Series-parallel mixed true time delay architecture of near-field uniform circular antenna array and beam forming method

Abstract

The invention discloses a series-parallel mixed true time delay architecture and beam forming method for a near-field uniform circular antenna array, and the method comprises the steps: 1, building a spherical coordinate model of the uniform circular antenna array in a near-field downlink broadband transmission scene, and determining a channel expression from each unit of the antenna array to a user; 2, initializing a phase shift matrix of the phase shifter and a time delay value under the condition that a real time delay unit is parallel, and obtaining an optimal initial unit index under an ideal circuit; and step 3, according to the serial-parallel mixed true time delay architecture provided by the invention, determining a serial starting unit, initializing a serial true time delay unit and compensating time delay values of the true time delay unit. And 4, carrying out iterative optimization on the phase shift matrix, the time delay value of the true time delay unit and the compensation time delay value by adopting an alternating optimization algorithm. According to the invention, the maximum time delay demand of the real time delay unit can be reduced, so that the whole architecture can still obtain ideal beam focusing and signal transmission effects under the condition that the real time delay unit is limited.

Description

Technical Field

[0001] The present invention belongs to the field of wireless communication technologies, and particularly relates to a hybrid real-time architecture and beamforming method for a near-field uniform circular antenna array. Background Art

[0002] In traditional far-field communication modeling, the plane wave assumption generally applies to most scenarios, especially in the low-frequency band and under far-field propagation conditions. However, as wireless communication systems develop towards higher frequency bands, the applicability of the far-field model is gradually limited. Future 6G communication will further expand towards millimeter wave and higher frequency bands to meet the requirements of ultra-high speed, ultra-low latency, and ultra-large connection. In the near-field propagation environment of millimeter wave communication, since the distance between the user and the antenna array is short, the electromagnetic wave exhibits spherical wave characteristics, resulting in the inapplicability of the traditional far-field plane wave assumption. In order to accurately describe the beamforming and channel characteristics, a near-field channel modeling method based on the spherical wave model is required to depict the true signal propagation paths from different antenna elements to the user. Compared with a uniform linear array, a uniform circular array can better adapt to the near-field environment and provide more uniform beam coverage due to its omnidirectional coverage ability, making it an ideal antenna architecture choice in near-field millimeter wave communication.

[0003] In addition, in a broadband communication environment, the system usually adopts multi-carrier transmission technologies such as orthogonal frequency division multiplexing. However, due to the large frequency span of broadband signals, there will be a certain degree of deviation in the propagation paths of different sub-carriers, and this phenomenon is called the spatial broadband effect. In the case of millimeter waves, this effect is particularly obvious, resulting in inconsistent beam directions of different sub-carriers, and further causing the problem of beam defocusing. This means that the signal energy of some sub-carriers cannot be effectively focused on the target user, resulting in a decrease in the signal gain of the system, an increase in the bit error rate, and a reduction in communication quality.

[0004] After a literature search of the prior art, it was found that Akram Najjar et al. proposed a true time delay unit-phase shifter structure based on the time delay difference using a selection switch in the paper "Hybrid Delay-Phase Precoding in Wideband UM-MIMO Systems Under True Time Delay and Phase Shifter Hardware Limitations" published in 《IEEE Transactions on Wireless Communications, vol.23, no.7, July.2024. (IEEE Transactions on Wireless Communications, Volume 23, Issue 7, July 2024)》. This structure can significantly reduce the hardware requirements of true time delay units. However, the paper considered a uniform linear array in the far-field case and did not consider the beam defocusing phenomenon in the case of a near-field uniform circular array. Another search found that Yunhui Guo et al. published a paper titled "Wideband Beamforming for Near-Field Communications with Circular Arrays" in 《IEEE Transactions on Wireless Communications, vol.23, no.12, pp.19065-19082, Dec.2024. (IEEE Transactions on Wireless Communications, Volume 23, Issue 12, Pages 19065-19082, December 2024)》. In this paper, it was proven that the beam focus points in a wideband uniform circular array only exist at specific frequencies, resulting in a large beamforming loss. And a parallel true time delay unit-phase shifter architecture was proposed to suppress the beam defocusing phenomenon, but the maximum time delay value required by its true time delay unit is extremely large. Currently, in the millimeter-wave band, the maximum time delay that a single MEMS and CMOS true time delay unit can achieve is less than 300 ps, and the present invention can well solve the above problems. Summary of the Invention

[0005] The object of the present invention is to address the defects and deficiencies of the above prior art, and propose a serial-parallel hybrid true time delay architecture and beamforming method for a near-field uniform circular antenna array. This method makes full use of the symmetry of the uniform circular array and proposes a true time delay control architecture based on time delay difference with two groups in parallel and serial within the group to optimize the beam synthesis performance. While suppressing beam defocusing and ensuring the array gain, it effectively reduces the maximum time delay value required by the true time delay unit, so that the overall architecture can still obtain relatively ideal beam focusing and signal transmission effects under the limitation of the maximum time delay of the true time delay unit.

[0006] The technical solution adopted by the present invention to solve its technical problems is: a serial-parallel hybrid true time delay architecture and beamforming method for a near-field uniform circular antenna array, and the method includes the following steps:

[0007] Step 1: In the near-field downlink broadband transmission scenario, establish a spherical coordinate model of the uniform circular antenna array, and determine the channel expression of each unit of the antenna array to the user.

[0008] Step 2: Initialize the phase shift matrix of the phase shifter and the time delay value in the case of parallel true time delay units to obtain the optimal starting unit index under the ideal circuit.

