Spatial optimization method for multi-antenna array in HT-6DMA base station, computer equipment and base station

Abstract

The invention belongs to the technical field of communication, and particularly relates to a spatial optimization method for a multi-antenna array in an HT-6DMA base station, computer equipment and a base station. According to the method, firstly, the elevation angle and azimuth angle of an antenna array in a global spherical coordinate system with the center of a base station as an original point are used for representing the position of the antenna array, and the elevation angle and azimuth angle of an external normal vector of the antenna array in a local spherical coordinate system with the center of the antenna array as the original point are used for representing the orientation of the antenna array; and a new HT-6DMA base station space model capable of independently adjusting the position and the orientation is redesigned. And iteratively optimizing the position / orientation of each antenna array through a hierarchical adjustment mode of firstly adjusting the position of the antenna array and then adjusting the orientation of the array and an alternate optimization method in combination with constraint conditions of space state optimization of the reconstructed antenna array. According to the scheme provided by the invention, the complexity of the optimization problem is greatly reduced, the optimization efficiency is improved, the performance of the base station after the space layout is optimized is improved, and the method can be suitable for various scenes.

Description

Technical Field

[0001] The present invention belongs to the field of communication technology, and in particular relates to a spatial optimization method for a multi-antenna array in an HT-6DMA base station, and corresponding computer equipment, computer program products and an HT-6DMA base station. Background Art

[0002] The sixth generation (6G) of wireless networks has attracted extensive research interest from both industry and academia. In addition to providing enhanced communication capabilities beyond the fifth generation (5G) network, 6G is also designed to support new application scenarios, including artificial intelligence (AI) and communication fusion, integrated communication and perception (ISAC), and ubiquitous connectivity. In 6G technology, multiple-input multiple-output (MIMO) technology is widely adopted, which not only significantly improves the efficiency and reliability of wireless communication systems through spatial multiplexing and diversity gain, but also improves the accuracy and resolution of wireless perception through the waveform and spatial diversity gain provided.

[0003] However, the above-mentioned traditional MIMO technology relies on fixed position antennas (FPAs) whose positions cannot be adjusted after deployment, and cannot adapt to channel fluctuations in space and time and changes in wireless tasks. To overcome this shortcoming, movable antennas (MA) have been introduced as an emerging technology, which can adjust the antenna position according to instantaneous or statistical channels and specific wireless tasks with different design goals. A large number of existing studies have demonstrated the advantages of MA over FPA in wireless communications, demonstrating its superiority in enhancing channel gain, improving MIMO channel capacity, flexible beamforming, interference suppression, and transmit power optimization. At present, research on MA mainly focuses on reconfiguring wireless channels through antenna position optimization, but more and more research interests are focused on using antenna position and direction / rotation adjustment to further improve the performance of different wireless tasks.

[0004] In order to obtain better communication performance, six-dimensional movable antenna (6DMA) technology was proposed. This type of antenna technology can more flexibly adjust the three-dimensional (3D) spatial position and 3D rotation of the antenna array. Compared with previous types of MA, 6DMA introduces several new features. First, 6DMA can naturally adapt to changes in the 3D wireless environment, which is difficult to achieve in traditional MA when the antenna can only move within a specific 1D or 2D area. Second, 6DMA adjusts the antenna position and rotation at the array level rather than the individual antenna level. These new features enable 6DMA to adapt more flexibly and efficiently to the slowly changing 3D large-scale user / target distribution in the network, thereby achieving more efficient deployment and reducing the frequency of antenna movement.

[0005] Although 6DMA technology has many advantages as mentioned above, there are still many challenges in the design and implementation of 6DMA that need to be solved urgently. First, since the 6DMA system usually needs to jointly optimize the 3D position and 3D rotation of all antenna arrays, the number of design variables in such problems will increase significantly with the increase in the number of arrays; on this basis, how to reduce the complexity of the spatial optimization problem of large-scale antenna arrays in base stations is the primary technical problem to be solved. Secondly, in the existing 6DMA model, the position and rotation variables of each array are defined in the Cartesian global coordinate system (CCS), which leads to a complex coupling relationship between them in the design objectives and the actual array motion constraints. This, in turn, will bring difficulties to the solution of related optimization problems; for example, in the existing optimization strategy based on the 6DMA model, the constraints related to the array orientation make the adjustment of the array position more limited, which not only increases the difficulty of convergence of the iterative results during the optimization process, but also causes the optimization of the global position and global rotation of the 6DMA array to easily fall into a poor local optimal solution. Summary of the invention

[0006] In order to solve the problem that the spatial optimization solution of the antenna array in the existing 6DMA base station is difficult to solve and the optimization result is easy to fall into the local optimal solution; the present invention provides a spatial optimization method of a multi-antenna array in a HT-6DMA base station, and its corresponding computer equipment, computer program product and base station.

[0007] The present invention is implemented by the following technical solutions: A method for spatial optimization of a multi-antenna array in an HT-6DMA base station, comprising: In the global spherical coordinate system with the center of the base station as the origin, the elevation angle of the antenna array is used and azimuth Define the position vector representing its position t : In the local spherical coordinate system with the center of the antenna array as the origin, the elevation angle of the antenna array's external normal vector relative to its own local spherical coordinate system is used and azimuth Define a rotation vector representing its orientation u ; ;based on t and u A spatial model of the HT-6DMA base station is constructed that can characterize the position and orientation of any antenna array in the base station.

[0008] Based on the new base station spatial model, the minimum distance constraint required for spatial optimization of the antenna array and the rotation constraint to avoid signal reflection are reconstructed; t and uThe decision variables are combined with the corresponding optimization objectives and scenario constraints to generate a mathematical model that characterizes the antenna array spatial optimization problem.

[0009] The mathematical model that characterizes the spatial optimization problem of antenna arrays is solved hierarchically. The process includes: first, fixing the local orientation of each antenna array to a spherical surface whose external normal vector is perpendicular to the global coordinate system, and then solving the mathematical model through an alternating optimization strategy to obtain the optimal position of each antenna array that satisfies the minimum distance constraint. . 2. In the best position The orientation of each antenna array on the system is rotated and adjusted, and the optimal orientation that satisfies the rotation constraint to avoid signal reflection is further solved through the alternating optimization strategy. .

