Millimeter wave MIMO radar imaging method based on sparse array

Abstract

The invention provides a millimeter wave MIMO radar imaging method based on a sparse array. The millimeter wave MIMO radar imaging method comprises the following steps: S1, carrying out dimension reduction on collected original five-dimensional echo signal data of a target scene to obtain low-dimensional echo signal data; s2, extracting low-dimensional echo signal data of any receiving channel; s3, processing the extracted low-dimensional echo signal data to obtain a target scene reflectivity image of any receiving channel; and S4, completing MIMO radar imaging according to the target scene reflectivity graph of each receiving channel.

Description

Technical Field

[0001] The present disclosure relates to the field of radar technology, and particularly to a millimeter-wave MIMO radar imaging method based on a sparse array. Background Art

[0002] Due to the ability of wideband millimeter-wave radar imaging to reflect the spatial distribution characteristics of high-energy scatterers of a target, it can well achieve three-dimensional reconstruction of the target. In particular, active millimeter-wave radar imaging has higher spatial resolution and dynamic range compared to passive millimeter-wave radar imaging, and it has been widely applied in various fields, including non-destructive testing, non-cooperative security inspection, medical diagnosis, intelligent vehicle driving, etc.

[0003] However, the current millimeter-wave radar imaging methods face problems such as complexity and poor real-time performance. Summary of the Invention

[0004] In view of this, in order to at least partially solve at least one of the above-mentioned technical problems, the present disclosure provides a millimeter-wave MIMO radar imaging method based on a sparse array.

[0005] To achieve the above object, the technical solution of the present disclosure is as follows:

[0006] According to an embodiment of the present disclosure, there is provided a millimeter-wave MIMO radar imaging method based on a sparse array, including operations S1 - S4.

[0007] Operation S1: Reducing the dimension of the original five-dimensional echo signal data of the collected target scene to obtain low-dimensional echo signal data;

[0008] Operation S2: Extracting the low-dimensional echo signal data of any receiving channel;

[0009] Operation S3: Processing the extracted low-dimensional echo signal data to obtain the reflectivity image of the target scene of any receiving channel; and

[0010] Operation S4: Completing MIMO radar imaging according to the reflectivity maps of the target scene of each receiving channel.

[0011] According to an embodiment of the present disclosure, operation S3 includes the following sub-steps:

[0012] S3.1: Performing azimuth two-dimensional Fourier transform on the echo signal data obtained after step S2;

[0013] S3.2: Performing range migration correction and phase correction at the target center on the echo signal data obtained after sub-step S3.1;

[0014] S3.3: Perform phase correction on the echo signal data obtained after sub-step S3.2 at positions other than the center of the target scene.

[0015] S3.4: Perform range inverse Fourier transform on the echo signal data obtained after sub-step S3.3;

[0016] S3.5: Perform further range migration correction on the echo signal data obtained after sub-step S3.4; and

[0017] S3.6: Perform range inverse Fourier transform on the low-dimensional echo signal data obtained after sub-step S3.5.

[0018] S3.7: Perform phase correction on the echo signal data obtained after sub-step S3.6;

[0019] S3.8: Perform range inverse Fourier transform on the echo signal data obtained after sub-step S3.7;

[0020] S3.9: Perform phase compensation on the echo signal data obtained after sub-step S3.8;

[0021] S3.10: Perform azimuth inverse Fourier transform on the echo signal data obtained after sub-step S3.9; and

[0022] S3.11: Perform spatial domain phase compensation on the low echo signal data obtained after sub-step S3.10.

[0023] According to an embodiment of the present disclosure, the five-dimensional echo signal data includes transmitting antenna position information, receiving antenna position information, and frequency information.

[0024] According to an embodiment of the present disclosure, the five-dimensional echo signal data is reduced to three-dimensional echo signal data by converting the MIMO system into a MISO system.

[0025] According to an embodiment of the present disclosure, the transceiver antenna array of the MIMO radar is distributed in a two-dimensional rectangular shape.

