Unmanned aerial vehicle target direction finding method and device based on multi-band sparse array

Abstract

The application provides a UAV target direction finding method and device based on a multi-band sparse array, comprising: collecting signals of a target through a non-uniform sparse antenna array arranged on the side of a UAV, and sampling the signals to obtain multi-band radio frequency data; obtaining the attitude angle of the UAV at each sampling time, and correcting the multi-band radio frequency data based on the attitude angle; constructing a virtual differential common array based on the corrected multi-band radio frequency data, and performing virtual aperture fusion on the virtual differential common array to obtain a non-hole virtual differential common array; constructing a multi-band joint observation model based on the non-hole virtual differential common array, obtaining a virtual observation vector, and performing weighted sparse reconstruction on the virtual observation vector to obtain the direction finding result of the target. The application improves the problem of limited space on the side of the UAV, and improves the target resolution accuracy.

Description

Technical Field

[0001] This invention relates to the field of UAV target orientation technology, and in particular to a UAV target orientation method and apparatus based on a multi-band sparse array. Background Technology

[0002] In recent years, with the rapid development of modern electronic information technology, the electromagnetic spectrum has become the fifth strategic high ground after land, sea, air, and space. Against this backdrop, unmanned aerial vehicle (UAV) systems, with their advantages of zero casualties, high maneuverability, low radar cross-section, and flexible deployment, have demonstrated irreplaceable tactical value in tasks such as electronic reconnaissance, passive positioning, communication relay, and jamming suppression. Currently, mainstream passive direction finding systems mainly include amplitude comparison direction finding, interferometer direction finding, and spatial spectrum estimation techniques based on array signal processing, which have broad application prospects in radar countermeasures and radio monitoring. However, when adapting these techniques to miniaturized UAV platforms, they still face severe physical limitations and technical bottlenecks. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a UAV target direction finding method and device based on a multi-band sparse array, so as to improve the problem of limited lateral space of UAV and improve the target resolution accuracy.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a UAV target direction finding method based on a multi-band sparse array, comprising: acquiring target signals by deploying a non-uniform sparse antenna array on the side of the UAV, and sampling the signals to obtain multi-band radio frequency data; acquiring the attitude angle of the UAV at each sampling time, and correcting the multi-band radio frequency data based on the attitude angle; constructing a virtual differential co-array based on the corrected multi-band radio frequency data, and performing virtual aperture fusion on the virtual differential co-array to obtain a hole-free virtual differential co-array; constructing a multi-band joint observation model based on the hole-free virtual differential co-array to obtain a virtual observation vector, and performing weighted sparse reconstruction on the virtual observation vector to obtain the target direction finding result.

[0005] Optionally, the attitude angles of the UAV at each sampling time are acquired, and the multi-band radio frequency data is corrected based on the attitude angles. This includes: acquiring the attitude angles of the UAV relative to the geodetic coordinate system through an onboard inertial measurement unit; determining the attitude rotation matrix from the geodetic coordinate system to the body coordinate system based on the Euler angle rotation theorem and the attitude angles, and determining the direction cosine matrix based on the attitude rotation matrix; calculating the attitude deviation projection of the UAV based on the direction cosine matrix; performing phase inversion rotation on the multi-band radio frequency data, and calculating the diagonal elements of the compensation matrix based on the phase inversion rotated multi-band radio frequency data and the attitude deviation projection to obtain the compensation matrix; and correcting the multi-band radio frequency data based on the compensation matrix.

[0006] Optionally, a virtual differential co-array is constructed based on the corrected multi-band radio frequency data, including: calculating the theoretical covariance matrix of the corrected multi-band radio frequency data; performing column vectorization on the theoretical covariance matrix to obtain the virtual differential co-array, and determining the position set of each virtual array element in the virtual differential co-array.

[0007] Optionally, virtual aperture fusion is performed on the virtual differential co-array to obtain a hole-free virtual differential co-array, including: determining a reference frequency based on the corrected multi-band radio frequency data, mapping the data of other frequencies in the corrected multi-band radio frequency data other than the reference frequency to the virtual space of the reference frequency to obtain a hole-free virtual differential co-array, and determining the location set of the hole-free virtual differential co-array.

[0008] Optionally, a multi-band joint observation model is constructed based on a hole-free virtual differential array to obtain virtual observation vectors. The virtual observation vectors are then weighted and sparsely reconstructed to obtain the target's direction finding results. This includes: constructing a multi-frequency joint virtual steering matrix based on the position set of the hole-free virtual differential array; constructing a multi-band joint observation model based on the multi-frequency joint virtual steering matrix to obtain virtual observation vectors; discretizing a preset spatial angle range into multiple grid points, and constructing an overcomplete dictionary matrix based on the spatial angles corresponding to the grid points and the multi-frequency joint virtual steering matrix; constructing a sparse representation model of the virtual observation vectors based on the overcomplete dictionary matrix; and solving the sparse representation model to obtain the target's direction finding results.

