Antenna array layout method and device and storage medium

Abstract

The invention discloses an antenna array layout method, which is used for adjusting a plurality of array element positions of a plurality of antenna array elements in an antenna array. The method comprises the following steps: constructing a plurality of steering vectors according to a plurality of arrival angles and a plurality of array element positions; the at least one transmitting antenna array transmits signals to a target to form a first phase difference, the at least one receiving antenna array receives signals to form a second phase difference, and each of the plurality of steering vectors is constructed according to the first phase difference and the second phase difference. The plurality of arrival angles are angles relative to the antenna array when the target is at different positions; constructing a confusion matrix according to the correlation among the plurality of steering vectors; constructing a target function according to the confusion matrix; calculating a plurality of gradient vectors of the target function at a plurality of array element positions; and adjusting the plurality of array element positions according to the plurality of gradient vectors to obtain a plurality of updated array element positions. In addition, the invention also provides an antenna array layout device and a storage medium.

Description

Technical Field

[0001] This disclosure relates to methods for antenna array layout, and particularly to a method, apparatus, and storage medium for antenna array layout based on gradient optimization. Background Technology

[0002] Generally speaking, 4D millimeter-wave radar, in addition to measuring range, velocity, and azimuth, adds the ability to measure elevation angle, enabling it to detect the height information of obstacles. The main purpose of using multiple-input multiple-output (MIMO) antenna arrays in 4D millimeter-wave radar is to calculate the target angle, that is, to achieve sufficiently high angular resolution and low angular confusion within the expected field of view (FOV). Reflecting on the antenna array design, the core requirement is to maximize the phase difference of signals incident from different directions on each antenna within the expected FOV and angular resolution.

[0003] According to traditional signal processing theory (e.g., the Nyquist sampling theorem), antenna layouts based on uniform arrays can be designed, arranging multiple antenna elements at half-wavelength intervals and resolving angles using Fast Fourier Transform (FFT). However, antenna layouts based on uniform arrays require a large number of antenna elements to achieve high angular resolution, resulting in high costs and the need for additional shielding to handle crosstalk that may occur between closely spaced antenna elements. To reduce the number of antenna elements and reduce crosstalk, antenna layouts based on sparse arrays can be designed to increase the spacing between these elements, achieving a similar virtual antenna aperture with fewer elements. Furthermore, antenna layouts based on sparse arrays remove some antenna elements from a uniformly spaced full array while still allowing for angle resolution using the FFT algorithm.

[0004] Generally, antenna layout based on sparse arrays requires optimization to determine the positions of multiple antenna elements. However, traditional optimization methods, such as genetic algorithms and simulated annealing algorithms, suffer from high computational cost and slow optimization speed. Given the initial conditions of the circuit board and antenna elements, it often takes several days of optimization to find a relatively optimal solution that meets the required angular resolution, making it impractical. Summary of the Invention

[0005] This disclosure provides a method, apparatus, and storage medium for antenna array layout, which can find the optimal configuration of multiple antenna elements in an antenna array through gradient optimization under given antenna array settings. Compared with methods based on genetic algorithms or simulated annealing, the antenna array layout method of this disclosure achieves faster optimization speed and lower overall computational cost.

[0006] In a first aspect, embodiments of this disclosure provide a method for arranging an antenna array, used to adjust the positions of multiple antenna elements in the antenna array. The multiple antenna elements include at least one transmitting antenna element and at least one receiving antenna element. This method for arranging the antenna array includes:

[0007] Multiple steering vectors are constructed based on multiple angles of arrival and the positions of the multiple array elements. At least one transmitting antenna element transmits a signal toward the target, forming a first phase difference; at least one receiving antenna element receives the signal, forming a second phase difference; and each of the multiple steering vectors is constructed based on the first phase difference and the second phase difference. The multiple angles of arrival are angles relative to the antenna array when the target is at different positions.

[0008] Construct a confusion matrix based on the correlation between the multiple guidance vectors;

[0009] Construct the objective function based on the confusion matrix;

[0010] Calculate the multiple gradient vectors of the objective function at the locations of the multiple array elements; and

[0011] The positions of the multiple array elements are adjusted according to the multiple gradient vectors to obtain multiple updated array element positions.

[0012] In a second aspect, embodiments of this disclosure provide an apparatus for antenna array layout, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the antenna array layout method provided in embodiments of this disclosure.

[0013] Thirdly, embodiments of this disclosure provide a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method for antenna array layout as provided in embodiments of this disclosure.

[0014] Based on the above disclosure, in the antenna array layout method described above, multiple steering vectors are constructed according to multiple angles of arrival and multiple element positions, and a confusion matrix is ​​constructed based on the correlation between the multiple steering vectors. Next, an objective function is constructed based on the confusion matrix, multiple gradient vectors of the objective function at multiple element positions are calculated, and the multiple element positions are adjusted based on the multiple gradient vectors to obtain multiple updated element positions. The antenna array layout method disclosed above can find the optimal configuration of multiple antenna elements in an antenna array through gradient optimization under given antenna array settings. When a radar uses this optimized antenna array, it can improve the radar's angular resolution and reduce angular confusion. Compared with methods based on genetic algorithms or simulated annealing, the antenna array layout method of this disclosure embodiment can achieve a faster optimization speed, and even after multiple iterations in the optimization process, the required computational load remains low. Attached Figure Description

[0015] The accompanying drawings exemplify embodiments and form part of the specification, serving together with the textual description to explain exemplary implementations of the embodiments. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort. In all the drawings, the same reference numerals refer to similar but not necessarily identical elements.