[0026] According to an embodiment of the present disclosure, coherent accumulation summation is performed on the target scene reflectivity map of each receiving channel to complete MIMO radar imaging.

[0027] According to an embodiment of the present disclosure, decoupling of range and bandwidth is achieved through Taylor expansion. Description of the Drawings

[0028] Through the following description of the embodiments of the present disclosure with reference to the drawings, the above and other objects, features, and advantages of the present disclosure will become clearer. In the drawings:

[0029] Figure 1Flowchart of the millimeter-wave MIMO radar imaging method based on a sparse array according to an embodiment of the present disclosure.

[0030] Figure 2 Specific process architecture diagram of the millimeter-wave MIMO radar imaging method based on a sparse array according to an embodiment of the present disclosure.

[0031] Figure 3 Schematic diagram of the MIMO radar array according to an embodiment of the present disclosure.

[0032] Figure 4 Schematic diagram of the layout of the boundary-type MIMO array according to an embodiment of the present disclosure.

[0033] Figure 5 Schematic diagram of the ideal point target imaging experiment scenario according to an embodiment of the present disclosure.

[0034] Figure 6 Schematic diagram of the point target simulation results of the two-dimensional XY projection and the two-dimensional XZ projection of the radar imaging method based on the BP algorithm.

[0035] Figure 7 Schematic diagram of the point target simulation results of the two-dimensional XY projection and the two-dimensional XZ projection of the radar imaging method based on the RMA algorithm.

[0036] Figure 8 Schematic diagram of the point target simulation results of the two-dimensional XY projection and the two-dimensional XZ projection of the radar imaging method based on the RSA algorithm.

[0037] Figure 9 Schematic diagram of the point target simulation results of the two-dimensional XY projection and the two-dimensional XZ projection of the radar imaging method according to an embodiment of the present disclosure. Detailed implementation manners

[0038] The present disclosure provides a millimeter-wave MIMO radar imaging method based on a sparse array. This method first reduces the five-dimensional data to three-dimensional data by converting the MIMO array into multiple MISO arrays, thereby reducing the memory usage. Then, for the echo data of each receiving antenna, the decoupling of distance and bandwidth is achieved through Taylor expansion, and the target reflectivity map is calculated through two fast Fourier transforms in the azimuth direction, three fast Fourier transforms in the range direction, and six complex multiplications. Finally, the target image is obtained by coherently accumulating and summing the reflectivity maps calculated by different receiving channels.

[0039] In the process of implementing the present disclosure, the inventors found that in the prior art, in order to achieve high-resolution three-dimensional imaging of a target, traditional radars arrange antenna transceivers uniformly and densely on a two-dimensional plane, and obtain full-aperture echo data through beam synthesis, that is, the single-input single-output (SISO) transceiver unit method. However, at high frequencies, if a single-station array is arranged according to the traditional half-wavelength spatial sampling requirement, a large number of transceiver units will be required, resulting in high antenna costs. To reduce the hardware cost, a one-dimensional SISO one-dimensional scanning scheme has been proposed, and some research institutions have developed the first scanning one-dimensional SISO array radar and conducted research on related imaging methods.

[0040] To reduce the system cost, based on the principle of equivalent phase center, multi-input multi-output (MIMO) radar design has become increasingly popular in recent years. Inspired by the one-dimensional SISO array mechanical scanning scheme, a one-dimensional MIMO array can be combined with one-dimensional mechanical scanning to achieve high-resolution and large-aperture imaging in the scanning direction through the principle of synthetic aperture radar. In addition, a one-dimensional MIMO array can also be used to scan the target area along a circular or other specific path to achieve high-resolution azimuth imaging through synthetic aperture technology. However, the mechanical scanning in the data acquisition process inevitably increases the time cost of imaging and reduces the real-time imaging ability.