[0009] Optionally, the sparse representation model is solved to obtain the target's direction finding results, including: correcting the sparse representation model based on the deviation between the target and grid points; and using weighted averages. The norm minimization algorithm is used to solve the modified sparse representation model to obtain the direction finding results of the target.

[0010] Optionally, the non-uniform sparse antenna array adopts a coprime linear array structure, consisting of two sparse uniform subarrays; the first subarray includes a first number of array elements, and the second subarray includes a second number of array elements; wherein the first number and the second number are coprime numbers, and the first subarray and the second subarray share the first array element.

[0011] Secondly, the present invention provides a UAV target direction finding device based on a multi-band sparse array, comprising: a data acquisition module for acquiring target signals through a non-uniform sparse antenna array deployed on the side of the UAV, and sampling the signals to obtain multi-band radio frequency data; a correction module for acquiring the attitude angle of the UAV at each sampling time, and correcting the multi-band radio frequency data based on the attitude angle; a virtual array construction and filling module for constructing a virtual differential co-array based on the corrected multi-band radio frequency data, and performing virtual aperture fusion on the virtual differential co-array to obtain a hole-free virtual differential co-array; and a direction finding result generation module for constructing a multi-band joint observation model based on the hole-free virtual differential co-array, obtaining a virtual observation vector, and performing weighted sparse reconstruction on the virtual observation vector to obtain the target direction finding result.

[0012] Thirdly, the present invention provides an electronic device including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the steps of the method provided in any of the first aspects above.

[0013] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, performs the steps of the method provided in any of the first aspects above.

[0014] This invention brings the following beneficial effects: The present invention provides a UAV target direction finding method and apparatus based on a multi-band sparse array. First, a non-uniform sparse antenna array deployed on the side of the UAV is used to collect the target signal and sample the signal to obtain multi-band radio frequency data. Then, the attitude angle of the UAV at each sampling time is obtained, and the multi-band radio frequency data is corrected based on the attitude angle. Next, a virtual differential co-array is constructed based on the corrected multi-band radio frequency data, and virtual aperture fusion is performed on the virtual differential co-array to obtain a hole-free virtual differential co-array. Finally, a multi-band joint observation model is constructed based on the hole-free virtual differential co-array to obtain a virtual observation vector, and the virtual observation vector is weighted and sparsely reconstructed to obtain the target direction finding result. In the above method, deploying a non-uniform sparse antenna array on the side of the UAV can maintain the large aperture observation capability while significantly reducing the number and weight of physical antennas, thus improving the problem of limited lateral space for the UAV. By constructing a virtual differential array using multi-band radio frequency data and performing virtual aperture fusion, the direction finding results of the target are obtained by using a weighted sparse reconstruction method. This effectively improves the ambiguity problem of sparse arrays under single frequency, ensures the uniqueness and robustness of the direction finding results in complex electromagnetic environments, and improves the target resolution accuracy.

[0015] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 A flowchart of a UAV target direction finding method based on a multi-band sparse array provided in an embodiment of the present invention; Figure 2 A schematic diagram of a dual-sided distributed sparse array system device provided in an embodiment of the present invention; Figure 3 A schematic diagram of a coprime array physical structure and a multi-frequency virtual aperture expansion principle provided in an embodiment of the present invention; Figure 4A flowchart illustrating the implementation of a multi-band array direction finding method is provided in this embodiment of the invention. Figure 5 A comparison diagram of single-band grating lobe blurring and the multi-frequency suppression effect of the present invention is provided for an embodiment of the present invention; Figure 6 A convergence curve of direction finding error before and after attitude compensation for a UAV is provided as an embodiment of the present invention; Figure 7 A super-resolution direction finding effect diagram based on sparse reconstruction is provided for an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of a UAV target direction finding device based on a multi-band sparse array, provided in an embodiment of the present invention. Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Currently, adapting mainstream passive direction finding systems to miniaturized UAV platforms still faces severe physical limitations and technical bottlenecks. First, UAV platforms impose extremely strict constraints on payload size, weight, and power consumption (SWaP), resulting in very limited physical space available for antenna placement on the fuselage or wing sides. Constrained by the Rayleigh limit, the direction finding resolution of small-aperture uniform linear arrays is severely insufficient, making it difficult to distinguish radiation sources with similar spatial angles. Second, although sparse arrays can expand the physical aperture without increasing the number of array elements, they are prone to generating high-level grating lobes when processing single-band narrowband signals, leading to severe direction finding ambiguity and spurious peaks. Furthermore, UAVs are inevitably subject to yaw, pitch, and roll motions during flight due to airflow disturbances. These dynamic changes in platform attitude introduce non-stationary phase errors, and without an effective compensation mechanism, array manifold mismatch will occur, drastically degrading direction finding performance.

[0021] Based on this, the present invention provides a UAV target direction finding method and device based on a multi-band sparse array, which can improve the problem of limited lateral space of UAVs and improve target resolution accuracy.