[0016] Figure 1 This is a schematic diagram illustrating the steps of the antenna array layout method in an embodiment of this disclosure;

[0017] Figure 2 This is a schematic diagram of the azimuth, elevation, and azimuth and elevation angles in this disclosure;

[0018] Figure 3 A schematic diagram illustrating the formation of an equivalent virtual antenna array for a multi-input multi-output antenna array;

[0019] Figure 4 A visualization of a portion of the four-dimensional confusion matrix of a conventional antenna array;

[0020] Figure 5 This is a schematic diagram of the antenna array before adjusting the positions of multiple array elements in an embodiment of this disclosure;

[0021] Figure 6 This is a schematic diagram of the antenna array after multiple iterations and adjustments of the positions of multiple array elements in an embodiment of this disclosure;

[0022] Figure 7 This is a schematic diagram of the antenna array and its equivalent virtual antenna array before adjusting the positions of multiple array elements in the embodiments of this disclosure;

[0023] Figure 8 This is a schematic diagram of the antenna array and its equivalent virtual antenna array after multiple iterations and adjustments of the positions of multiple array elements in the embodiments of this disclosure.

[0024] Figure 9 A visualization of a portion of a four-dimensional confusion matrix simulated in an embodiment of this disclosure;

[0025] Figure 10 This is a schematic diagram of the digital beamforming pattern of the antenna array in the azimuth dimension according to an embodiment of the present disclosure;

[0026] Figure 11 This is a schematic diagram of the digital beamforming pattern of the antenna array in the elevation dimension according to an embodiment of the present disclosure;

[0027] Figure 12 A visualization of a portion of a four-dimensional confusion matrix simulated for a comparative embodiment;

[0028] Figure 13 for Figure 12 A schematic diagram of the digital beamforming pattern of the antenna array in the azimuth dimension of the comparative embodiment;

[0029] Figure 14 for Figure 12 A schematic diagram of the digital beamforming pattern of the antenna array in the elevation dimension of the comparative embodiment;

[0030] Figure 15 This is a structural block diagram of a computer device according to an embodiment of the present disclosure. Detailed Implementation

[0031] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0032] In this disclosure, the term "a plurality of" refers to two or more unless otherwise stated. In this disclosure, unless otherwise stated, the terms "first," "second," etc., are used to distinguish similar objects and are not intended to limit their positional, temporal, or importance relationships. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of the invention described herein can be implemented in ways other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatuses. To enable those skilled in the art to better understand this application, some technical terms appearing in the embodiments of this application are explained below:

[0033] 4D millimeter-wave radar: A millimeter-wave radar that can measure distance, velocity, azimuth, and elevation.

[0034] MIMO: Multiple Input Multiple Output, an antenna array with multiple inputs and multiple outputs.

[0035] FOV: Field of View. In millimeter-wave radar, it refers to the expected radar scanning area.

[0036] ADAM: Adaptive Moment Estimation, a gradient-based optimizer primarily used in neural network optimization.

[0037] In this embodiment, the MIMO antenna system can be used in vehicle-mounted radar, such as a vehicle-mounted 4D millimeter-wave radar, and can be implemented using a single chip or a cascaded multi-chip approach. The antenna array in the MIMO antenna system is composed of multiple antennas arranged in a row. The individual antenna elements that make up the antenna array are called array elements. Specifically, the MIMO antenna system may include multiple transmit antenna array elements, multiple receive antenna array elements, and multiple cascaded radio frequency chips, wherein these transmit antenna array elements and these receive antenna array elements constitute the antenna array.

[0038] Figure 2This is a schematic diagram illustrating the azimuth, elevation, and azimuth / elevation angles in this disclosure. Specifically, a two-dimensional plane coordinate system can be established on the plane where the antenna array is located. One coordinate axis is called the elevation dimension or elevation axis, and the other coordinate axis perpendicular to it is called the azimuth dimension or azimuth axis. When a vehicle-mounted radar using a MIMO antenna system is in use, the plane formed by the antenna array will be perpendicular to the ground plane (i.e., the ground). At this time, the elevation dimension or elevation axis is perpendicular to the ground plane, and the azimuth dimension or azimuth axis is parallel to the ground plane.

[0039] In this disclosure, "azimuth" and "elevation" refer to directional dimensions defined in a geographic coordinate system with the radar center point as the coordinate center. Azimuth refers to a direction on the ground plane (or a direction parallel to the ground plane), and correspondingly, the azimuth angle refers to the angle from true north, rotating counterclockwise around the horizon to the target direction line. Elevation refers to a direction on a plane perpendicular to the ground plane (or a direction perpendicular to the ground plane), and correspondingly, the elevation angle refers to the angle formed by the target direction line and the ground plane. A positive elevation angle occurs when the target direction line is above the ground plane; a negative depression angle occurs when the target direction line is below the ground plane. Figure 2 As shown, for a target direction line, i.e. Figure 2 The vector indicated by the thick arrow in the diagram, when rotated counterclockwise from true north on the ground plane, has an azimuth angle θ. The angle from which the vector's projection onto the ground plane is rotated is the elevation angle.