[0041] Considering the time limitation of the mechanical scanning scheme, a fully electronic two-dimensional MIMO array provides a more efficient solution. This array uses fully electronic scanning, significantly improving the data acquisition speed and making real-time imaging possible. Through the principle of equivalent phase center, the two-dimensional MIMO array can synthesize a larger virtual array with fewer actual antenna elements, while taking into account the control of hardware costs and spatial resolution.

[0042] However, two-dimensional MIMO arrays still face challenges in imaging. For example, although the traditional back-projection (BP) algorithm can accurately reconstruct the target scene and has high flexibility, its computational complexity shows an obvious bottleneck in high-resolution imaging applications, and it is difficult to meet the real-time imaging requirements even with GPU acceleration. The range migration algorithm (RMA) uses the fast Fourier transform method to accelerate imaging. However, this algorithm requires interpolation to obtain uniformly gridded sampled data, inevitably increasing the time overhead. The phase shift migration algorithm (PSM) realizes target reconstruction by accumulation in the range direction, but for large-depth-of-field scenes, its time complexity is too high. The compensation method based on the principle of equivalent phase center can convert MIMO imaging into SISO imaging, significantly improving the computational efficiency, but it is also only applicable to narrow field-of-view and small-depth scenes.

[0043] Due to the inherent coupling relationship between bandwidth and range in the millimeter-wave three-dimensional imaging process, researchers have proposed various decoupling schemes. Among them, the method based on Taylor series expansion linearizes the wave number at the reference range and combines the central beam approximation technique to effectively avoid complex interpolation operations. However, this approximation inevitably introduces phase errors. Especially in the region far from the reference range, the accumulated phase deviation may lead to a decrease in range resolution and imaging artifacts. Another important class of fast algorithms is the chirp scaling algorithm (CSA) and the range scaling algorithm (RSA). They precisely correct the wave number domain distribution of the target through a cleverly designed frequency domain transformation sequence and phase scaling operation.

[0044] However, although these algorithms perform well in SAR imaging, they face unique challenges in two-dimensional MIMO millimeter-wave systems: When processing five-dimensional echo data (transmitting antenna position, receiving antenna position, and frequency), a large number of zero-padding operations are generally required, which not only significantly increases the computational complexity but also causes the memory requirement to increase exponentially, making it extremely difficult to achieve real-time processing on conventional hardware platforms.

[0045] Therefore, the present disclosure mainly designs a fast imaging method with low memory occupancy and low computational complexity for a millimeter-wave two-dimensional sparse boundary MIMO array.

[0046] To make the objectives, technical solutions, and advantages of the present disclosure clearer and more understandable, the following further elaborates on the present disclosure in detail with reference to specific embodiments and the accompanying drawings.

[0047] In the embodiments of the present disclosure, in combination with Figure 1 and Figure 2 , and Figure 3 as shown, a millimeter-wave MIMO radar imaging method based on a sparse array is provided, including:

[0048] Operation S1: Reducing the dimension of the original five-dimensional echo signal data of the target scene collected to obtain low-dimensional echo signal data;

[0049] Operation S2: Extracting the low-dimensional echo signal data of any receiving channel;

[0050] Operation S3: Processing the extracted low-dimensional echo signal data to obtain the reflectivity image of the target scene of any receiving channel; and

[0051] Operation S4: Completing MIMO radar imaging based on the reflectivity maps of the target scene of each receiving channel.

[0052] According to the embodiments of the present disclosure, in combination with Figure 2 and Figure 1 , Figure 3 as shown, operation S3 includes the following sub-steps:

[0053] S3.1: Perform two-dimensional Fourier transform in the azimuth direction on the echo signal data obtained after step S2;

[0054] S3.2: Perform range migration correction and phase correction at the target center on the echo signal data obtained after sub-step S3.1;

[0055] S3.3: Perform phase correction at positions other than the target scene center on the echo signal data obtained after sub-step S3.2.