[0010] The spatial optimization method of the multi-antenna array in the HT-6DMA base station provided by the present invention can be used to jointly optimize the position and orientation of all antenna arrays in the HT-6DMA base station in the uplink communication scenario, thereby maximizing the long-term average communication sum total rate of the user. It can also be used in air route perception to jointly optimize the position and orientation of all antenna arrays in the 6DMA base station and the covariance matrix of the transmitted signal, thereby maximizing the minimum received perception signal power on the flight segment.

[0011] The present invention also includes a computer device, which includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the spatial optimization method of the multi-antenna array in the HT-6DMA base station as described above is implemented, thereby realizing the joint optimization of the position and orientation of all antenna arrays in the HT-6DMA base station according to the scene requirements.

[0012] A computer program product comprises a computer program. When the computer program is executed by a processor, the computer program implements the aforementioned spatial optimization method for a multi-antenna array in an HT-6DMA base station.

[0013] The present invention also includes an HT-6DMA base station, which includes: a plurality of antenna arrays, an actuator and a controller.

[0014] The antenna arrays are distributed on the surface of a spherical space, and each antenna array includes a plurality of directional antenna units. The actuator includes a first actuator unit and a second actuator unit. The first actuator unit is used to move and adjust the position of each antenna array on the surface of the spherical space; the second actuator unit is used to further rotate and adjust the orientation of each antenna array at its respective position.

[0015] A computer program product as described above is deployed in the controller; the controller runs the computer program product according to the scene requirements, and then issues instructions to the actuator based on the jointly optimized optimal position and optimal orientation generated by the computer program product, thereby adaptively adjusting each antenna array to the optimal spatial state.

[0016] The technical solution provided by the present invention has the following beneficial effects: The spatial optimization method of multiple antenna arrays in the HT-6DMA base station provided by the present invention is used to jointly optimize the position and orientation of multiple antenna arrays that can be flexibly moved and rotated in spherical space. The scheme uses the elevation angle and azimuth angle of the antenna array in the global spherical coordinate system with the center of the base station as the origin to characterize its position, and uses the elevation angle and azimuth angle of the external normal vector of the antenna array in the local spherical coordinate system with its own center as the origin to characterize its orientation, thereby generating a new hierarchically adjustable six-dimensional movable antenna base station spatial model (i.e., HT-6DMA). Based on the new base station spatial model, the present invention reconstructs the constraints for optimizing the spatial state of the antenna array, and generates the optimal position and orientation of each antenna array through a two-stage optimization strategy in which the position and rotation of each antenna array are adjusted separately in sequence while one of them is fixed; thereby greatly reducing the optimization complexity and improving the quality of the solution.

[0017] The solution provided by the present invention has good universality in a variety of scenarios; in multi-user uplink communication scenarios and aerial perception scenarios, the optimization efficiency of the solution of the present invention is higher, the quality of the optimization results is also better, and it can avoid falling into the local optimum. Compared with fixed-position antennas, the present invention has significant improvements in both perception and communication performance. In addition, in communication scenarios, compared with existing solutions that can achieve joint optimization of position and orientation, the present invention can converge within a smaller number of iterations and has a higher optimization efficiency. In addition, this additional gain will increase further as the minimum spacing requirements that need to be met between arrays increase. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 It is a schematic diagram of a typical communication scenario using an HT-6DMA base station disclosed in Embodiment 1 of the present invention.

[0019] Figure 2 Schematic diagram of the method for converting between the local coordinates and the global coordinates of the antenna array in Example 1 of the present invention.

[0020] Figure 3This is a schematic diagram of the same air route in a three-dimensional Cartesian coordinate system and a two-dimensional angular coordinate system in an air route perception scenario according to Example 1 of the present invention.

[0021] Figure 4 Schematic diagram of the horizontal radiation pattern adopted by the antenna unit in the test experiment and its corresponding 3D directional gain.

[0022] Figure 5 For testing experiments, B =16, N =4 when the convergence curves of different schemes.

[0023] Figure 6 The spatial state distribution diagram of the antenna unit after optimization of different schemes for the test experiment.

[0024] Figure 7 The following is a performance comparison chart of the average total rate and execution time of different schemes in the uplink communication scenario in the test experiment.

[0025] Figure 8 This is a curve showing how the average total rate achievable by different schemes in the uplink communication scenario changes with the sparsity of user distribution in the test experiment.

[0026] Fig. 9 In order to test the HT-6DMA and HT-6DMA with PA schemes in different d min Comparison of the average total rate under the conditions of θ and the additional gain brought by local rotation of the array.

[0027] Fig.10 This is the spatial distribution map of the two routes included in the airspace perception scenario in the test experiment.

[0028] Fig.11 It is the convergence curve of three HT-6DMA schemes in the test experiment.

[0029] Fig.12 To test the distribution status of the three HT-6DMA schemes in the optimized antenna units in the experiment.

[0030] Fig.13 The beam pattern diagram after optimization of three HT-6DMA schemes in the test experiment.

[0031] Fig.14 In order to test the HT-6DMA in the experiment, the covariance matrix of the transmitted signal was optimized. R d The beam pattern after .

[0032] Fig.15 The following are the curves showing the changes in received power for different schemes under two routes in the test experiment.

[0033] Fig.16 In order to test the HT-6DMA and HT-6DMA with PA schemes in different d min Comparison of the minimum received power under the conditions and the additional gain brought by the local rotation of the array. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0035] Example 1 like Figure 1 As shown in the figure, the HT-6DMA (hierarchically adjustable six-dimensional movable antenna) base station is equipped with multiple uniform planar arrays (UPAs), each antenna array consists of multiple antenna units, and all antenna arrays can be moved within a radius of R The HT-6DMA base station not only adjusts the position of each antenna array on the surface of the spherical space, but also supports adjusting the orientation of each antenna array so that the antenna array can provide more accurate communication coverage for the target user hotspot area in the corresponding direction. The function of the spatial optimization method of the multi-antenna array in the HT-6DMA base station provided in this embodiment is to optimize and adjust the spatial state (including position and orientation) of each antenna array in the HT-6DMA base station in different scenarios in combination with the corresponding scenario requirements.