[0040] Figure 3 This diagram illustrates the formation of an equivalent virtual antenna array for a MIMO antenna array. The principle of the antenna array in a MIMO antenna system includes using N... tx One transmitting antenna element and N rx An antenna pair composed of 10 receiving antenna elements serves to generate N tx ×N rx The function of each virtual antenna element. For example, Figure 3 The antenna array in the image consists of four transmit antenna elements. These transmit antenna elements are arranged at intervals of half the signal wavelength in both the x-axis and z-axis. Figure 3 The antenna array also includes eight receiving antenna elements arranged at half-wavelength intervals along the x-axis. In this schematic diagram, this antenna array forms an equivalent of 32 virtual antenna elements. In embodiments of this disclosure, the x-axis is, for example, the azimuth dimension described above, and the z-axis is the elevation dimension described above. In other embodiments, the x-axis may also be the elevation dimension, and the z-axis may be the azimuth dimension.

[0041] When a radar using an antenna array detects a target, the transmitting antenna element emits a signal toward the target, and the receiving antenna element receives the reflected signal. The target has an azimuth angle θ relative to the antenna array in the azimuth dimension and an elevation angle in the elevation dimension. The whole is represented as the angle of arrival. by Figure 3 In the example antenna array, the coordinates of these transmitting antenna elements can be represented as follows: Where N t This indicates the number of these transmitting antenna elements. The coordinates of multiple receiving antenna elements can be represented as follows: Where N r This indicates the number of these receiving antenna array elements.

[0042] When these transmitting antenna elements of the antenna array transmit signals with wavelength λ toward the target, a phase difference will exist. Phase difference Specifically, this can be expressed by formula (1):

[0043]

[0044] When these receiving antenna elements of the antenna array receive the reflected signal, a phase difference will exist. Phase difference Specifically, this can be expressed by formula (2):

[0045]

[0046] Based on the above formulas (1) and (2), for the formed virtual antenna array, the phase difference of the signal from transmission to reception can be expressed as a vector. vector Specifically, this can be expressed by formula (3):

[0047]

[0048] Specifically, vector The steering vector, for example, can represent the overall response of multiple elements of an antenna array to a signal. When considering any two different angles of arrival... The phase difference presented across these antenna elements will be different. Maximizing this difference helps improve the angular resolution for target detection. Mathematically, an antenna design that improves angular resolution is one where, for any two different angles of arrival... The correlation coefficient between the two corresponding steering vectors should be kept as low as possible. In one implementation, a confusion matrix can be constructed based on the correlation between multiple steering vectors. For example, it can be traversed over specific angles of arrival. The confusion matrix is ​​obtained by identifying all steering vectors within the specified range and calculating the correlation coefficients between these steering vectors. Confusion Matrix Specifically, this can be expressed by formula (4):

[0049]

[0050] Where i, j, k, l, r, e are the parameters used for the above traversal. Confusion matrix For example, a four-dimensional confusion matrix. Figure 4 A visualization of a portion of the four-dimensional confusion matrix of a conventional antenna array. Figure 4 The system is divided into multiple zones along several discrete azimuth angles (in degrees °) and several discrete elevation angles (in degrees °). For each zone, the horizontal and vertical axes range, for example, from 0 degrees to 180 degrees. The color variations within each zone reflect the specific angle of arrival. The corresponding degree of autocorrelation. Specifically, for two very close angles of arrival... The two corresponding steering vectors will have a high correlation, resulting in a high degree of autocorrelation. Therefore, for each region, the angle of arrival corresponding to that region will be used. The location of this peak, characterized by high autocorrelation, is called the main lobe. For example, in... Figure 4 In region B, the main lobe will form at a position corresponding to the same angle of arrival (90°, 90°). Figure 4 In region C, the main lobe will form at positions corresponding to the same angle of arrival (150°, 90°). The smaller the width of the main lobe, the higher the angular resolution of the antenna array.

[0051] exist Figure 4 There are two bright spots in region A, located at the corresponding angles of arrival (90°, 150°) and (90°, 30°) respectively. This means that a target located at the angle of arrival (90°, 150°) may be confused with another target located at the angle of arrival (90°, 30°). Figure 4 The size of the bright spot in the image can represent the angular resolution. If one of the aforementioned regions has multiple bright spots (such as in region A), it indicates that in addition to the main lobe, there is also a side lobe with relatively high autocorrelation. The existence of side lobes means that the antenna array will obtain similar steering vectors for different angles of arrival, causing confusion in target detection.

[0052] Figure 5 This is a schematic diagram of the antenna array before adjusting the positions of multiple array elements in an embodiment of this disclosure. Figure 6 This is a schematic diagram of an antenna array after multiple iterations and adjustments to the positions of multiple antenna elements, as described in this disclosure. This disclosure provides a method for antenna array layout that, under given antenna array settings, uses gradient optimization to find the optimal configuration of multiple antenna elements in the antenna array. The antenna array layout method of this disclosure is used to adjust the positions of multiple antenna elements in an antenna array. The antenna array can be used, for example, in 4D millimeter-wave radar and can meet the angular resolution requirements of imaging radar.