[0056] S3.4: Perform inverse Fourier transform in the range direction on the echo signal data obtained after sub-step S3.3;

[0057] S3.5: Perform further range migration correction on the echo signal data obtained after sub-step S3.4;

[0058] S3.6: Perform inverse Fourier transform in the range direction on the low-dimensional echo signal data obtained after sub-step S3.5.

[0059] S3.7: Perform phase correction on the echo signal data obtained after sub-step S3.6;

[0060] S3.8: Perform inverse Fourier transform in the range direction on the echo signal data obtained after sub-step S3.7;

[0061] S3.9: Perform phase compensation on the echo signal data obtained after sub-step S3.8;

[0062] S3.10: Perform inverse Fourier transform in the azimuth direction on the echo signal data obtained after sub-step S3.9; and

[0063] S3.11: Perform spatial domain phase compensation on the low echo signal data obtained after sub-step S3.10.

[0064] According to an embodiment of the present disclosure, the five-dimensional echo signal data includes transmitting antenna position information, receiving antenna position information, and frequency information.

[0065] According to an embodiment of the present disclosure, the five-dimensional echo signal data is reduced to three-dimensional echo signal data by converting the MIMO system into a MISO system.

[0066] According to an embodiment of the present disclosure, the transceiver antenna array of the MIMO radar is distributed in a two-dimensional rectangular shape.

[0067] According to an embodiment of the present disclosure, coherent accumulation summation is performed on the target scene reflectivity map of each receiving channel to obtain the final target reflectivity map, completing MIMO radar imaging.

[0068] According to an embodiment of the present disclosure, decoupling of distance and bandwidth is achieved through Taylor expansion.

[0069] According to an embodiment of the present disclosure, the interval between the transmitting antenna and the receiving antenna is between 2 cm and 5 cm.

[0070] Specifically, for a 2D MIMO boundary array model as Figure 3 shown, the transceiver units are rectangularly distributed in space. Assuming there are a total of transmitting antennas and receiving antennas, where the transmitting antennas are uniformly distributed above and below the array and parallel to the ground. The receiving antennas are uniformly distributed on the left and right sides of the array and perpendicular to the ground. The position coordinates of the transmitting antennas and the receiving antennas in the (spatial three-dimensional) coordinate system (with the center of the array as the origin) can be expressed as and , the transmitting antenna sequentially transmits FMCW signals for the target scene. For each transmission, the receiving antennas respectively receive echo signals from serial number 1 to , where is the number of receiving antennas.

[0071] Assuming that the scattering process satisfies the Born approximation, the frequency-domain echo signal received by the scattered receiving antenna can be expressed as:

[0072] ;

[0073] where is the reflectivity map of the target scene of interest, which is the superposition of all targets in the region, where x′y′z′ respectively represent the position coordinate values of the target scene. represents the spatial beam corresponding to different frequencies, f represents the frequency, and c represents the propagation speed of the echo signal. R T represents the distance from the transmitting antenna to the target, and R R represents the distance from the receiving antenna to the target, which are respectively expressed as:

[0074] ;

[0075] ;

[0076] To solve the problem of large memory and time overhead caused by the original 5D echo signal , here the MIMO imaging problem is transformed into imaging problems of multiple single-transmit multi-receive (MISO) systems, and then the target reflectivity maps under each MISO are solved separately. Finally, the reflectivity map of MIMO is obtained through coherent accumulation. Among them, x t , y tDenote the horizontal and vertical position coordinates corresponding to the transmitting antenna, \(x\) r and \(y\) r which denote the horizontal and vertical position coordinates corresponding to the receiving antenna. Then, for any receiving antenna at a certain position , the low-dimensional echo signals received corresponding to different transmitting antennas can be expressed as:

[0077] ;

[0078] where: \(i\) represents the ordinal number, for example, the \(i\)-th receiving antenna, and \(R\) Ri denotes the distance of the \(i\)-th receiving antenna relative to the target. Then:

[0079] ;

[0080] \(x\) ri and \(y\) ri respectively denote the horizontal coordinate and vertical coordinate corresponding to the receiving antenna with ordinal number \(i\).