[0022] To facilitate understanding of this embodiment, a detailed description of the UAV target direction finding method based on a multi-band sparse array disclosed in this invention will be provided first. This method can be executed by the UAV, or the UAV can return the collected data to ground-based electronic devices, such as computers, smartphones, and tablets, for execution. See also... Figure 1 The flowchart shown illustrates a UAV target direction finding method based on a multi-band sparse array, indicating that the method mainly includes the following steps S101 to S104: Step S101: Acquire the target signal by deploying a non-uniform sparse antenna array on the side of the UAV, and sample the signal to obtain multi-band radio frequency data.

[0023] In one implementation, see Figure 2 The diagram illustrates a dual-sided distributed sparse array system. A non-uniform sparse antenna array is deployed laterally on a UAV, covering a large physical aperture with a number of elements far less than the Rayleigh limit requirement. To obtain the maximum virtual aperture within the limited lateral space of the UAV, the non-uniform sparse antenna array adopts a coprime linear array structure, consisting of two sparse uniform subarrays. The first subarray includes a first number of elements, and the second subarray includes a second number of elements; wherein the first and second numbers are coprime, and the first and second subarrays share the first element.

[0024] Consider a sparse array consisting of M omnidirectional antenna elements, whose physical positions no longer satisfy a uniform distribution. In this embodiment of the invention, the non-uniform sparse antenna array is formed by interleaving two sparse uniform subarrays: First subarray (subarray A): contains There are array elements, and the spacing between the array elements is... ; Second subarray (subarray B): contains There are array elements, and the spacing between the array elements is... .

[0025] in, One and The numbers are coprime (i.e., their common divisor is only 1), and . The reference half-wavelength unit is typically taken as half the wavelength corresponding to the system's highest operating frequency. The two subarrays share the first element (located at zero). Therefore, the total number of elements in the physical array is... .

[0026] The set of physical positions of array elements in the body coordinate system It can be represented as:

[0027] These positions are fixed physical coordinates, all located on a one-dimensional baseline to the side of the UAV.

[0028] Furthermore, after detecting the target's signal using a non-uniform sparse antenna array deployed laterally on the UAV, multi-band radio frequency data can be simultaneously acquired by utilizing the broadband characteristics of the target signal or its frequency agility at multiple frequency points. Specifically, it is assumed that there exists in the far field... A non-coherent narrowband signal source (or multiple frequency components of a wideband signal decomposed by a narrowband filter bank), respectively at angles Incident on a non-uniform sparse antenna array. The UAV receiver is... Discrete frequency points Synchronous sampling is performed on the upper part.

[0029] For the Frequency Its corresponding wavelength is ( (The speed of light). At time... The snapshot data vector received by the physical array It can be represented as:

[0030] in, It is the first Under each frequency point The complex envelope vector of each information source.

[0031] It is an additive white Gaussian noise vector, assuming its mean is zero and its variance is... Furthermore, they are unrelated in both space and time.

[0032] It is the first Array manifold matrix at each frequency point.

[0033] Guide vector It is frequency-dependent: for positions where The Each array element, relative to the reference array element ( The phase delay depends on the frequency. and angle of incidence Therefore, the steering vector is expressed as:

[0034] Among them, phase factor This indicates that changing the frequency This property, which is equivalent to changing the electrical length position of the array element, is the physical basis for subsequent multi-band virtual aperture synthesis.

[0035] Step S102: Obtain the attitude angle of the UAV at each sampling time, and correct the multi-band radio frequency data based on the attitude angle.

[0036] In one implementation, the attitude angles (roll, pitch, and yaw) of the UAV at each sampling moment are obtained through an attitude sensing module (IMU / INS), and the steering vector of the received signal is phase corrected in real time to eliminate the non-stationary error introduced by the moving base.

[0037] Step S103: Construct a virtual differential co-array based on the corrected multi-band RF data, and perform virtual aperture fusion on the virtual differential co-array to obtain a hole-free virtual differential co-array.

[0038] In one implementation, the corrected multi-band data is processed by second-order statistics (covariance matrix vectorization) to construct a virtual difference co-array. By utilizing the frequency diversity characteristics, the sparse co-arrays at different frequencies are mapped to a unified reference wavelength space, and the holes caused by the sparse array are filled to each other, thus synthesizing a virtual uniform linear array with a huge aperture and no grating lobes, i.e., a hole-free virtual difference co-array.

[0039] Step S104: Construct a multi-band joint observation model based on a hole-free virtual differential co-array to obtain virtual observation vectors, and perform weighted sparse reconstruction on the virtual observation vectors to obtain the target direction finding results.

[0040] In one implementation, a multi-band joint observation model is constructed using a hole-free virtual differential co-array to obtain virtual observation vectors, and then applied based on... The norm-optimized sparse reconstruction algorithm performs weighted sparse reconstruction on the virtual observation vector to obtain the target's direction finding results.

[0041] The UAV target direction finding method based on multi-band sparse array provided in this invention deploys a non-uniform sparse antenna array on the side of the UAV, which can maintain large-aperture observation capability while significantly reducing the number and weight of physical antennas, thus improving the problem of limited lateral space for the UAV. By constructing a virtual differential array using multi-band radio frequency data and performing virtual aperture fusion, the direction finding results of the target are obtained by using a weighted sparse reconstruction method, which effectively improves the ambiguity problem of sparse arrays under single frequency, ensures the uniqueness and robustness of the direction finding results in complex electromagnetic environments, and improves the target resolution accuracy.