[0053] The antenna array can be, for example, a MIMO antenna array. The multiple antenna elements of the antenna array include at least one transmit antenna element (e.g., Figure 5 Multiple transmit antenna elements 402) and at least one receive antenna element (e.g., Figure 5 (Multiple receiving antenna elements 404). Specifically, such as... Figure 5 As shown, multiple array element positions are multiple initial array element positions, and multiple antenna array elements are arranged to form a rectangle on a circuit board. At least three of the multiple antenna array elements are arranged in a line with equal spacing. In some embodiments of this disclosure, the initial setting of arranging them into a rectangle is beneficial for subsequent optimization. However, in some embodiments, the multiple initial array element positions can also be randomly arranged on the circuit board. The circuit board can be, for example, a printed circuit board (PCB).

[0054] In this embodiment, at least one transmitting antenna element (multiple transmitting antenna elements 402) transmits a signal toward the target, forming a first phase difference. At least one receiving antenna element (multiple receiving antenna elements 404) receives the signal, forming a second phase difference. The electromagnetic wave frequency band of the signal can be the millimeter wave band. The method of antenna array layout includes: constructing multiple steering vectors based on multiple angles of arrival and the positions of the multiple elements; and constructing a confusion matrix (e.g., a four-dimensional confusion matrix) based on the correlation between the multiple steering vectors. The multiple angles of arrival are the angles relative to the antenna array when the target is at different positions. Furthermore, each of the multiple steering vectors is constructed based on the first phase difference and the second phase difference. In detail, the step of constructing the confusion matrix includes: traversing the multiple steering vectors and calculating the correlation coefficient between any two of the multiple steering vectors to obtain multiple correlation coefficients; and constructing the confusion matrix based on the multiple correlation coefficients. For the construction of the multiple steering vectors and the construction of the confusion matrix, at least refer to the section on... Figure 3-4 Related descriptions.

[0055] In this embodiment, in order to make the confusion matrix By reducing the width of the main lobe and suppressing the side lobes, the confusion matrix can be reduced. The overall reduction in autocorrelation is used as the optimization objective for the antenna array. In this embodiment, the antenna array layout method further includes: constructing an objective function based on the confusion matrix. Specifically, in order to efficiently reduce the confusion matrix... The overall autocorrelation value can be calculated using LogSumExp(LSE) to construct an objective function L suitable for gradient optimization. target Objective function L target Specifically, this can be expressed by formula (5):

[0056]

[0057] Here, 's' serves as a smoothing factor, and an appropriate value can be set according to the needs of the optimization experiment. For example, in this embodiment, setting the smoothing factor to 10 can achieve better optimization results.

[0058] In this embodiment, the antenna array layout method further includes: calculating multiple gradient vectors of the objective function at multiple array element positions, and adjusting the multiple array element positions according to the multiple gradient vectors to obtain multiple updated array element positions. Specifically, due to some physical limitations in the actual arrangement of the multiple antenna elements, the antenna array layout method in this embodiment further includes constructing a constraint loss associated with the multiple array element positions. Furthermore, the step of calculating multiple gradient vectors of the objective function at multiple array element positions includes calculating the multiple gradient vectors of the objective function at multiple array element positions while ensuring that the constraint loss is substantially zero.

[0059] Please continue to refer to this. Figure 5-6 Specifically, multiple antenna elements (e.g., multiple transmitting antenna elements 402 and multiple receiving antenna elements 404) are disposed on a rectangular circuit board 408, and the multiple antenna elements are connected to at least one chip 406. Figure 5-6 The various elements are represented by boxes, which briefly show their approximate size in space. Figure 5-6 The specific dimensions and shapes of these components can be set according to actual needs. The aforementioned constraint loss is used to measure the degree to which at least one of the following conditions (e.g., conditions one to five) is satisfied.

[0060] Condition 1: Multiple antenna elements do not overlap with each other in the direction perpendicular to the rectangular circuit board 408. The aforementioned constraint loss may include the first sub-constraint loss L. AA This is used to measure the degree to which condition one is satisfied. The first sub-constraint loss L... AA This can be expressed by formula (6):

[0061]

[0062] In the scenario described in condition one, N represents the number of antenna elements. Each antenna element has a width w and a length h. For any two antenna elements, their center points have corresponding element positions (x, y) in the coordinate system. i ,y i ) and (x j ,y j At this time, I i,j This can be expressed by formula (7):

[0063]

[0064] Where min represents the minimum value and max represents the maximum value.

[0065] Condition 2: Multiple antenna array elements and at least one chip 406 do not overlap with each other in a direction perpendicular to the rectangular circuit board 408. The aforementioned constraint loss may include a second sub-constraint loss L. AC This is used to measure the degree to which condition two is satisfied. The second sub-constraint loss L... AC This can be expressed by formula (8):

[0066]

[0067] In the second scenario, N represents the number of antenna elements. M represents the number of at least one chip. Each of the antenna elements has a width w1 and a length h1. Each of the at least one chip has a width w2 and a length h2. For any two antenna elements, their center points have corresponding element positions (x, y) in the coordinate system. i ,y i ) and (x j ,y j At this time, J i,j This can be expressed by formula (9):

[0068]

[0069] Where min represents the minimum value and max represents the maximum value.

[0070] Condition 3: The placement area of ​​multiple antenna array elements does not exceed the area defined by the rectangular circuit board 408. The above constraint loss may include the third constraint loss L. PCB This is used to measure the degree to which condition three is satisfied. In the context of condition three, the two vertices of the diagonal of the rectangular circuit board 408 in the coordinate system are (0,0) and (P). x ,P y For any one of the multiple antenna array elements, the center point of the antenna array element has an element position (x) in the coordinate system. i ,y i The third constraint loss L)PCB This can be expressed by formula (10):

[0071]

[0072] Where N is the number of the plurality of antenna array elements, and max represents taking the maximum value.