[0081] Process the extracted echo signal data to obtain the target scene reflectivity image of any receiving channel; specifically, it includes the following sub-steps:

[0082] S3.1: Perform two-dimensional Fourier transform in the azimuth direction on the echo signal data obtained after step S2;

[0083] Perform Fourier transform on both the left and right sides of the formula with respect to . The wavenumber domain data can be obtained:

[0084] ;

[0085] \(k\) xt and \(k\) yt respectively denote the wavenumbers in the wavenumber domain corresponding to the result of performing Fourier transform on .

[0086] Using the principle of stationary phase (POSP), let , and let . Then there is a stationary phase point:

[0087] ;

[0088] denotes the phase term related to when performing Fourier transform in the dimension.

[0089] Substitute the stationary phase point back into Then there is:

[0090] ;

[0091] It represents the amplitude change term of the echo signal after Fourier transform in the dimension.

[0092] Using the PSP again, let , let , then there are stationary phase points:

[0093] ;

[0094] It represents the phase term related to when performing Fourier transform in the dimension.

[0095] Substitute the stationary phase points back into Then there is:

[0096] ;

[0097] Among them:

[0098] ;

[0099] It represents the amplitude change term of the echo signal after Fourier transform in the dimension.

[0100] Next, perform a third-order Taylor expansion on , and we can get:

[0101] ;

[0102] So there is:

[0103] ;

[0104] Among them:

[0105] ;

[0106] ;

[0107] It represents when performing R in ri R corresponding to the third-order Taylor expansion ri Higher-order terms

[0108] For For the last term in and To decouple, let where is the beam corresponding to the center frequency, k z represents the wavenumber corresponding to the range direction, k b represents the wavenumber offset relative to the wavenumber at the center frequency. For Perform polynomial expansion, applicable to the case of low squint and relatively narrow relative bandwidth:

[0109] ;

[0110] Among them:

[0111] ;

[0112] ;

[0113] D represents the coefficient related to the azimuth wavenumber, and g0 represents the coefficient related to the wavenumber offset in the range wavenumber.

[0114] S3.2: Perform consistent range migration correction and phase correction at the target center on the echo signal data obtained after sub-step S3.1.

[0115] To achieve consistent range migration correction, define the target scene center reference function as:

[0116] ;

[0117] Multiply the target echo signal by the target scene center reference function:

[0118] ;

[0119] Among them is the center position of the target plane. Since the target scene is not large, and it is in the millimeter wave band, and there is: , so for terms perform approximation.

[0120] The above process performs amplitude attenuation and phase correction at the target center.

[0121] S3.3: Perform phase correction for positions other than the center of the target scene on the echo signal data obtained after sub-step S3.2.

[0122] To further correct the phase at other positions, the following frequency scaling correction is performed on the above formula. First, introduce the residual error phase function ;

[0123] ;

[0124] ;

[0125] where represents the chirp rate, where is the bandwidth, is the frequency sampling interval, and α represents the chirp rate in the residual error. Multiply both ends of by the residual video error phase function respectively:

[0126] ;

[0127] S3.4: Perform range inverse Fourier transform on the echo signal data obtained after sub-step S3.3.

[0128] Perform range inverse Fourier transform on both ends of :

[0129] ;

[0130] where C represents the phase integration term related to the range wavenumber during the range inverse Fourier transform process:

[0131] ;

[0132] z″ represents the Fourier transform pair corresponding to the offset wavenumber k b .

[0133] Through POSP, the following definitions are made:

[0134] ;

[0135] represents the phase term related to the offset wavenumber k b , and α represents the chirp rate in the residual error.

[0136] The stationary phase point is calculated as:

[0137] ;

[0138] Ignoring the constant term, we can obtain:

[0139] ;

[0140] ;

[0141] It represents the phase integration term related to the spatial distance during the range Fourier transform process.