[0036] For the corresponding problem, existing solutions usually use a six-dimensional feature vector To characterize the spatial state of each antenna array in the base station. Among them, the six-dimensional feature vector ( x , y , z ) represents the three-dimensional coordinates of the center of each antenna array in the same Cartesian coordinate system; and , , Then they represent the deflection angles of the antenna array relative to the X-axis, Y-axis and Z-axis in the global coordinate system. Based on the above base station spatial model represented by the six-dimensional eigenvector, it can be seen that the existing scheme needs to combine the optimization objectives and use various optimization algorithms to solve the optimal solution of each individual six-dimensional eigenvector; and in the solution process, the feasible domain of the six-dimensional eigenvector of each antenna array needs to simultaneously satisfy the minimum distance constraint between antenna arrays, the first rotation constraint to avoid signal reflection, and the second rotation constraint to avoid signal blocking. Among them, the minimum distance constraint means that any two adjacent antenna arrays should avoid overlapping after the position is adjusted. The first selection constraint means that any two antenna arrays should avoid mutual signal reflection after the position and orientation are adjusted. The second selection constraint means that each antenna array should avoid signal blocking after the position and orientation are adjusted. It can be seen that the existing scheme needs to satisfy the above three constraints at the same time when solving the spatial state optimization problem of each antenna array in the base station. Considering that in the mathematical model corresponding to the optimization problem, the position and orientation of the antenna array contained in the decision variables are coupled with each other, each constraint is also relatively harsh, and the corresponding mathematical model is extremely difficult to solve.

[0037] In view of the difficulties existing in the existing solutions in the optimization of the spatial state of the antenna array, this embodiment proposes a new solution. In this solution, the base station spatial model that characterizes the spatial state of each antenna array contained therein is redesigned in combination with the inherent structural characteristics of the base station. The new base station (referred to as HT-6DMA) uses a two-dimensional position vector and a two-dimensional rotation vector to jointly characterize the spatial state of each antenna array, and supports hierarchical adjustment of the position and orientation of the antenna array. Combined with the newly designed base station spatial model, this embodiment further reconstructs the three types of constraints that need to be met in the antenna array optimization stage. The above improvements ultimately achieve dimensionality reduction of the high-dimensional features contained in the decision variables of the optimization problem; and simplify the constraints in the optimization problem, and expand the feasible domain of variables in the optimization process, reducing the difficulty of problem solving. The new base station spatial model and constraints realize the decoupling of the position characteristics and orientation characteristics of the antenna unit, and then support the hierarchical adjustment of the position and orientation of each antenna array by using an alternating optimization strategy, which can not only greatly improve the optimization efficiency, but also overcome the defect that the optimization result falls into a local optimal solution with poor performance.

[0038] In order to more clearly reflect the details and advantages of the solution of the present invention, the following chapters will describe in detail the spatial modeling process of the base station in the solution, the constraint reconstruction process, and the solution process of the optimization problem in multiple scenarios.

[0039] 1. HT-6DMA base station spatial model The HT-6DMA base station is equipped with Buniform planar antenna arrays, denoted as B = {1, 2, ..., B}, each array consists of N ≥1 antenna unit, N={1, 2, ..., N}. Figure 1 As shown, all antenna arrays can be R The spherical surface moves, and the spherical space is recorded as C .

[0040] On this basis, any antenna array b exist C The position of the surface is determined by its center relative to the origin of the global spherical coordinate system (SCS) (assumed to be the center of the base station, i.e. C The elevation angle of the center of and azimuth Represented as position vector : , in, , .

[0041] Arbitrary Antenna Array b The direction of the antenna is determined by the elevation angle of its external normal vector relative to the origin of its local coordinate system (SCS) (assumed to be the center of the antenna array). and azimuth Represented as the rotation vector : , in, , .

[0042] So far, this embodiment transforms the spatial state of each antenna array including position and orientation from the six-dimensional feature Dimensionality reduction to four-dimensional features .

[0043] In the new base station space model, it is assumed that b The first n The position of an antenna element in its local coordinate system and its position in its global coordinate system are recorded as and , , , then in the given and In the case of and The conversion equation between them is as follows: , in, Indicates that the bThe original local coordinates of the antenna array are transformed into The transformation matrix of the specified rotated local coordinates; Indicates that the b The local coordinates of the antenna array after rotation are further transformed into The transformation matrix of the specified global coordinates; in order to obtain the final global coordinates , that is, b The radius of the antenna array in the global coordinate system is R The position on the spherical surface C needs to add the corresponding distance offset , where the antenna array b The normalized distance from the center of the base station to the center of the base station Given by: .

[0044] In this embodiment, define represents the rotation matrix along the y-axis, Represents the rotation matrix along the z-axis. The expressions of the two are: .

[0045] Then, the transformation matrix and The expression is as follows: .

[0046] The following, combined Figure 2 The transformation process between the global coordinates and local coordinates of any antenna array, and how to generate the required transformation matrix and The process is described as follows: Before local rotation, the unit external normal vector of the original local coordinates of the antenna array is assumed to be n = [0,0,1] T Aligned with the z-axis of the local coordinate system, such as Figure 2 As shown in part (a) of . Then after the local rotation, n according to Rotate to ,in Represents the new unit external normal vector of the antenna array after local rotation. Among them, the overall local rotation consists of two separable rotations. The first is a rotation around the y-axis of the local coordinate system. , followed by a rotation around the z-axis of the local coordinate system , based on this, we can get Accordingly, after After the specified local rotation, Figure 2 Part (a) ofb The first n The original local coordinates of the antennas is converted to its rotated local coordinates, such as Figure 2 As shown in part (b) of , and is given by: , Then, in order to convert the local coordinate system Converted to the global coordinate system , further around the y-axis of the global coordinate system Rotate an angle , and then rotate around the z-axis of the global coordinate system ,like Figure 2 As shown in part (c) of In addition, It can be further expressed in the global coordinate system as ,like Figure 2 As shown in part (c) of Adds a distance offset after the specified global rotation The global coordinates can be obtained .