[0073] Condition 4: Multiple antenna array elements include multiple first antenna array elements. These multiple first antenna array elements are connected to the same chip 406 (e.g., [chip name missing]). Figure 6 One of the four chips 406 is connected to the first antenna array element. Multiple first antenna array elements form multiple interconnects 502 with the chip 406, and the included angle γ between any two interconnects 502 is greater than or equal to 15 degrees. The aforementioned constraint loss may include a fourth sub-constraint loss L. γ This is used to measure the degree to which condition four is satisfied. The fourth sub-constraint loss L... γ It can be expressed by formula (11):

[0074]

[0075] In the scenario of condition four, K is the number of multiple first antenna elements, and max represents taking the maximum value.

[0076] Condition 5: The plurality of first antenna array elements connected to chip 406 include at least one first transmitting antenna array element and at least one first receiving antenna array element (e.g., Figure 6 The chip 406 in the upper left corner is connected to multiple transmit antenna elements 402 and multiple receive antenna elements 404, and there is a dividing line such that at least one first transmit antenna element and at least one first receive antenna element are located on opposite sides of the dividing line (for example, there is a dividing line such that...). Figure 6 The chip 406 in the upper left corner is connected to multiple transmit antenna elements 402 and multiple receive antenna elements 404, which are located on opposite sides of the dividing line. This avoids the situation where transmit and receive antenna elements corresponding to the same chip are arranged in an alternating pattern. The aforementioned constraint loss may include the fifth sub-constraint loss L. Cross This is used to measure the degree to which condition five is satisfied. In the scenario of condition five, the multiple first antenna array elements connected to chip 406 form multiple interconnections 502 with chip 406. The angle between any one of the multiple interconnections 502 and the aforementioned boundary line is... Fifth Sub-Constraint Loss L cross This can be expressed by formula (12):

[0077]

[0078] The average angle between at least one first transmitting antenna element connected to chip 406 and the aforementioned boundary line is: The average angle between at least one first receiving antenna element connected to chip 406 and the aforementioned boundary line is: T represents the number of at least one first transmitting antenna elements connected to chip 406. R represents the number of at least one first receiving antenna elements connected to chip 406. max indicates the maximum value.

[0079] In this embodiment, the constraint loss may include a weighted sum of multiple sub-constraint losses. These sub-constraint losses may include, for example, a first sub-constraint loss L. AA Second sub-constraint loss L AC Third contract constraint loss L PCB Fourth Sub-Constraint Loss L γ and the fifth sub-constraint loss L cross These are used to measure the degree to which conditions one, two, three, four, and five are satisfied, respectively. More specifically, the objective function L can be synthesized. target With the relevant constraints, the optimization equation is obtained, as shown in formula (13):

[0080] L = L target +0.01L AA +0.01L AC +0.01L PCB +0.1L γ +0.1L cross Formula (13)

[0081] In the antenna array layout method of this embodiment, when calculating multiple gradient vectors of the objective function at multiple array element positions, the coefficients of the above-mentioned constraints (e.g., the weighting coefficients multiplied by each sub-constraint loss in formula (13)) need to be set to values ​​that make the above-mentioned constraint loss equal to 0. If the coefficients are set too large, a large gradient will be generated when gradient optimization goes out of bounds, which may destroy the optimization of the main objective function. If the coefficients are set too small, the constraint loss may become ineffective. In detail, appropriate weighting coefficients can be set according to experiments.

[0082] Please refer to Figure 5 In this embodiment, the step of calculating multiple gradient vectors of the objective function at multiple array element positions, as described above, includes differentiating the objective function at each of the multiple array element positions to obtain multiple gradient vectors. Specifically, this can be achieved by differentiating the objective function L at each array element position while ensuring that the constraint loss is substantially zero. target exist Figure 5 For each of the multiple array element positions, the derivative is calculated to obtain multiple gradient vectors corresponding to these element positions. These gradient vectors can be used to guide the optimization of these element positions.

[0083] In this embodiment, the step of adjusting the positions of multiple array elements based on multiple gradient vectors to obtain multiple updated array element positions includes, for each of the multiple array element positions, adding a movement vector 401 to the array element position to obtain an updated array element position. The movement vector 401 is the product of the gradient vector in the opposite direction and a preset step size. Figure 5 As can be seen, for each of the multiple transmitting antenna elements 402 and multiple receiving antenna elements 404, a movement vector 401 can be obtained to guide the update of the element position of that antenna element.

[0084] Furthermore, the antenna array layout method includes constructing multiple updated steering vectors based on multiple angles of arrival and multiple updated element positions, and iterating to adjust the multiple updated element positions. In other words, the process of updating the multiple element positions of multiple antenna elements can be performed multiple times to seek convergence of the calculation results, and thus find the optimal configuration of the antenna array. Figure 6 This refers to the antenna array after multiple iterations and adjustments to the positions of various array elements.