[0142] Substitute back into We can get:

[0143] ;

[0144] S3.5: Perform range migration correction on the echo signal data obtained after sub-step S3.4.

[0145] Next, perform uniform range migration correction by defining a frequency scaling function. Define:

[0146] ;

[0147] It represents the phase compensation term related to the frequency scaling function, It represents the frequency scaling factor.

[0148] Multiply both the left and right sides of by the frequency scaling function:

[0149] ;

[0150] S3.6: Perform inverse range Fourier transform on the echo signal data obtained after sub-step S3.5.

[0151] Performing Fourier transform on both the left and right ends in the range direction can obtain:

[0152] ;

[0153] ;

[0154] Among them, C1 represents the phase integration term related to the range wavenumber during the range Fourier transform process.

[0155] Through POSP, the following definitions are made:

[0156] ;

[0157] The stationary phase points are calculated as:

[0158] ;

[0159] ;

[0160] Bring back to to obtain:

[0161] ;

[0162] S3.7: Perform phase correction on the echo signal data obtained after sub-step S3.6.

[0163] Next, define the phase correction function :

[0164] ;

[0165] Multiply the complex number by the phase correction function, and substitute to obtain: Substitute to get:

[0166] ;

[0167] S3.8: Perform range inverse Fourier transform on the echo signal data obtained after sub-step S3.7.

[0168] Perform range inverse Fourier transform on both the left and right sides:

[0169] ;

[0170] S3.9: Perform phase compensation on the echo signal data obtained after sub-step S3.8.

[0171] Next, multiply both ends of by the phase compensation function respectively:

[0172] ;

[0173] ;

[0174] S3.10: Perform azimuth inverse Fourier transform on the echo signal data obtained after sub-step S3.9.

[0175] For Performing azimuth inverse Fourier transform on both the left and right ends respectively, we can obtain:

[0176] ;

[0177] where R Ri represents the distance from the i-th receiving antenna to the target.

[0178] S3.11: Perform spatial domain phase compensation on the echo signal data obtained after sub-step S3.10.

[0179] After that, perform phase compensation on both the left and right ends respectively:

[0180] ;

[0181] ;

[0182] where represents the amplitude compensation function.

[0183] Operation S4: Complete MIMO radar imaging based on the target scene reflectivity map of each receiving channel.

[0184] Specifically, perform coherent accumulation on the target scene reflectivity map of each receiving channel to complete MIMO radar imaging. Repeat the above sub-steps to obtain the reflectivity maps for all receiving antennas respectively and then perform coherent summation to obtain the final target reflectivity map.

[0185] ;

[0186] where is the number of receiving antennas.

[0187] Computational complexity analysis:

[0188] Assume that the size of the target scene of interest is , is the number of frequency points. are the numbers of the horizontal and vertical coordinates of the transmitting antennas respectively, represents the number of receiving antennas, represents the number of the horizontal coordinates of the receiving antennas, represents the number of the vertical coordinates of the receiving antennas, Indicates the number of wavenumber points.

[0189] Operations S1 and S2 involve data dimensionality reduction and extraction, without complex calculations. They only involve reading the stored data and padding with zeros, so the complexity calculation is not performed here.

[0190] Sub - steps in operation S3:

[0191] In S3.1, a two - dimensional Fourier transform is involved in the azimuth direction. The computational complexity here is:

[0192] ;

[0193] In S3.2, implementing range migration correction and propagation loss compensation is involved. Here, complex multiplication needs to be performed on the wavenumber - domain data after the two - dimensional Fourier transform in the azimuth direction. The computational complexity here is:

[0194] ;

[0195] In S3.3, introducing the residual video error phase and multiplying the phase function with the wavenumber - domain data after S3.2 is involved. The computational complexity here is:

[0196] ;

[0197] In S3.4, performing an inverse Fourier transform in the range direction is involved. The computational complexity here is:

[0198]