[0047] Based on the above description of the base station space model of the present invention, it can be seen that, assuming that in a R =1, there is an antenna array whose center is located at , and the antenna's external normal vector is perpendicular to the spherical surface. Then in HT-6DMA, the spatial state of the antenna array can be expressed as , , while in the original 6DMA, it needs to be expressed as: , .

[0048] It can be seen that the solution provided in this embodiment reduces the dimension of the base station space model that characterizes the spatial state of the antenna array from 6 dimensions to 4 dimensions. This lays a foundation for this embodiment to ultimately achieve more efficient optimization of the spatial state of the antenna array.

[0049] 2. Constraints for Antenna Array Spatial Optimization As mentioned above, during the spatial optimization of the antenna array, the position and orientation adjustments need to satisfy three types of constraints, namely, the minimum distance constraint, the first rotation constraint to avoid signal reflection, and the second rotation constraint to avoid signal blocking.

[0050] 2.1 Minimum distance constraint In the existing scheme, the minimum distance constraint requires that the distance between any two antenna arrays in the base station should be far enough to avoid overlap; that is, the distance between the two centers should be greater than a preset minimum safety distance. d min In the existing scheme, two antenna arrays i and j The center distance is the Euclidean distance of the three-dimensional coordinates of the two. In this embodiment, based on the improved base station space model, the constraint condition can be expressed by the following formula: , in, and Respectively i and j The normalized distance from the center of the antenna array to the center of the base station; R Indicates the radius of the spherical area where the antenna array position can be adjusted; d min Indicates the preset minimum safe distance between any two antenna array centers; Operator representing the Euclidean norm of a vector.

[0051] 2.2. Rotation Constraints to Avoid Signal Reflection During the process of adjusting the position and orientation of the antenna array, the rotation constraint of avoiding mutual signal reflection between any two arrays must be satisfied. Obviously, in the existing scheme, the six-dimensional eigenvector of each antenna array needs to be used to make a comprehensive judgment on whether this constraint is satisfied. In this embodiment, this rotation constraint is reconstructed as: , in, Indicates i The unit external normal vector of the antenna array relative to the global coordinate system satisfies: , in, n is the initial unit external normal vector of the antenna array relative to the local coordinate system, n =[0,0,1] T , and for all antenna arrays n Therefore, this constraint can be further simplified as follows: , in, express The third column vector of .

[0052] 2.3. Rotation Constraints to Avoid Signal Blockage When rotating and adjusting the orientation of each antenna array, it is necessary not only to avoid reflection of the transmitted signal between the antenna arrays, but also to avoid blocking of the transmitted signal of each antenna array itself. Specifically, this constraint means that the orientation of each antenna array cannot point to the center of the base station.

[0053] In the existing scheme, in order to meet the constraints, it is necessary to comprehensively judge the orientation of the antenna array by combining the three-dimensional coordinates of each antenna array and the deflection angles relative to each axis of the Cartesian coordinate system, and set the corresponding combination criteria as follows: , In the new scheme, the rotation vector in the new HT-6DMA base station spatial model is taken into account In , that is, the feasible domain of the rotation vector naturally excludes the possibility that the antenna array is facing one side of the base station center. Therefore, the rotation constraint to avoid signal blocking is automatically and implicitly satisfied in the new model without additional consideration.

[0054] In addition, in the proposed HT-6DMA base station spatial model, the position and rotation of the antenna array are hierarchically adjustable, that is, the position of each array on the sphere can be adjusted first according to the actual minimum distance constraint without any local rotation, and then the orientation of each antenna array is designed according to the rotation constraint to avoid signal reflection. This hierarchical adjustable structure can greatly simplify the algorithm design for joint optimization of array position and rotation, while achieving better performance than the existing 6DMA scheme.

[0055] 3. Uplink Communication In the uplink communication scenario, it is assumed that the K users communicate with the base station at the same time; each user is equipped with an omnidirectional fixed position antenna (FPA). and Represent the position vector and rotation vector of the stack respectively.

[0056] 3.1 3D Steering Vector Let K = {1, 2, ..., K} represent the user set, and Respectively represent k The elevation and azimuth of each user relative to the center of the HT-6DMA base station in the global coordinate system, k ∈K; then its unit pointing vector in the global Cartesian coordinate system is given by The following formula gives: , Further combining the conversion relationship between global coordinates and local coordinates, we can know that b The antenna array corresponds to thek The steering vector of each user for: , in, Indicates the carrier wavelength.

[0057] 3.2 Effective Antenna Gain Order k Users relative to b The elevation angle and azimuth angle of the antenna array center in its local spherical coordinate system are expressed as and ,in Indicates that the unit pointing vector is b 3D coordinates in the local Cartesian coordinate system of the array. k Users relative to b The effective antenna gain (on a linear scale) of an array is given by: , in, Indicates that each antenna array is and Directional gain in direction.

[0058] 3.3 Effective Channel Assuming that the signal between each user and the base station is a line-of-sight (LoS) channel, k The channel between the user and the base station It can be expressed as: , In the above formula, is the path gain.

[0059] Base station from all K The signal received by the user y It can be expressed as: , in, z It means that the mean is zero and the variance is Circular symmetric complex Gaussian noise vector of ; p Indicates the transmit power; x Indicates all K The vector corresponding to the signal sent by the user; x 1~ x K Respectively No. 1~ K The signal sent by the antenna array.

[0060] Maximum achievable aggregate rate for all users in uplink communications with perfect signal state information (CSI) at the base station, using Gaussian signals, and using multi-user joint decoding based on MMSE-SIC receivers C ( t , u ) is given by: ; in, I NB The dimension is NB To facilitate the solution, this embodiment uses the standard Monte Carlo method to approximate the average total rate of the user, and the average total rate obtained is Approximately: , In the above formula, Represents a given channel sample Time s The total rate achieved; S is based on H Total number of signal samples for distribution or user space status.