[0085] Specifically, the above formula (13) is differentiable, and therefore suitable for optimization using gradient descent. In this embodiment, the PyTroch framework or other neural network computing frameworks can be used to calculate the derivative of the objective function at multiple array element positions. During iteration, Adaptive Moment Estimation (ADAM) can be used as an optimizer to assist in adjusting the parameters of the above formula (13) to accelerate the convergence of the calculation results. The above preset step size is also called the learning rate. By setting an appropriate learning rate, the equation can be optimized through multiple iterations to converge it and find a local optimum. The above preset step size can be reasonably set according to the computational requirements. In addition, since the above optimization formula (formula (13)) is nonconvex, there will be a large number of local optima, which makes the optimization result roughly fine-tuned around the initial setting state. Therefore, a good initial setting is very important. In this embodiment, the initial setting arranged to form a rectangle can obtain better optimization results.

[0086] Figure 7 This is a schematic diagram of the antenna array and its equivalent virtual antenna array before adjusting the positions of multiple array elements in the embodiments of this disclosure; Figure 8 This is a schematic diagram of the antenna array and its equivalent virtual antenna array after multiple iterations and adjustments of the positions of multiple array elements in the embodiments of this disclosure. Figure 7-8 The multiple transmitting antenna elements 602, multiple receiving antenna elements 604, and circuit board 608 are respectively similar to Figure 5-6The system includes multiple transmitting antenna elements 402, multiple receiving antenna elements 404, and a circuit board 408. Figure 7-8 The multiple virtual antenna array elements 606 in the middle are similar to Figure 3 Multiple virtual antenna array elements in the array. Figure 7-8 The description of the multiple components can be found at least in [reference]. Figure 3 , 5 Description of the multiple components corresponding to -6. Figure 7-8 Multiple elements are represented by solid rectangles, solid circles, or hollow triangular rectangles, which briefly show their approximate location in space, while omitting their specific size and shape. Figure 7-8 The specific dimensions and shapes of these components can be set according to actual needs. Furthermore, to avoid affecting the presentation of multiple virtual antenna elements 606, Figure 7-8 The chip is omitted in the middle.

[0087] In detail, Figure 7 This presents the initialization settings for optimizing multiple antenna elements using the antenna array layout method of embodiments of this disclosure. Figure 7 In this configuration, the antenna elements are arranged in a rectangular shape. The transmitting antenna elements 602 are arranged in the z-axis, and the receiving antenna elements 604 are arranged in the x-axis. The x-axis is, for example, the azimuth dimension described above, and the z-axis is the elevation dimension described above. In other embodiments, the x-axis may also be the elevation dimension, and the z-axis may be the azimuth dimension. In this embodiment, the antenna array forms an equivalent plurality of virtual antenna elements 606.

[0088] Figure 8 The present invention presents the setup of multiple antenna array elements after multiple iterations and adjustments, through the antenna array layout method of the embodiments of this disclosure. Figure 7-8 The changes in the element positions of these virtual antenna elements 606 are presented based on the changes in the element positions of these transmitting antenna elements 602 and these receiving antenna elements 604.

[0089] Figure 9 A visualization of a portion of a four-dimensional confusion matrix simulated for an embodiment of this disclosure. Figure 9 Multiple subplots are presented based on different discrete azimuth and elevation angles. The color changes in each subplot reflect a specific angle of arrival. The corresponding degree of autocorrelation. To clearly present each of these small plots, Figure 9 The expected results for all discrete azimuth and pitch angles within the field of view were not presented. Overall... Figure 9In each of these small images, we can see that apart from a bright spot at the main lobe location, there are no obvious side lobes elsewhere, indicating that the side lobes are well suppressed. The peak value of the correlation coefficient is also displayed below each of these small images. Figure 9 In one embodiment, the average peak value of the correlation coefficients corresponding to all small plots within the expected field of view (e.g., within an azimuth angle of 120° and a pitch angle of 30°) is 0.316.

[0090] Figure 10 This is a schematic diagram of the digital beamforming pattern of the antenna array in the azimuth dimension according to an embodiment of the present disclosure. Figure 11 This is a schematic diagram of the digital beamforming pattern of the antenna array in the elevation dimension, according to an embodiment of this disclosure. Figure 10-11 In the diagram, the horizontal axis represents angle, and the vertical axis represents normalized gain. Antenna radiation patterns typically have two or more lobes, with the lobe having the highest radiation intensity being the main lobe and the others being side lobes. The angle between two points on either side of the main lobe's maximum radiation direction, where the radiation intensity decreases by 3 dB (power density is halved), is defined as the beamwidth (also known as the lobe width, main lobe width, or half-power angle). The beamwidth reflects the angular resolution of the radar employing the antenna array of this disclosure.

[0091] In this embodiment, the 3dB beamwidth of the digital beamforming pattern in the azimuth dimension is 1.04°, and the sidelobes are -11.7dB. The 3dB beamwidth of the digital beamforming pattern in the elevation dimension is 1.14°, and the sidelobes are -10.6dB. That is, the theoretically identifiable azimuth resolution of the radar using the antenna array of this embodiment is 1.04°, and the elevation resolution is 1.14°.

[0092] Figure 12 This is a visualization of a portion of a simulated four-dimensional confusion matrix from a comparative embodiment. This comparative embodiment is not optimized using the antenna array layout method of the embodiments of this disclosure. Figure 12 and Figure 9 It can be seen that the four-dimensional confusion matrix of this comparative embodiment contains multiple vertical stripe-like side lobes. Figure 12 In the study, the average peak value of the correlation coefficients corresponding to all small plots within the expected field of view (e.g., an azimuth angle of 120° and an elevation angle of 30°) is as high as 0.579. Clearly, by optimizing the antenna array layout method according to embodiments of this disclosure, the overall correlation coefficient in the four-dimensional confusion matrix can be reduced.