[0199] In S3.5, introducing a frequency scaling function and multiplying the function with the result after S3.4 is involved. The computational complexity here is:

[0200] ;

[0201] In S3.6, performing a Fourier transform in the range direction is involved. The computational complexity here is:

[0202] ;

[0203] In S3.7, introducing a phase correction function and multiplying the phase correction function with the data after S3.6 is involved. The computational complexity here is:

[0204] ;

[0205] In S3.8, an inverse Fourier transform in the range direction is involved, and the computational complexity here is:

[0206] ;

[0207] In S3.9, introducing a phase compensation function and performing complex multiplication on the data after the inverse Fourier transform in S3.8 is involved, and the computational complexity here is:

[0208] ;

[0209] In S3.10, a two-dimensional inverse Fourier transform in the azimuth direction is involved. To improve the imaging accuracy, zero-padding is performed on the transceiver array and then the inverse Fourier transform is carried out, and the computational complexity here is:

[0210] ;

[0211] In S3.11, compensation in the spatial domain is involved, and the computational complexity here is:

[0212] ;

[0213] The computational complexity of the above sub-steps is:

[0214] ;

[0215] To obtain the reflectivity map corresponding to each receiving channel, the above sub-steps need to be repeated, and finally, coherent accumulation summation is performed on the reflectivity map of the target scene for each receiving channel. The time complexity of the coherent summation step is:

[0216] ;

[0217] So the total time complexity is:

[0218] ;

[0219] Table 1: Comparison of algorithm complexities of different algorithms

[0220]

[0221] For the convenience of analysis, assume , , and the computational complexities of different algorithms can be expressed as:

[0222] Computational complexity of back-projection algorithm: ;

[0223] Computational complexity of range migration algorithm: ;

[0224] Computational complexity of range scaling algorithm: ;

[0225] Algorithm complexity involved in the method of the present disclosure: ;

[0226] It can be seen that the computational complexity of the method proposed in the present disclosure C has one less dimension of computational complexity compared to other advanced algorithms.

[0227] Example 1

[0228] In the example of the present disclosure, by setting specific simulation parameters, the feasibility of the algorithm is simulated and analyzed.

[0229] (1) Set the simulation parameters as shown in Table 2:

[0230] Table 2 Simulation parameter settings

[0231]

[0232] (2) The boundary-type MIMO array is set as Figure 4 shown:

[0233] For the point target scene reconstruction, assume there are 32 transmit channels (Tx) and 32 receive channels (Rx), a total of 1024 channel data. Assume the frequency sampling points are 256, then set the number of received echo signal points to 1024×256, and the size of the imaging scene is 0.6m×0.6m.

[0234] Place 9 point targets with a reflectivity of 1 in the space at a relative position of 3.5 - 4.5m from the array. The three-dimensional scene diagram of the specific distribution positions is as Figure 5 shown. Then, image the target scene through the imaging methods based on the BP algorithm, RMA algorithm, RSA algorithm and the imaging method proposed in the present disclosure respectively. The imaging results are respectively as Figures 6 to 9 shown.

[0235] The imaging results show that the method proposed in the present disclosure can well achieve the reconstruction of the target scene.

[0236] The imaging times corresponding to the algorithms involved in different methods are shown in Table 3. The results show that the proposed algorithm has a low computational time complexity.

[0237] Table 3 Time consumption of different imaging algorithms

[0238]

[0239] The radar imaging method proposed in this disclosure mainly designs a fast imaging method with low memory occupancy and low computational complexity for a millimeter-wave two-dimensional sparse boundary MIMO array. This method first reduces the five-dimensional data to three-dimensional data by converting the MIMO array into multiple MISO arrays, reducing the memory usage. Then, for the echo data of each receiving antenna, the decoupling of range and bandwidth is achieved through Taylor expansion, and the target reflectivity map is calculated through two azimuth fast Fourier transforms, three range fast Fourier transforms, and six complex multiplications. Finally, the target image is obtained by coherently accumulating and summing the reflectivity maps calculated by different receiving channels. Compared with advanced algorithms, the computational complexity of this method is one dimension less, with faster computational speed and imaging speed. The computational memory occupancy is low, and interpolation is not required in the computational process compared with other algorithms.