[0061] 4. Air Route Perception In the space route perception scenario, the HT-6DMA base station is designed to improve the perception performance of the route in its far field area. Assume that the base station is located at the origin of the Cartesian coordinate system, and there is a straight route segment above the base station, starting from , the end point is . Then the elevation angles and azimuth angles of the starting point and the end point relative to the center of the base station are: , Each point on the route segment It can be expressed as: , in, , since the flight path is located in the far field area of ​​the base station, a specific point on a given flight path , the corresponding steering vector of each antenna array is related only to its angular direction relative to the base station.

[0062] like Figure 3 As shown, each point in the route in the three-dimensional Cartesian coordinate system After conversion to two-dimensional coordinates represented by elevation and azimuth relative to the center of the base station, it can be expressed as : , in, , .

[0063] Due to the direction The corresponding pointing vector for: , No. b The steering vector corresponding to the antenna array is:

[0064] make x r represents the transmitted perceptual signal, whose mean is zero and the covariance matrix R d is full rank, that is:

[0065] Assume that the maximum transmit power budget of the HT-6DMA base station is P 0 ; then transmit the perception signal x r It must be designed under the total power constraint, which is: .

[0066] Since the airway is usually located at a high position above the ground, it is assumed that the signal between the base station and all points on the route segment is a line-of-sight signal. In order to enhance the performance of various sensing tasks along the route, the received signal power distribution along the route is adopted. As an evaluation index of route sensing performance, its expression is: , in, Indicates any point on the route interval The expression of the sensing channel between the base station is: , In the above formula, represents a constant related to the reference channel power; represents the path loss index in the route sensing scenario. Here it is assumed This is the same for all antenna arrays of a base station, since in practice the (shortest) distance between a base station and a route is much larger than the radius of the spherical space in which the antenna array can move. R .

[0067] 5. Position and Rotation Optimization in Uplink Communication Scenario In the uplink communication scenario, the optimization goal is to maximize the average total rate by jointly optimizing the positions and orientations of all antenna arrays while satisfying the constraints of minimizing distance and avoiding rotation constraints of signal reflection. Therefore, the optimization problem can be expressed as: , It should be emphasized that: by comparing the optimization problem statements of the 6DMA model in the existing solution and the HT-6DMA model in this case, it can be found that the rotation vector in the HT-6DMA model u It is defined relative to the local spherical coordinate system of each antenna array, while the rotation variables in the existing 6DMA model are defined relative to the global Cartesian coordinate system. This not only allows the present invention to essentially eliminate the signal blocking constraint, but also realizes the position vector in the optimization process. t and the rotation vector u decoupling.

[0068] Specifically, when solving the above optimization problem, the present invention can first adjust the position vector while satisfying only the minimum distance constraint. t , and then adjust the local rotation variable by only satisfying the rotation constraint to avoid signal reflection u Secondly, the position vector in the optimization problem of the new HT-6DMA model t and the rotation vector u The HT-6DMA model reduces the number of variables in the overall representation of each array from 6 to 4.

[0069] In addition, the moving area C of the antenna array is a cube in the existing 6DMA, but a sphere in HT-6DMA. Although C is non-convex in HT-6DMA, the algorithm design can be simplified by taking advantage of the fact that the centers of all arrays are at the same distance from the center of the base station. These differences pave the way for designing simpler and more efficient algorithms to solve the corresponding problems.

[0070] 5.1. Fixation u hour t Optimization First, all antenna arrays are rotated by the vector u Fixed to: In this case, on the left side of the constraint, we get: , Then, according to the Cauchy-Schwarz inequality, we can get: , Therefore, when optimizing the position vector, the rotation constraint to avoid signal reflection is always satisfied and can be removed. The optimization problem at this stage is simplified to:

[0071] So far, this embodiment obtains the optimal position vector by alternating optimization, and the position vectors of the remaining antenna arrays are given in each iteration. To optimize the position vector of the currently selected antenna array Since the minimum distance constraint is imposed on On, and and There is a one-to-one mapping relationship between them. In this example, we choose to optimize To facilitate calculation. The optimization problem at this stage is equivalent to:

[0072] in, To ensure has unit length. In order to reduce the difficulty of solving, this embodiment further Relaxation is: . And for , by It is expressed as a unit-length feasible solution after iteration t-1, and then the constraint is linearized and solved by applying the successive convex approximation method. At this point, the optimization problem is further transformed into:

[0073] At a given iteration t -1 hour Finally, in the inner layer, the Frank-Wolfe (conditional gradient descent) algorithm is applied to obtain Finally, Normalize to Iteration t At the same time, in the outer layer, alternate optimization , until the target value converges or reaches the predefined maximum number of outer layer iterations. Finally, after alternating optimization of the inner and outer layers, the final optimal position vector can be obtained, denoted as .

[0074] 5.2. In the given Next Optimization u This embodiment has been determined Based on this, we continue to optimize the orientation of each antenna array (i.e. the rotation vector u ); The optimization problem at this stage can be expressed as:

[0075] Similar to the previous step, this step can also continue to use a similar alternating optimization strategy to solve the optimization problem. Based on this, we get the final optimal rotation vector, recorded as .

[0076] 6. Position and rotation optimization in air route perception In the space route perception scenario, the optimization goal is to achieve joint optimization. t , u and the emission covariance matrix R d , on the basis of satisfying the aforementioned minimum distance constraint and rotation constraint, further satisfying the total power constraint of the transmitted signal , thereby maximizing the minimum received power on the spatial path segment. The general statement of this optimization problem is:

[0077] Note that this optimization problem is highly non-convex. To solve it, first fix R d and optimizes all antenna arrays t and u , and then based on the optimized t and u optimization R d Specifically, based on the unique advantages of the HT-6DMA model, the optimization in this scenario t and u The overall process is the same as that in the uplink communication link scenario. The only difference is that the objective function in the current scenario is not differentiable, that is: , To solve this problem, the line of this embodiment is fixed R d Optimize the case t and u , and convert the non-differentiable objective function into a differentiable one. Then based on the optimized t and u optimization R d .

[0078] 6.1. Fixation R d Time Optimization t and u First, the emission covariance matrix R d Fixed to , which satisfies the total power constraint, and then starts the optimization t and u , the optimization problem at this time is:

[0079] The objective function of the above formula is not differentiable. In order to solve this problem, we first discretize the continuous interval [0,1] into an interval containing D A set of equally spaced discrete points ,Right now: .