[0042] In one implementation, in practical applications, unmanned aerial vehicles (UAVs) are not ideally stationary platforms. Airflow disturbances, engine vibrations, and predetermined flight maneuvers can cause yaw, pitch, and roll movements of the aircraft. These attitude changes cause instantaneous changes in the actual angle of the antenna array relative to the target, affecting the preset array manifold matrix. Mismatch. Without compensation, virtual aperture synthesis will fail due to phase incoherence. Based on this, this embodiment of the invention establishes a real-time attitude compensation model based on Euler angles.

[0043] Specifically, for the aforementioned step S102, that is, when acquiring the attitude angle of the UAV at each sampling moment and correcting the multi-band radio frequency data based on the attitude angle, the following methods can be used, including but not limited to: (1) Obtain the attitude angle of the UAV relative to the geodetic coordinate system through the airborne inertial measurement unit.

[0044] (2) Based on Euler's angle rotation theorem and attitude angle, determine the attitude rotation matrix from the geodetic coordinate system to the body coordinate system, and determine the direction cosine matrix based on the attitude rotation matrix.

[0045] In practice, the geodetic coordinate system (n-frame) adopts the North-East-Earth (NED) coordinate system as the reference system for the target's orientation. The body coordinate system (b-frame) is fixed to the UAV body, with its origin at the center of gravity, the X-axis pointing towards the nose, the Y-axis towards the right wing, and the Z-axis pointing vertically downwards. A sparse array is mounted on the body, and its element coordinates... It is constant within a b-frame.

[0046] In this embodiment of the invention, the attitude angles of the UAV relative to the geodetic coordinate system, namely the roll angle, are obtained in real time using an airborne inertial measurement unit. Rotation around the X-axis; pitch angle Rotation about the Y-axis; yaw angle : Rotate around the Z-axis.

[0047] According to Euler's angle rotation theorem (in ZYX order, i.e., the 3-2-1 rotation sequence), the attitude rotation matrix from the geodetic coordinate system to the body coordinate system is... Defined as:

[0048] Expanding it yields the Direction Cosine Matrix (DCM): .

[0049] (3) Calculate the attitude deviation projection of the UAV based on the direction cosine matrix.

[0050] (4) Perform phase inversion rotation on the multi-band radio frequency data, and calculate the diagonal elements of the compensation matrix based on the phase inversion rotation of the multi-band radio frequency data and the attitude deviation projection to obtain the compensation perception matrix; (5) Correct the multi-band radio frequency data based on the compensation matrix.

[0051] In practical implementation, it is assumed that the direction of arrival vector of the target in the geodetic coordinate system is... (in, It is the azimuth angle. (Pitch angle), affected by attitude, the equivalent incident direction vector of the target in the body coordinate system. for:

[0052] For the Each array element, in Time, frequency The actual phase delay below It is no longer a simple linear relationship, but rather:

[0053] This means that the drone's attitude changes distort the one-dimensional linear array's observation model into three-dimensional space. To compensate for this effect in signal processing, this embodiment of the invention employs a manifold reconstruction method to correct multi-band radio frequency data, specifically including: First, the signal model of the multi-band radio frequency data is rewritten into a form that includes attitude terms:

[0054] in, Let t be the attitude vector of the UAV at time t.

[0055] Then, phase derotation will be performed on each snapshot data. For the lateral array (assuming it is distributed along the Y-axis of the body, i.e.) Its phase is mainly affected by The impact of the second line.

[0056] Finally, construct the compensation matrix. This makes the compensated data Data approximating the ideal horizontal flight attitude. The diagonal elements of the compensation matrix are:

[0057] in, It is by The calculated attitude deviation projection is then used. By fine-tuning the phase of the snapshot data, the dynamically changing array manifold is pulled back to the reference attitude (such as the horizontal attitude), thus ensuring the coherence of the subsequent covariance matrix accumulation.

[0058] In one implementation, for the aforementioned step S103, i.e. when constructing a virtual differential array based on the corrected multi-band radio frequency data, the following methods may be adopted, including but not limited to: first, calculating the theoretical covariance matrix of the corrected multi-band radio frequency data; then, performing column vectorization on the theoretical covariance matrix to obtain the virtual differential array, and determining the position set of each virtual array element in the virtual differential array.

[0059] In specific implementation, for the first Received signal at each frequency point Calculate its sampling covariance matrix (Assuming data was collected) (Data from individual snapshots)

[0060] In ideal circumstances ( The theoretical covariance matrix is:

[0061] in, It is a diagonal matrix of source power. It is the first A source at frequency The power below.

[0062] right Perform column vectorization to extract differential co-array information and obtain a virtual differential co-array. :

[0063] in, Indicates the Kronecker product; It is The matrix is ​​called the virtual steering matrix; It is the source power vector; This is the corresponding noise term.