[0093] Figure 13 for Figure 12 A schematic diagram of the digital beamforming pattern of the antenna array in the azimuth dimension of the comparative embodiment. Figure 14 for Figure 12A schematic diagram of the digital beamforming pattern of the antenna array in the elevation dimension of the comparative embodiment is provided. In this comparative embodiment, the 3dB beamwidth of the digital beamforming pattern in the azimuth dimension is 1.59°. The 3dB beamwidth of the digital beamforming pattern in the elevation dimension is 1.37°. The theoretically identifiable azimuth resolution of the radar using the antenna array of this comparative embodiment is 1.59°, and the elevation resolution is 1.37°. Specifically, the lower the angular resolution value, the higher the accuracy of target resolution, i.e., the higher the angular resolution. Obviously, by optimizing the antenna array layout method of the embodiments of this disclosure, the angular resolution of the radar can be improved.

[0094] Specifically, in the antenna array layout method of this disclosure embodiment, multiple steering vectors are constructed based on multiple angles of arrival and multiple element positions, and a confusion matrix is ​​constructed based on the correlation between the multiple steering vectors. Next, an objective function is constructed based on the confusion matrix, multiple gradient vectors of the objective function at multiple element positions are calculated, and the multiple element positions are adjusted based on the multiple gradient vectors to obtain multiple updated element positions. Therefore, the antenna array layout method can find the optimal configuration of these antenna elements in the antenna array under given antenna array settings, such as an initialized multiple antenna element configuration, through gradient optimization, for example, reducing the correlation coefficient value in the confusion matrix to facilitate a reduction in main lobe width and suppression of sidelobes. When a radar uses this optimized antenna array, it can improve the radar's angular resolution and reduce angular confusion. When the radar is, for example, an automotive 4D millimeter-wave radar, it can improve driving performance in many autonomous driving or assisted driving scenarios. Compared with methods based on genetic algorithms or simulated annealing, the antenna array layout method of this disclosure embodiment can achieve a faster optimization speed. Even after multiple iterations in the optimization process, the required computational load remains low.

[0095] For example, when optimizing an antenna array using a genetic algorithm with multiple antenna elements initially set up, a high computational load and several days are required to achieve the desired antenna element configuration. However, with the same computing power, when optimizing the antenna array using the antenna array layout method of this disclosure embodiment, even with tens of thousands of iterations during the optimization process, the overall computational load remains low, and the desired antenna element configuration can be obtained in just a few minutes.

[0096] In detail, the antenna array layout method of this disclosure optimizes the antenna array, which can be widely used in automotive radar, such as 4D millimeter-wave radar. 4D millimeter-wave radar serves as an effective sensor for perceiving the external environment in many autonomous driving or assisted driving scenarios, including Adaptive Cruise Control (ACC), Automatic Emergency Braking (AEB), Blind Spot Monitoring (BSD), Lane Change Assist (LCA), and Rear CrossTraffic Alert (RCTA). Improving the performance of 4D millimeter-wave radar is crucial for driving safety.

[0097] Figure 1 This diagram illustrates the steps of an antenna array layout method according to an embodiment of the present disclosure. This method can be executed by a device for antenna array layout. This device can be implemented by software and / or hardware, and is generally integrated into a computer device. The computer device can be integrated into a movable object, or it can be externally connected to or communicate with the movable object. Figure 1 For detailed descriptions or extended solutions of each step, please refer to the relevant descriptions of the embodiments of this disclosure above. Please refer to... Figure 1 The antenna array layout method of this disclosure includes:

[0098] Step S100: Construct multiple guidance vectors based on multiple angles of arrival and multiple array element positions;

[0099] After completing step S100, proceed to step S102. Step S102: Construct a confusion matrix based on the correlation between multiple steering vectors;

[0100] After completing step S102, proceed to step S104. Step S104: Construct the objective function based on the confusion matrix;

[0101] After completing step S104, proceed to step S106. Step S106: Calculate multiple gradient vectors of the objective function at multiple array element positions.

[0102] After completing step S106, proceed to step S108. Step S108: Adjust the positions of multiple array elements according to multiple gradient vectors to obtain multiple updated array element positions.

[0103] The antenna array layout method described in this embodiment can at least achieve the technical effects of the antenna array layout method mentioned above. It can find the optimal configuration of multiple antenna elements in the antenna array through gradient optimization, achieve a faster optimization speed, and has a lower overall computational load.

[0104] This disclosure provides a computer device in which the means for integrating the antenna array layout provided in this disclosure can be integrated. Figure 15 This is a structural block diagram of a computer device provided in an embodiment of this disclosure. The computer device 100 may include: a memory 101, a processor 102, and a computer program stored in the memory 101 and executable on the processor. When the processor 102 executes the computer program, it implements the antenna array layout method as described in the embodiment of this disclosure. The computer device can be integrated into a movable object; in this case, the computer device can also be considered the movable object itself, such as a vehicle. The computer device can also be externally connected to the movable object or communicate with the movable object.

[0105] The computer device provided in this disclosure embodiment can implement the antenna array layout method described in this disclosure embodiment, and can at least achieve the above-mentioned technical effects of the antenna array layout method. It can find the optimal configuration of multiple antenna elements in the antenna array through gradient optimization, achieve a faster optimization speed, and have a lower overall computational load.