[0080] when D When it is large enough, there are:

[0081] On this basis, the entropy regularization method is applied to further replace the function approximate : , in, is a hyperparameter used to adjust the accuracy of the approximation. It is easy to prove that , and when hour, That is, there is no need to directly maximize , but rather maximizes the alternative function , it is The lower bound of When is larger, the gap between the two can be ignored. Through this approximation, the objective function becomes differentiable, and the optimization problem is transformed into:

[0082] This optimization problem can be solved by the same alternating optimization strategy as before, just replace the objective function with Specifically, first fix Alternate optimization in the case of t . After optimization, we get back, u It is also optimized in an alternating manner, and finally the optimized .

[0083] 6.1. Fixation and hour R d Optimization After obtaining all antenna arrays and After that, continue to optimize the above optimization problem R d The corresponding problem statement is , in, is introduced as an auxiliary variable. Here, for the convenience of optimization, the discrete set given in the previous article is also used Instead of the continuous interval [0,1]. Note that this optimization problem is a standard separable semidefinite programming (SSDP), which can be optimally solved by standard numerical solvers such as CVX.

[0084] Example 2 On the basis of Embodiments 1 and 2, this embodiment further provides a computer device; the computer device includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the spatial optimization method of the multi-antenna array in the HT-6DMA base station in Embodiment 1 is implemented, and then the position and orientation of all antenna arrays in the HT-6DMA base station are jointly optimized according to the scene requirements.

[0085] In practical applications, the computer device may be an embedded device that can be integrated into the base station. It may also be a computing device independent of the base station device. For example, various intelligent terminals capable of executing programs, tablet computers, laptop computers, desktop computers, rack servers, blade servers, tower servers or cabinet servers (including independent servers, or server clusters consisting of multiple servers), etc.

[0086] In practical applications, the spatial optimization method for multiple antenna arrays in an HT-6DMA base station provided in Example 1 may also exist in the form of a computer program product or storage medium storing a computer program. When the computer program is executed by a processor, the spatial optimization method for multiple antenna arrays in an HT-6DMA base station in Example 1 is implemented.

[0087] Example 3 On the basis of Examples 1 and 2, this embodiment further provides an HT-6DMA base station, which includes: multiple antenna arrays, an actuator and a controller. Among them, each antenna array is distributed on the surface of a spherical space, and each antenna array includes multiple antenna units. The actuator includes a first actuator unit and a second actuator unit, the first actuator unit is used to move and adjust the position of each antenna array on the surface of the spherical space; the second actuator unit is used to rotate and adjust the orientation of each antenna array at its respective position. A computer program product as in Example 2 is deployed in the controller; the controller runs the computer program product according to the needs of the scene, and then, in combination with the optimal position and optimal orientation after joint optimization generated by the computer program product, issues instructions to the actuator, thereby adaptively adjusting each antenna array to the optimal spatial state.

[0088] That is, the HT-6DMA base station provided in this embodiment has the technical effect of adaptively adjusting the spatial state of its own antenna array according to demand.

[0089] Verification experiment In order to verify the performance of the spatial optimization method of the multi-antenna array in the HT-6DMA base station provided by the present invention, the technicians conducted simulation tests in combination with two typical scenarios of uplink communication and air route perception.

[0090] 1. Simulation Conditions and Control Group In this experiment, the antenna array in each base station is considered as a uniform planar array with half-wavelength antenna spacing and a sphere C The radius is set to R =1 meter. Directional antenna gain (in dB) is given by: , In the above formula, and are the 3-dB beamwidths in the horizontal and vertical screens, respectively, which are set to 65° in this experiment. G s and G v Respectively represent the front-to-back ratio and sidelobe level limit. G s = G v =30dB. G max is the maximum directional gain of each antenna in the main lobe direction. G max Set to 8dBi. Accordingly, the horizontal radiation pattern and 3D directional gain Divided into Figure 4 Part (a) and Figure 4 As shown in part (b) of .

[0091] This experiment also compared the scheme of the present invention (referred to as HT-6DMA) with the following multiple control group schemes: 1. FPA: The base station evenly places three sector antenna arrays on the equator of C. Each antenna array has (NB / 3) antennas. The direction of each antenna array is fixed to .

[0092] 2. 6DMA: In this scheme, B, N and C are set to be the same as those of HT-6DMA base station. Antenna array position and rotation are jointly optimized by the algorithm proposed in the previous existing scheme.

[0093] 3. HT-6DMA with position adjustment only (denoted as HT-6DMAwith PA): This scheme only optimizes the array position through the HT-6DMA algorithm without any local rotation.

[0094] 4. HT-6DMA with rotation adjustment only (denoted as HT-6DMAwith RA): The antenna array in this scheme is evenly arranged on the surface of the sphere, and the orientation of the antenna array is optimized only by the HT-6DMA algorithm.

[0095] 2. Performance comparison in uplink communication scenarios In the experiment, the simulation results of each scheme in the uplink communication scenario are first tested. In the experiment, a non-homogeneous Poisson process is used to model the spatial state of the user. Assume that all users are distributed in the 3D coverage area represented by A, which is a 3D spherical ring with a radial distance from 50 meters to 120 meters from the center of the HT-6DMA base station. In A, there are three hot spots A1, A1 and A3, which are spheres with centers at (30, -60, -50) meters, (-40, 0, 60) meters and (0, 100, 20) meters, respectively, with a radius of 15 meters. The area excluding these hot spots is represented as A0. The number of users in each channel implementation in each area is represented as K i ; i ∈{0,1,2,3}, they are all modeled as Poisson random variables. S = 100 Monte Carlo simulations, and set the average noise power is -50 dBm, carrier frequency f c =2.4GHz, the transmission power of each user p = 30 mW. Based on the corresponding user space distribution, this experiment compares the performance of the proposed HT-6DMA with the existing 6DMA and FPA.