[0064] Each row of the virtual steering matrix corresponds to the difference between the positions of two sensors in the physical array. Specifically, it represents the set of positions of the virtual array elements. (i.e., differential co-array) is:

[0065] For coprime arrays Include A unique virtual array element position, whose aperture is much larger than the physical aperture.

[0066] In one implementation, for the aforementioned step S103, i.e., when performing virtual aperture fusion on the virtual differential co-array to obtain a hole-free virtual differential co-array, the following methods may be adopted, including but not limited to: First, a reference frequency is determined based on the corrected multi-band radio frequency data, and the data of other frequencies in the corrected multi-band radio frequency data other than the reference frequency are mapped to the virtual space of the reference frequency to obtain a hole-free virtual differential co-array, and the location set of the hole-free virtual differential co-array is determined.

[0067] In practice, the corrected multi-band radio frequency data is used to fill in the gaps. The holes in the middle. First, define a reference frequency. (Usually the lowest frequency or center frequency is chosen), and other frequencies are... The observed data (i.e. radio frequency data) is mapped into the virtual space of the reference frequency.

[0068] In frequency Below, physical location difference The resulting phase difference is This is equivalent to the reference frequency. Below, by a location Phase difference generated by virtual sensors:

[0069] Therefore, frequency The equivalent set of virtual array element positions contributed is:

[0070] Through joint processing For each frequency, the obtained multi-frequency virtual co-array location set (i.e., the location set of the hole-free virtual differential co-array) is the union of all single-frequency virtual sets and their scaled versions, that is:

[0071] In this embodiment of the invention, see Figure 3 The diagram illustrates a coprime array physical structure and a multi-frequency virtual aperture expansion principle, demonstrating how a reasonable frequency set can be designed. (For example, making the frequency ratio) It can generate complementary virtual positions, filling holes in single-frequency co-arrays, and even expanding the virtual aperture several times the physical aperture. For example, if there are positional differences in the physical array... In single-frequency operation, position 1 is missing; if frequency is introduced... Then the position difference 2 is in The lower equivalent is Position 1 below, thus filling the hole.

[0072] After completing multi-band virtual aperture fusion and attitude compensation, a high-dimensional virtual observation vector was obtained. To fully utilize its sparsity properties and further improve resolution, embodiments of the present invention employ weighted averages. The weighted l_1-norm minimization algorithm is used for DOA estimation. Specifically, it includes: (1) Based on the position set of the holeless virtual differential co-array, a multi-frequency joint virtual steering matrix is ​​constructed.

[0073] (2) Construct a multi-band joint observation model based on the multi-frequency joint virtual steering matrix to obtain the virtual observation vector.

[0074] In practical implementation, a multi-frequency joint virtual steering matrix (i.e., a global virtual manifold matrix) is constructed based on the position set of the hole-free virtual differential co-array. Combined with the position sets of each virtual array element, a multi-frequency joint observation model is constructed to obtain the virtual observation vector. The multi-frequency joint observation model is as follows:

[0075] in, For virtual observation vectors, It is a selection matrix used to remove duplicate virtual elements (redundant averaging). Based on The constructed global virtual manifold matrix. At this point, the system is equivalent to a virtual uniform linear array with a huge aperture and extremely dense density, which completely solves the grating lobe ambiguity problem and greatly improves the degrees of freedom.

[0076] (3) Discretize the preset spatial angle range into multiple grid points, and construct an overcomplete dictionary matrix based on the spatial angles corresponding to the grid points and the multi-frequency joint virtual steering matrix.

[0077] (4) Construct a sparse representation model of virtual observation vectors based on an overcomplete dictionary matrix.

[0078] In practical implementation, the first step is to construct a sparse representation model. The spatial angular range of interest (i.e., the defined spatial angular range, such as the lateral range) is then defined. arrive Discretize into A fine grid of dots ,in (The number of grids is much greater than the number of sources).

[0079] Based on multi-frequency joint virtual steering matrix A complete dictionary matrix has been constructed. :

[0080] in, It is a virtual array at an angle The guide vector at that location.

[0081] At this point, the DOA estimation problem is transformed into solving the sparse solution of the following system of linear equations (i.e., the sparse representation model of the virtual observation vector):

[0082] in, It is a sparse vector, where the position of its non-zero elements corresponds to the source angle and the magnitude corresponds to the source power; This is the residual term (which includes noise and model error).

[0083] (5) Solve the sparse representation model to obtain the direction finding results of the target.

[0084] In one implementation, the sparse representation model is first modified based on the deviation between the target and the grid points; then a weighted average is applied. The norm minimization algorithm is used to solve the modified sparse representation model to obtain the direction finding results of the target.

[0085] In practice, the traditional sparse reconstruction assumption is that the target falls exactly on the grid points, but the actual target angle is different. Often located between two grids, i.e. This off-grid error leads to decreased reconstruction accuracy and spurious peaks. Therefore, in this embodiment of the invention, a first-order Taylor expansion is used for correction. Specifically, it includes: Assuming the true angle of the target nearest grid point The deviation is (Right now Then the true steering vector can be approximated as:

[0086] in, It is the first derivative of the guiding vector.