[0240] So far, the embodiments of the present disclosure have been described in detail with reference to the accompanying drawings. It should be noted that the implementation manners not illustrated or described in the accompanying drawings or the text of the specification are all forms known to those of ordinary skill in the art and have not been described in detail. In addition, the definitions of the above elements and methods are not limited to the specific structures, shapes, or manners mentioned in the embodiments, and those of ordinary skill in the art can make simple changes or substitutions to them.

[0241] In addition, unless specifically described or steps that must occur in sequence, the order of the above steps is not limited to those listed above and can be changed or rearranged according to the required design. And the above embodiments can be used in combination with each other or combined with other embodiments based on considerations of design and reliability, that is, the technical features in different embodiments can be freely combined to form more embodiments.

[0242] The specific embodiments described above further illustrate the purpose, technical solutions, and beneficial effects of the present disclosure. It should be understood that the above are only specific embodiments of the present disclosure and are not used to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present disclosure should be included within the protection scope of the present disclosure.

Claims

1. A millimeter wave MIMO radar imaging method based on a sparse array, comprising: Operation S1: reducing the dimension of the collected original five-dimensional echo signal data of the target scene to obtain low-dimensional echo signal data; Operation S2: extracting low-dimensional echo signal data of any receiving channel; Operation S3: Processing the extracted low-dimensional echo signal data to obtain a target scene reflectivity image of any receiving channel; as well as Operation S4: completing MIMO radar imaging according to the target scene reflectivity map of each receiving channel.

2. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 1, operation S3 comprises the following sub-steps: S3.1: performing azimuth two-dimensional Fourier transform on the echo signal data obtained after step S2; S3.2: performing range migration correction and phase correction at the target center on the echo signal data obtained after sub-step S3.1; S3.3: Perform phase correction on the echo signal data obtained after sub-step S3.2 at positions other than the center of the target scene.

3. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 1, operation S3 further comprises: S3.4: performing range-direction inverse Fourier transform on the echo signal data obtained after sub-step S3.3; S3.5: further performing range migration correction on the echo signal data obtained after sub-step S3.4; and S3.6: Perform range-direction inverse Fourier transform on the low-dimensional echo signal data obtained after sub-step S3.

5.

4. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 3, operation S3 further comprises: S3.7: performing phase correction on the echo signal data obtained after sub-step S3.6; S3.8: Performing range-direction inverse Fourier transform on the echo signal data obtained after sub-step S3.7; S3.9: Perform phase compensation on the echo signal data obtained after sub-step S3.8; S3.10: Performing an azimuth inverse Fourier transform on the echo signal data obtained after sub-step S3.9; and S3.11: Perform spatial domain phase compensation on the low echo signal data obtained after sub-step S3.

10.

5. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 1, the five-dimensional echo signal data includes transmitting antenna position information, receiving antenna position information, and frequency information.

6. The millimeter wave MIMO radar imaging method based on sparse array according to claim 1, wherein the five-dimensional echo signal data is reduced to three-dimensional echo signal data by converting the MIMO system into a MISO system.

7. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 1, the transceiver antenna array of the MIMO radar is distributed in a two-dimensional rectangular shape.

8. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 1, coherent accumulation and summation are performed on the target scene reflectivity map of each receiving channel to complete MIMO radar imaging. 9 . The millimeter-wave MIMO radar imaging method based on sparse array according to claim 1 , wherein the decoupling of distance and bandwidth is achieved through Taylor expansion.

10. According to the sparse array-based millimeter-wave MIMO radar imaging method of claim 7, the interval between the transmitting antenna and the receiving antenna is between 2 cm and 5 cm.

Images