[0096] During the experiment, the proposed HT-6DMA algorithm and the existing 6DMA algorithm were compared in ( B =16, N =4) when the convergence behavior is as follows Figure 5 The number of iterations in the figure corresponds to the number of updates of the position variables or rotation variables of each array. For example, B = 16 iterations corresponding to array global position update, followed by the existing 6DMA B= 16 array global direction updates. From the data in Observation 5, it can be seen that the HT-6DMA proposed in the present invention converges to a better local optimal solution in a smaller number of iterations, which is better than the existing 6DMA algorithm. The reasons for this advantage include the following two points: First, when optimizing the array position in the existing 6DMA algorithm, the orientation of the array relative to the global coordinate system remains unchanged, while in the HT-6DMA algorithm, although there is no rotation of the array relative to the local coordinate system when optimizing the array position, the orientation of the array relative to the global coordinate system will change as the array moves on C, thereby more effectively searching for the optimal array position / orientation. Secondly, compared with the existing 6DMA algorithm, the constraints for array position and array direction optimization in HT-6DMA are simpler, thereby obtaining a better solution at a faster convergence speed. During the experiment, the optimization results of different schemes were compared. Figure 6 As shown, Figure 6 Part (a) is the optimization result of the existing 6DMA scheme; Figure 6 Part (b) shows the optimization result of the HT-6DMA with PA solution; Figure 6 Part (c) shows the optimization result of HT-6DMA with RA scheme; Figure 6 Part (d) in the figure is the optimization result of the HT-6DMA solution. Combining the results in the figure, it can be seen that the optimization results of HT-6DMA with PA and HT-6DMA are slightly different, which reflects that the performance gap between the two is small.

[0097] This study further compared HT-6DMA with other control groups in ( B= 16, N= 4) and ( B= 4, N= 16) under the conditions of the total rate and algorithm execution time, the results are as follows Figure 7 Part (a) and Figure 7 As shown in part (b) of FIG. 1 , it can be clearly seen from the figure that both the existing 6DMA and the HT-6DMA of the present invention achieve better performance than FPA because they can better adapt to the changes in the user space state. Figure 7 It can also be observed from the data that when the total number of antennas is kept constant, the overall performance of all HT-6DMA schemes generally increases with B The increase is better because now there are more arrays to adjust, which makes the design more flexible. However, as B =16, more variables and constraints need to be designed, and the execution time of all designs becomes longer. In particular, the existing 6DMA algorithm B = 16 takes longer to find a solution, but its actual performance is even slightly lower than B = 4, which may be due toB When is large, it is easy to fall into a bad local optimal solution. In contrast, it is observed that the HT-6DMA algorithm achieves better performance than the existing 6DMA algorithm with less execution time in both settings. Finally, it can also be observed from the figure that even in ( B= 4, N= 16), the performance improvement of HT-6DMA mainly comes from position optimization rather than local rotation optimization.

[0098] Figure 8 Shown in B =16 and N =4, the total rate that can be achieved is proportional to the user sparseness ratio It can be observed in the figure that when When changing from 0 to 1, the performance of FPA improves, while the performance of HT-6DMA and existing 6DMA decreases. This is because the ability to align antenna resources with hot spots becomes less important when users are more sparsely distributed. For FPA, even if the positions and directions of its three sectors are fixed, when When is larger, more users can be properly served. In addition, it is observed that the performance gap between different schemes becomes negligible when the user distribution is more sparse, because At larger sizes, proper design of the array becomes less important due to the lack of user clusters.

[0099] Fig. 9 Demonstrating the minimum distance between different arrays of HT-6DMA and HT-6DMA with PA d min The performance difference between them reflects the gain of local rotation adjustment of the array in HT-6DMA. First, from Fig. 9 It can be seen that with d min As the value of increases, the overall rate performance of both designs decreases due to the stricter inter-array spacing requirement. In addition, it can be observed that when d min As increases, the performance gain from rotation adjustment also increases. This is because when d min When is larger, it becomes more difficult to position all arrays so that their default orientation can accurately point to the corresponding hotspot area.,Therefore, additional local orientation adjustment to make each array face the hotspot area can bring higher,performance gains.

[0100] 3. Performance comparison in aerial perception scenarios This experiment further tests the simulation results of each solution in an aerial perception scenario, which contains two intersecting routes. Fig.10 As shown in the figure, the starting coordinates of the first route are (-30, -50, 30) m, and the ending coordinates are (60, 40, 30) m. The starting coordinates of the second route are (50, -40, 30) m, and the ending coordinates are (5, 40, 30) m. Fig.11 Demonstrates different HT-6DMA algorithms in proxy functions f a The convergence behavior of the two aspects can be seen in the figure: f a Most of the improvements are achieved by adjusting the position of the array on the sphere rather than by local rotation adjustments, which is similar to the uplink communication scenario.

[0101] Fig.12 The optimized array positions and rotations of different HT-6DMA schemes are shown as viewed from above. Fig.12 (c) in the HT-6DMA with PA ( Fig.12 The antenna arrays in (b) are all oriented toward the two airspace route segments, showing the effectiveness of the proposed algorithm based on the HT-6DMA model. In addition, compared with HT-6DMA with PA, HT-6DMA shows that the array further maximizes the minimum received power along the airspace route segment through additional rotation. The figure also shows that although some arrays are optimized to face the airspace route segment, due to HT-6DMA with RA ( Fig.12 Part (a) in the figure cannot achieve position adjustment, so the performance is still relatively poor.

[0102] In order to compare the perceived performance of different schemes, such as Fig.13 As shown, this experiment shows the FPA in a two-dimensional angle diagram ( Fig.13 (a) in the HT-6DMA with RA ( Fig.13 (b) in the HT-6DMA withPA ( Fig.13 The received power distribution in part (c) of Figure 1. In this experiment, the transmit covariance matrix R d According to the results, HT-6DMA achieves the highest minimum perceived signal power on the two airspace routes, which proves the effectiveness of the scheme proposed in the present invention. In addition, if the proposed HT-6DMA algorithm is further optimized R d , received power distribution Fig.14 As shown in the figure, it can be seen that the signal power is more concentrated on the airspace route, thereby improving the perception performance.

[0103] Under the same setting, this experiment further Fig.15 The received powers of different schemes along the first and second airspace routes are shown in Figure 1, and their minimum received powers are listed in Table 1. It can be observed that, similar to the uplink communication scenario, the performance gain mainly comes from the array position optimization rather than the array local rotation optimization, because Table 1 shows that the performance gap between the HT-6DMA and HT-6DMA with PA schemes is very small.