[0087] Therefore, the corrected sparse representation model is as follows:

[0088] in, It is a derivative matrix. It is an auxiliary sparse vector (This is the Hadamard product). This model utilizes both the sparsity of the signal and the sparsity of the grid bias.

[0089] Finally, weighted average is used. The norm minimization algorithm solves the modified sparse representation model, specifically including: To improve the sparsity of the solution and suppress noise, the following joint optimization objective function is constructed:

[0090] Among them, the first item ( ) represents the data fitting error term.

[0091] The second item ( ) represents the weighted sparsity constraint term. It is a weighting matrix, and the weight design is based on the Capon spectrum or preliminary beamforming results, i.e. This results in a smaller penalty for angles where signals are likely to be present, while a larger penalty is imposed on noisy areas.

[0092] The third item ( ) is a grid deviation constraint term to prevent Excessive size leads to near-failure.

[0093] In this embodiment of the invention, the aforementioned non-convex optimization problem can be solved by alternating iterative directional multiplier method or block coordinate descent method, with alternating updates. and The process continues until convergence, ultimately yielding the target's direction finding results.

[0094] This invention provides a physically sparse, virtually dense space-frequency-time joint processing architecture. Considering the limited lateral (wing or fuselage side) installation space and the continuous motion of UAVs, a sparse configuration such as a coprime array or nested array is employed. See also... Figure 4 The flowchart shown here illustrates a multi-band array direction finding implementation, which mainly includes the following processes: (1) The direction finding mission is started.

[0095] (2) Lateral sparse array configuration optimization design.

[0096] Physical layer: Deploy a non-uniform sparse antenna array on the side of the UAV to cover a larger physical aperture with a number of array elements that is much smaller than the Rayleigh limit required.

[0097] (3) Multi-band observation and attitude data acquisition.

[0098] (4) Construct a multi-band sparse array signal model.

[0099] Signal layer: The front-end receiver utilizes the broadband characteristics of the target signal or the frequency agility characteristics at multiple frequency points to simultaneously acquire multi-band radio frequency data.

[0100] Preprocessing layer: An attitude perception module is introduced to obtain the roll, pitch and yaw angles of the UAV at each sampling moment, and to perform real-time phase correction on the guidance vector of the received signal to eliminate non-stationary errors introduced by the moving base.

[0101] (5) Multi-band virtual differential co-array synthesis (extended equivalent aperture).

[0102] Virtual layer: The corrected multi-band data undergoes second-order statistical processing (covariance matrix vectorization) to construct a virtual differential co-array. Utilizing frequency diversity characteristics, the sparse co-arrays at different frequencies are mapped to a unified reference wavelength space, filling the holes created by the sparse array and synthesizing a virtual uniform linear array with a huge aperture and no grating lobes.

[0103] (6) High-resolution DOA estimation based on sparse reconstruction (suppressing grating ambiguity).

[0104] Algorithm layer: Applying algorithms based on the virtual array model The norm-optimized sparse reconstruction algorithm transforms DOA estimation into a sparse signal recovery problem within a compressed sensing framework, enabling accurate target orientation.

[0105] (7) Real-time compensation for platform motion posture error.

[0106] (8) Output the target’s precise lateral bearing.

[0107] The method provided in this embodiment of the invention expands the aperture by constructing a virtual differential co-array through multi-band signal fusion, and achieves high-precision mapping from the array manifold to the target angle using a sparse reconstruction algorithm combined with an attitude decoupling model. Compared to traditional single-frequency uniform linear arrays, it offers higher resolution and stronger anti-blurring capabilities. See details... Figure 5 The image shows a comparison between a single-band grating lobe blurring effect and the multi-frequency suppression effect of the present invention. Figure 6 The diagram shows the convergence curve of the direction finding error before and after attitude compensation for a UAV, and... Figure 7 The image shown is a super-resolution direction finding effect based on sparse reconstruction. Compared with the prior art, the present invention has the following advantages: (1) The airborne detection system based on sparse array has the advantages of lightweight and low load, and is suitable for micro and small UAV platforms. Through non-uniform sparse array design, the number and weight of physical antennas are greatly reduced while maintaining large aperture observation capability, which improves the problem of limited lateral space of UAV.

[0108] (2) The multi-band fusion direction finding method has strong anti-interference adaptability and can work effectively in scenarios with grating lobe ambiguity and low signal-to-noise ratio. By utilizing the wavelength diversity characteristic to cancel spurious spectral peaks, the ambiguity problem of sparse arrays under single frequency is effectively solved, ensuring the uniqueness and robustness of direction finding results in complex electromagnetic environments.

[0109] (3) The use of virtual aperture synthesis and motion compensation algorithms improved the target resolution accuracy and significantly reduced the impact of platform jitter on direction finding. By mathematically constructing virtual array elements, the physical Rayleigh limit was broken, and the body attitude error was corrected in real time, realizing stable and ultra-high resolution tracking of lateral targets under dynamic flight conditions.

