates of the velocity of each target According to the observation time (t) p ,t m Frequency correction is performed on the echoes from all channels of each target to achieve range alignment; Step 9: Acquiring the distance cell signals of all targets: Echo signal after phase alignment Obtain the target location Complex signals on a distance unit Step 10: Accurate joint estimation of target velocity and azimuth: The spectral estimation method is used to obtain the target's velocity and high-resolution angle estimation results; Step 11, Final result of the target radial velocity: The rough estimate of the velocity of each target With the results of the precise velocity estimation The summation yields an accurate estimate of the target's actual radial velocity.

2. The MIMO radar high-precision velocity-azimuth joint measurement method according to claim 1, characterized in that: In step one, the specific method is as follows: the transmitting antenna Tx transmits linear frequency modulated signals sequentially, and all receiving antennas Rx are used simultaneously to receive the target echo signal; Suppose a target in a certain observation scenario is located at an angle θ away from the radar's direct front, with an initial distance of R0, and its radial velocity relative to the radar is v; and Transmitting antenna and Time-division transmission of signals, located in and Receiving antenna and Simultaneously receive echo signals from the target; For amplitude A t A linear frequency modulated continuous wave radar signal s with an initial frequency of f0 and a frequency modulation slope of γ is transmitted. t The relationship between (t) and time t is expressed as: Among them, the imaginary unit For a target at a distance R from the radar, the echo signal s r (t) is Among them, A r The echo amplitude is given by τ = 2(R0 - vt) / c, the signal delay time is given by c, and c is the speed of light. The intermediate frequency (IF) output signal is obtained after mixing the echo signal and the transmitted signal. For MIMO radar, taking into account the target echo direction, the path difference of each channel, and the target's velocity, and assuming the amplitude of the IF echo signal of each target is 1, then the IF signal... Represented as Where, x m and x n These represent the positions of the transmitting and receiving antennas, with the electromagnetic wave wavelength λ as a reference. This indicates that the displacement of the m-th transmitting antenna relative to the reference transmitting antenna is half a wavelength. Times, N Tx The number of transmitting antennas, This indicates that the displacement of the nth receiving antenna relative to the reference receiving antenna is half a wavelength. times, t p Let t be the start time of the p-th frame during observation. p = (p-1)·T(p=1,2,…,N) Loop ), T = T LFM ·N Tx t represents the duration of each observation period. m For each frame located at x m The transmit antenna relative to the frame start time t p Signal transmission start time, t m =m·T LFM (m=1,2,…,N Tx -1), T LFM For the period of a linear frequency modulated signal, Then it is located at x n The fast time of signal reception by the receiving antenna; considering the radial motion of the target, assuming that the target motion within a single linear frequency modulation cycle is negligible, then the echo delay τ in equation (3) mnp Represented as Where, x m0 and x n0 These are the reference positions of the transmitting and receiving antennas when calculating the path difference of each channel, respectively, and R0 is the initial distance to the target. The phase of equation (3) is expressed as Meanwhile, the phase in equation (5) that does not change with fast time is denoted as Therefore, The frequency f of the intermediate frequency echo signal is obtained from equation (5). IF (x m ,x n ,t p ,t m )for Substituting equation (4) into equation (8) yields The target echo intermediate frequency f of the MIMO reference channel at the initial observation time. IF,ini,ref Frequency difference Δf of intermediate frequency signal IF (x m ,x n ,t p ,t m )for In one-dimensional high-resolution range profile calculation based on Fourier transform, the phase difference of the range cell containing the target caused by mid-frequency differences... Represented as Among them, T LFM Represents the period of a single linear frequency modulated signal; For a target located at an angle θ away from directly in front of the radar, with a relative radial velocity v, the constant phase component represented by R0 is removed, and the positions of the transmitting and receiving antennas x are set. m and x n relative displacement and This indicates that, obtained from equations (7) and (12), the total phase difference ΔΦ(x) of the distance cell in which it is located is... m ,x n ,t p ,t m ;θ,v) is represented as From the phase of each target range unit observed at different times as represented by Equation (13), it can be seen that the target's azimuth and radial velocity relative to the radar can be estimated simultaneously based on the model.

3. The high-precision velocity-azimuth joint measurement method for MIMO radar according to claim 1, characterized in that: In step three, 4. The MIMO radar high-precision velocity-azimuth joint measurement method according to claim 1, characterized in that: In steps four and five, 5. The high-precision velocity-azimuth joint measurement method for MIMO radar according to claim 1, characterized in that: In step six, 6. The MIMO radar high-precision velocity-azimuth joint measurement method according to claim 1, characterized in that: In step seven, Where Δr is the total number of observation periods P of the target. total Change in inner radial distance.

7. The high-precision velocity-azimuth joint measurement method for MIMO radar according to claim 1, characterized in that: In step eight, the first Compensation frequency for each target Represented as; Compensated signal for 8. The high-precision velocity-azimuth joint measurement method for MIMO radar according to claim 1, characterized in that: In step nine, the first The signal of each target in each channel Represented as 9. The high-precision velocity-azimuth joint measurement method for MIMO radar according to claim 1, characterized in that: In step ten, the specific method is as follows: For an array observation system consisting of M antenna channels, the observation data is represented as... x(t)=As(t)+n(t) (22) Where x(t) is the array sensor snapshot observation data vector with dimension M×1, s(t) is the target signal vector with dimension K×1, and n(t) is the noise vector with dimension M×1, respectively represented as follows: x(t)=[x1(t),…,x m (t),…,x M (t)] T (23) s(t)=[s1(t),…,s k (t),…,s K (t)] T (24) n(t)=[n1(t),…,n m (t),…,n M (t)] T (25) Where A is the guidance matrix, and for an observation scenario with K targets, the dimension of A is M×K, represented as... A=[a1,…,a k ,…,a K ] (26) Wherein, the guidance vector a of the k-th target k , represented as a k =[a 1,k ,…,a m,k ,…,a M,k ] T (27) When the direction of arrival of the k-th target echo signal is θ k When the first element of the sensor array is taken as the reference element, the steering vector a k Represented as Where λ is the wavelength of the electromagnetic wave, and d is the spacing between the antenna array elements; For the observation scenario represented by equation (22), when the azimuth angle of the kth target is θ k The radial velocity is v k When, the guiding vector a represented by equation (28) k Then equation (13) is updated to At this time, a k It is of dimension N Tx N Rx N Loop A vector of size 1, where N Tx and N Rx N represents the number of transmitting antennas and the number of receiving antennas, respectively. Loop It is the number of consecutive observation cycles, that is, the number of times each transmitting antenna transmits a signal; With R x =E{x(t)x H (t)} represents the covariance matrix of the observed data. The Capon method yields the result as a function of the echo angle. and speed Changing spatial spectrum Represented as Among them, equation (29)a(θ) k ,v k ) in θ k and v k Replace with and That is, get for The conjugate transpose of . For R x The inverse matrix; the angle range is The accurate speed estimation range is Therefore, the first From the perspective of a goal and speed The precise estimation result is 10. The high-precision velocity-azimuth joint measurement method for MIMO radar according to claim 1, characterized in that: In step eleven, the actual radial velocity of the target for

Images