[0104] Table 1: Comparison of minimum received power of different schemes

[0105] at last, Fig.16 Demonstrated the differences between HT-6DMA and HT-6DMA with PA d min The performance comparison of the minimum received signal power under HT-6DMA is shown in Figure 2, where their performance difference reflects the perceived gain of the local rotation of the array in HT-6DMA. To simplify the analysis, only the second airspace route is considered here. Similar to the uplink communication scenario, Fig.16 It can be seen that with d min As increases, the perceptual performance of the two design schemes decreases, while the performance gain brought by the rotation adjustment increases.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for spatial optimization of a multi-antenna array in an HT-6DMA base station, characterized in that: It includes: In the global spherical coordinate system with the center of the base station as the origin, the elevation angle of the antenna array is used and azimuth Define the position vector representing its position t : ; In the local spherical coordinate system with the center of the antenna array as the origin, the elevation angle of the antenna array's external normal vector relative to itself in the local spherical coordinate system is used and azimuth Define a rotation vector representing its orientation u ; ; based on t and u Construct a spatial model of the HT-6DMA base station that can characterize the position and orientation of any antenna array in the base station; Based on the new base station spatial model, the minimum distance constraint required for spatial optimization of the antenna array and the rotation constraint to avoid signal reflection are reconstructed; t and u As decision variables, a mathematical model representing the spatial optimization problem of antenna array is generated in combination with the corresponding optimization objectives and scenario constraints; Solve the mathematical model in layers: First, fix the local orientation of each antenna array to a spherical surface with an external normal vector perpendicular to the global coordinate system, and then solve the mathematical model through an alternating optimization strategy to obtain the optimal position of each antenna array that satisfies the minimum distance constraint. ; 2. To be in the best position The orientation of each antenna array on the system is rotated and adjusted, and the optimal orientation that satisfies the rotation constraint to avoid signal reflection is further solved through the alternating optimization strategy. .

2. The method for spatial optimization of a multi-antenna array in a HT-6DMA base station according to claim 1, characterized in that: The expression of the minimum distance constraint is as follows: ， in, and Respectively i and j The normalized distance from the center of the antenna array to the center of the HT-6DMA base station; R Indicates the radius of the spherical area where the antenna array position can be adjusted; d min represents the preset minimum safe distance between the centers of any two antenna arrays; B represents the set of all antenna arrays in the base station, B={1, 2, …, B}.

3. The spatial optimization method of a multi-antenna array in a HT-6DMA base station according to claim 2, characterized in that: The expression of the rotation constraint to avoid signal reflection is as follows: ， in, Indicates that the i The original local coordinates of the antenna array are transformed into The transformation matrix of the specified rotated local coordinates; express The third column vector of ; Indicates that the i The local coordinates of the antenna array after rotation are further transformed into Specifies the transformation matrix for global coordinates.

4. The method for spatial optimization of a multi-antenna array in a HT-6DMA base station according to claim 3, characterized in that: The process of the alternating optimization strategy is as follows: Firstly, the target space of the optimization result is convexly approximated using the successive convex approximation method; Then, a double-loop iterative algorithm with inner and outer nesting is designed. In the inner loop, the conditional gradient descent method is used to iterate one of the features related to the decision variable, while keeping the other features of the decision variable unchanged, until the target value of the inner loop converges or reaches the preset maximum iteration round; in the outer loop, each feature in the decision variable is alternately iterated and optimized until the target value of the outer loop converges or reaches the preset maximum iteration round; Finally, the desired optimal position or optimal orientation is generated according to the optimization results obtained in the inner and outer loops.

5. The method for spatial optimization of a multi-antenna array in a HT-6DMA base station as claimed in claim 3, characterized in that: In the base station space model, i The first n The position of the antenna element in its local coordinate system and its position in global coordinates The conversion equation between them is as follows: ， In the above formula, N represents the set of antenna elements in each antenna array, N={1, 2, ..., N}; the expression of the normalized distance from the center of any antenna array to the center of the base station is as follows: 。 6. The method for spatial optimization of a multi-antenna array in a HT-6DMA base station as claimed in claim 3, characterized in that: The transformation matrix and The expression is as follows: ， in, The rotation matrix represents the rotation along the y-axis, and its expression is given by the following formula: ， The rotation matrix represents the rotation along the z-axis, and its expression is given by the following formula: 。 7. The method for spatial optimization of a multi-antenna array in a HT-6DMA base station as claimed in claim 3, characterized in that: It is used to jointly optimize the position and orientation of all antenna arrays in the HT-6DMA base station in the uplink communication scenario, thereby maximizing the long-term average communication sum total rate of the user; The spatial optimization method is also used in air route perception to jointly optimize the positions and orientations of all antenna arrays in the HT-6DMA base station and the covariance matrix of the transmitted signal, thereby maximizing the minimum received perception signal power on the flight segment.

8. A computer device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: when the processor executes the computer program, it implements the spatial optimization method of a multi-antenna array in an HT-6DMA base station as described in any one of claims 1 to 7, thereby realizing joint optimization of the positions and orientations of all antenna arrays in the 6DMA base station according to scenario requirements.

9. A computer program product comprising a computer program, characterized in that: When the computer program is executed by a processor, the spatial optimization method of a multi-antenna array in an HT-6DMA base station according to any one of claims 1 to 7 is implemented.

10. An HT-6DMA base station, characterized in that: It includes: A plurality of antenna arrays are distributed on a surface of a spherical space, each antenna array comprising a plurality of directional antenna units; The actuator comprises a first actuator unit and a second actuator unit, wherein the first actuator unit is used to move and adjust the position of each antenna array on the surface of the spherical space; and the second actuator unit is used to further rotate and adjust the orientation of each antenna array at its respective position; A controller in which the computer program product as claimed in claim 9 is deployed; the controller runs the computer program product according to the scene requirements, and then issues instructions to the actuator in combination with the jointly optimized optimal position and optimal orientation generated by the computer program product, thereby adaptively adjusting each antenna array to the optimal spatial state.

Images