[0110] In addition to the UAV target direction finding method based on multi-band sparse arrays provided in the foregoing embodiments, this invention also provides a UAV target direction finding device based on multi-band sparse arrays. (See [link to previous document]). Figure 8 The schematic diagram shown illustrates the structure of a UAV target direction finding device based on a multi-band sparse array, indicating that the device mainly comprises the following parts: The data acquisition module 801 is used to acquire signals from the target through a non-uniform sparse antenna array deployed on the side of the UAV, and to sample the signals to obtain multi-band radio frequency data.

[0111] The correction module 802 is used to acquire the attitude angle of the UAV at each sampling time and correct the multi-band radio frequency data based on the attitude angle.

[0112] The virtual array construction and filling module 803 is used to construct a virtual differential common array based on the corrected multi-band radio frequency data, and to perform virtual aperture fusion on the virtual differential common array to obtain a hole-free virtual differential common array. The direction finding result generation module 804 is used to construct a multi-band joint observation model based on a hole-free virtual differential co-array, obtain virtual observation vectors, and perform weighted sparse reconstruction on the virtual observation vectors to obtain the direction finding results of the target.

[0113] The UAV target direction finding device based on a multi-band sparse array provided in this invention deploys a non-uniform sparse antenna array on the side of the UAV, which can maintain large-aperture observation capability while significantly reducing the number and weight of physical antennas, thus improving the problem of limited lateral space for the UAV. By constructing a virtual differential array using multi-band radio frequency data and performing virtual aperture fusion, the direction finding results of the target are obtained by using a weighted sparse reconstruction method, which effectively improves the ambiguity problem of sparse arrays under single frequency, ensures the uniqueness and robustness of the direction finding results in complex electromagnetic environments, and improves the target resolution accuracy.

[0114] In one embodiment, the correction module 802 is specifically used for: acquiring the attitude angle of the UAV relative to the geodetic coordinate system through the airborne inertial measurement unit; determining the attitude rotation matrix from the geodetic coordinate system to the body coordinate system based on the Euler angle rotation theorem and the attitude angle, and determining the direction cosine matrix based on the attitude rotation matrix; calculating the attitude deviation projection of the UAV based on the direction cosine matrix; performing phase inversion rotation on the multi-band radio frequency data, and calculating the diagonal elements of the compensation matrix based on the phase inversion multi-band radio frequency data and the attitude deviation projection to obtain the compensation matrix; and correcting the multi-band radio frequency data based on the compensation matrix.

[0115] In one embodiment, the virtual array construction and filling module 803 is specifically used to: calculate the theoretical covariance matrix of the corrected multi-band radio frequency data; perform column vectorization operation on the theoretical covariance matrix to obtain a virtual differential co-array, and determine the position set of each virtual array element in the virtual differential co-array.

[0116] In one embodiment, the virtual array construction and filling module 803 is specifically used to: determine a reference frequency based on the corrected multi-band radio frequency data, map the data of other frequencies in the corrected multi-band radio frequency data other than the reference frequency to the virtual space of the reference frequency, obtain a hole-free virtual differential common array, and determine the location set of the hole-free virtual differential common array.

[0117] In one embodiment, the direction finding result generation module 804 is specifically used for: constructing a multi-frequency joint virtual steering matrix based on the position set of the hole-free virtual differential co-array; constructing a multi-frequency band joint observation model based on the multi-frequency joint virtual steering matrix to obtain virtual observation vectors; discretizing a preset spatial angle range into multiple grid points, and constructing an overcomplete dictionary matrix based on the spatial angles corresponding to the grid points and the multi-frequency joint virtual steering matrix; constructing a sparse representation model of the virtual observation vectors based on the overcomplete dictionary matrix; and solving the sparse representation model to obtain the direction finding result of the target.

[0118] In one implementation, the direction finding result generation module 804 is further configured to: correct the sparse representation model based on the deviation between the target and the grid points; and employ weighted averages. The norm minimization algorithm is used to solve the modified sparse representation model to obtain the direction finding results of the target.

[0119] In one embodiment, the non-uniform sparse antenna array adopts a coprime linear array structure, which consists of two sparse uniform subarrays; the first subarray includes a first number of array elements, and the second subarray includes a second number of array elements; wherein the first number and the second number are coprime numbers, and the first subarray and the second subarray share the first array element.

[0120] It should be noted that the device provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment. The specific numerical values ​​provided in this embodiment are merely exemplary and are not intended to limit the scope of the invention.

[0121] This invention also provides an electronic device, specifically, the electronic device includes a processor and a storage device; the storage device stores a computer program, and the computer program, when run by the processor, executes the method described in any of the above embodiments.

[0122] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 100 includes: a processor 90, a memory 91, a bus 92, and a communication interface 93. The processor 90, the communication interface 93, and the memory 91 are connected through the bus 92. The processor 90 is used to execute executable modules, such as computer programs, stored in the memory 91.

[0123] The memory 91 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 93 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0124] Bus 92 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0125] The memory 91 is used to store programs. After receiving an execution instruction, the processor 90 executes the programs. The method executed by the device for defining the flow process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 90 or implemented by the processor 90.