[0106] This disclosure also provides a storage medium (computer-readable storage medium) containing computer-executable instructions that, when executed by a computer processor, implement the method for antenna array layout as provided in this disclosure.

[0107] Storage medium – any type of memory device or storage device. The term “storage medium” is intended to include: mounting media, such as CD-ROM, floppy disk, or magnetic tape devices; computer system memory or random access memory, such as DRAM, DDRRAM, SRAM, EDORAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (e.g., hard disk or optical storage); registers or other similar types of memory elements, etc. Storage medium may also include other types of memory or combinations thereof. Furthermore, storage medium may reside in a first computer system in which the program is executed, or it may reside in a different second computer system connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term “storage medium” can include two or more storage media that may reside in different locations (e.g., in different computer systems connected via a network). Storage medium may store program instructions (e.g., specifically implemented as a computer program) executable by one or more processors.

[0108] The antenna array layout apparatus, device, and storage medium provided in the above embodiments can execute the antenna array layout method provided in any embodiment of this disclosure, and have the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in the above embodiments can be found in the antenna array layout method provided in any embodiment of this disclosure.

[0109] While exemplary embodiments or examples of this disclosure have been described with reference to the accompanying drawings, it should be understood that the exemplary discussions above are not intended to be exhaustive or to limit the invention to the specific forms disclosed. Many modifications and variations are possible based on the teachings above. Therefore, the subject matter disclosed should not be limited to any single embodiment or example described herein, but should be interpreted in accordance with the breadth and scope of the appended claims.

Claims

1. A method for arranging an antenna array, used to adjust the positions of multiple antenna elements in the antenna array, wherein the multiple antenna elements include at least one transmitting antenna element and at least one receiving antenna element, characterized in that, The method includes: Multiple steering vectors are constructed based on multiple angles of arrival and the positions of multiple array elements, wherein at least one transmitting antenna element transmits a signal toward the target to form a first phase difference, and at least one receiving antenna element receives the signal to form a second phase difference, and each of the multiple steering vectors is constructed based on the first phase difference and the second phase difference, and the multiple angles of arrival are the angles relative to the antenna array when the target is at different positions; Construct a confusion matrix based on the correlation between the multiple guidance vectors; Construct the objective function based on the confusion matrix; Calculate the multiple gradient vectors of the objective function at the locations of the multiple array elements; and The positions of the multiple array elements are adjusted according to the multiple gradient vectors to obtain multiple updated array element positions.

2. The method according to claim 1, characterized in that, The calculation of the multiple gradient vectors of the objective function at the multiple array element positions includes: The objective function is differentiated at each of the plurality of array element positions to obtain the plurality of gradient vectors.

3. The method according to claim 1, characterized in that, The step of adjusting the positions of the plurality of array elements according to the plurality of gradient vectors to obtain the plurality of updated array element positions includes: For each of the plurality of element positions, a movement vector is added to the element position to obtain an updated element position, wherein the movement vector is the product of the gradient vector in the opposite direction and a preset step size.

4. The method according to claim 1, characterized in that, Including: Based on the multiple angles of arrival and the multiple updated element positions, multiple updated guidance vectors are constructed for iteration to adjust the multiple updated element positions.

5. The method according to claim 1, characterized in that, Including: Constructing a constraint loss associated with the multiple array element positions, wherein calculating the multiple gradient vectors of the objective function at the multiple array element positions includes: The objective function is computed at the plurality of gradient vectors at the plurality of array element locations while ensuring that the constraint loss is substantially zero.

6. The method according to claim 5, characterized in that, The plurality of antenna elements are disposed on a rectangular circuit board, and the plurality of antenna elements are connected to at least one chip, and the constraint loss is used to measure the degree to which at least one of the following conditions is satisfied: Condition 1: The plurality of antenna array elements do not overlap with each other in the direction perpendicular to the rectangular circuit board; Condition 2: The plurality of antenna array elements and the at least one chip do not overlap with each other in the direction perpendicular to the rectangular circuit board; Condition 3: The area where the multiple antenna array elements are set does not exceed the area defined by the rectangular circuit board; Condition 4: The plurality of antenna array elements includes a plurality of first antenna array elements, the plurality of first antenna array elements are connected to the same chip, the plurality of first antenna array elements and the chip form a plurality of first connection lines, and the included angle between any two of the plurality of first connection lines is greater than or equal to 15 degrees; and Condition 5: The plurality of first antenna array elements connected to the chip include at least one first transmitting antenna array element and at least one first receiving antenna array element, and there is a dividing line such that all at least one first transmitting antenna array element and all at least one first receiving antenna array element are located on opposite sides of the dividing line.

7. The method according to claim 1, characterized in that, The step of constructing the confusion matrix based on the correlation between the multiple steering vectors includes: The plurality of steering vectors are traversed, and the correlation coefficient between any two steering vectors is calculated to obtain multiple correlation coefficients; and The confusion matrix is ​​constructed based on the multiple correlation coefficients.

8. The method according to claim 1, characterized in that, The multiple array element positions are multiple initial array element positions, and the multiple antenna array elements are arranged to form a rectangle, with at least three antenna array elements among the multiple antenna array elements arranged in a line at equal intervals.

9. An apparatus for arranging an antenna array, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-8.

Images