[0126] The processor 90 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 90 or by instructions in software form. The processor 90 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 91. Processor 90 reads the information in memory 91 and, in conjunction with its hardware, completes the steps of the above method.

[0127] The computer program product of the readable storage medium provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the foregoing method embodiments. For specific implementation, please refer to the foregoing method embodiments, which will not be repeated here.

[0128] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0129] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A UAV target direction finding method based on a multi-band sparse array, characterized in that, include: Signals from the target are acquired by deploying a non-uniform sparse antenna array on the side of the UAV, and multi-band radio frequency data is obtained by sampling the signals. The attitude angle of the UAV at each sampling time is obtained, and the multi-band radio frequency data is corrected based on the attitude angle; A virtual differential common array is constructed based on the corrected multi-band radio frequency data, and virtual aperture fusion is performed on the virtual differential common array to obtain a hole-free virtual differential common array. A multi-band joint observation model is constructed based on the hole-free virtual differential co-array to obtain virtual observation vectors. The virtual observation vectors are then weighted and sparsely reconstructed to obtain the direction finding results of the target.

2. The method according to claim 1, characterized in that, Acquire the attitude angle of the UAV at each sampling time, and correct the multi-band radio frequency data based on the attitude angle, including: The attitude angle of the UAV relative to the geodetic coordinate system is obtained by the airborne inertial measurement unit; Based on Euler's angle rotation theorem and the attitude angle, the attitude rotation matrix from the geodetic coordinate system to the body coordinate system is determined, and the direction cosine matrix is ​​determined based on the attitude rotation matrix. Based on the direction cosine matrix, the attitude deviation projection of the UAV is calculated; The multi-band radio frequency data is phase-reversed, and the diagonal elements of the compensation matrix are calculated based on the phase-reversed multi-band radio frequency data and the attitude deviation projection to obtain the compensation perception matrix. The multi-band radio frequency data is corrected based on the compensation matrix.

3. The method according to claim 1, characterized in that, A virtual differential co-array is constructed based on the corrected multi-band radio frequency data, including: Calculate the theoretical covariance matrix of the corrected multi-band radio frequency data; The theoretical covariance matrix is ​​column-vectorized to obtain a virtual difference coarray, and the set of positions of each virtual array element in the virtual difference coarray is determined.

4. The method according to claim 3, characterized in that, Virtual aperture fusion is performed on the virtual differential co-array to obtain a hole-free virtual differential co-array, including: Based on the corrected multi-band radio frequency data, a reference frequency is determined, and data of other frequencies in the corrected multi-band radio frequency data other than the reference frequency are mapped to the virtual space of the reference frequency to obtain a hole-free virtual differential array, and the location set of the hole-free virtual differential array is determined.

5. The method according to claim 4, characterized in that, A multi-band joint observation model is constructed based on a hole-free virtual differential co-array to obtain virtual observation vectors. Weighted sparse reconstruction is then performed on these virtual observation vectors to obtain the direction-finding results for the target, including: Based on the location set of the holeless virtual differential co-array, a multi-frequency joint virtual steering matrix is ​​constructed. A multi-frequency joint observation model is constructed based on the multi-frequency joint virtual steering matrix to obtain the virtual observation vector; The preset spatial angle range is discretized into multiple grid points, and an overcomplete dictionary matrix is ​​constructed based on the spatial angles corresponding to the grid points and the multi-frequency joint virtual steering matrix. A sparse representation model of the virtual observation vector is constructed based on the overcomplete dictionary matrix; The sparse representation model is solved to obtain the direction finding results of the target.

6. The method according to claim 5, characterized in that, Solving the sparse representation model yields the direction finding results for the target, including: The sparse representation model is modified based on the deviation between the target and the grid points; Using weighted average The norm minimization algorithm is used to solve the modified sparse representation model to obtain the direction finding results of the target.

7. The method according to any one of claims 1-6, characterized in that, The non-uniform sparse antenna array adopts a coprime linear array structure, consisting of two sparse uniform subarrays; the first subarray includes a first number of array elements, and the second subarray includes a second number of array elements; wherein the first number and the second number are coprime numbers, and the first subarray and the second subarray share the first array element.

8. A UAV target direction finding device based on a multi-band sparse array, characterized in that, include: The data acquisition module is used to acquire signals from the target through a non-uniform sparse antenna array deployed on the side of the UAV, and to sample the signals to obtain multi-band radio frequency data. The correction module is used to acquire the attitude angle of the UAV at each sampling time and correct the multi-band radio frequency data based on the attitude angle; The virtual array construction and filling module is used to construct a virtual differential common array based on the corrected multi-band radio frequency data, and to perform virtual aperture fusion on the virtual differential common array to obtain a hole-free virtual differential common array. The direction finding result generation module is used to construct a multi-band joint observation model based on the hole-free virtual differential co-array, obtain virtual observation vectors, and perform weighted sparse reconstruction on the virtual observation vectors to obtain the direction finding results of the target.

9. An electronic device, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, The computer program is executed by the processor to perform the steps of the method described in any one of claims 1 to 7.

Images