[0009] Step 3: According to the serial-parallel hybrid true time delay architecture of the present invention, determine the starting unit of the serial connection, initialize the serial true time delay unit and the time delay value of the compensated true time delay unit, and calculate the cumulative time delay at each true time delay unit in the serial-parallel hybrid architecture.

[0010] Step 4: Use the alternating optimization algorithm to iteratively optimize the phase shift matrix, the time delay value of the true time delay unit, and the compensated time delay value to approximate the ideal beamforming effect, and design the digital beamforming matrix according to the equivalent channel.

[0011] Further, step 1 specifically includes:

[0012] In the near-field downlink broadband transmission scenario, the transmitting end is equipped with a uniform circular antenna array, and the orthogonal frequency division multiplexing technology of M subcarriers is used to serve K single-antenna users. The frequency f m of the m-th subcarrier is:

[0013]

[0014] where f c is the system center frequency, and B is the total system bandwidth.

[0015] The antenna array includes a total of N antenna elements, and the antenna element spacing d is set to half a wavelength. According to the geometric structure of the uniform circular array, the radius of the uniform circular array Near-field Rayleigh distance

[0016] Use the polar coordinate model to characterize the channel model between the antenna array elements and the user. Let the center of the uniform circle be the pole, and the position of the n-th transmitting antenna is expressed as p n =(R, ψ n , 0), and the position of the k-th user is represented by the polar coordinates u k =(r k , θ k , φ k ), r k represents the distance from the user to the center of the circle, θ k ∈[0, π] and φk ∈[0, 2π] represent the elevation angle and azimuth angle of the user to the center of the uniform circular array respectively.

[0017] The near - field channel \(h\) between user \(k\) and the \(n\)th antenna element at the \(m\)th sub - carrier m,k,n can be modeled as:

[0018]

[0019] where \(\beta\) m,k is the complex channel gain of the los path, satisfying represents the free - space gain, \(G\) r and \(G\) k represent the antenna gains at the transmitter and receiver respectively.

[0020] \(\alpha\) m \((r,\theta,\varphi)\) is the beam steering vector between the base station and user \(k\) at the \(m\)th sub - carrier:

[0021]

[0022] where \(c\) is the speed of light, represents the distance from the \(n\)th antenna element to the \(k\)th user.

[0023]

[0024] Furthermore, step 2 specifically includes:

[0025] Adopting a fully - connected architecture with true time - delay - phase shifters, \(N\) antenna elements are divided into \(Q\) sub - arrays, each sub - array has \(P\) antenna elements (\(P = N / Q\)). Assuming the number of RF chains is \(N\) RF , satisfying \(N\) RF \(=K\ll N\), each RF chain drives \(Q\) true time - delay units, and each true time - delay unit is connected to \(P\) phase shifters, thus forming an analog beamformer architecture based on TTD - PS:

[0026] \(W\) m \(=[W\) 1,m ,…, \(W\) k,m (5)

[0027]

[0028] where is the analog precoding matrix of the \(k\)th user's corresponding RF chain at the \(m\)th sub - carrier, is the phase - shift matrix of the phase shifters of the \(k\)th user's corresponding RF chain, is the phase - shift matrix implemented by the true time - delay unit of the \(k\)th user's corresponding RF chain at the \(m\)th sub - carrier, is the time delay value of the true time delay unit corresponding to the k-th user's radio frequency chain.

[0029] At the q-th subarray, the phase shift value that the analog beamformer needs to implement is:

[0030]

[0031] where represents the azimuth angle of the center point of the q-th subarray.

[0032] In Equation (7), the frequency-independent part is implemented by the phase shifter, and the frequency-dependent part is implemented by the true time delay unit.

[0033] The phase shift matrix [W k PS q for the k-th user at the phase shifter of the q-th subarray is:

[0034]

[0035] where a c (r k , θ k , φ k ) is the beam steering vector between the base station and user k at the center frequency, f c is the system center frequency, c is the speed of light, r k represents the distance from the user to the center of the circle in polar coordinates, θ k ∈[0, π] and φ k ∈[0, 2π] respectively represent the elevation angle and azimuth angle of user k to the center of the uniform circular array, R is the radius of the uniform circular antenna array, represents the azimuth angle from the center point of the q-th subarray to the center of the circle, N is the number of antenna elements, Q is the number of true time delay units, and P is the number of phase shifters connected to each true time delay unit.

[0036] The parallel delay t (q,k) of the k-th user at the true time delay unit of the q-th subarray is:

[0037]

[0038] According to Equation (9), it can be obtained that t (q,k) is positively correlated with the distance from the center point of the q-th subarray of the antenna to the user. Based on the symmetry of the uniform circular array, the Q true time delay units are evenly divided into two groups. Serial connection is used within the group, and parallel connection is used between the two groups.

[0039] Let the index of the true time delay unit with the lowest time delay value of the k-th user be Define the index set of the subarray is a forward serial set, and the starting index of the forward serial is while is a backward serial set, and the starting index of the backward serial is

[0040] In the ideal circuit case, the starting position of the serial can be any true delay unit. The starting index of the forward / backward serial is obtained according to the user's position, and a selection circuit is used to select the appropriate starting true delay unit, which can give full play to the advantage of the cumulative delay. and are the starting unit indexes in the ideal circuit case.

[0041] Furthermore, step 3 specifically includes:

[0042] may appear at any true delay unit. A Q-to-1 selection circuit is required between the RF chain and the true delay network to obtain At the same time, as the user's position changes, the index sets of the forward serial and the backward serial will also change frequently, resulting in a large number of dynamic switches required for the circuit. The change of the series-parallel structure may also lead to synchronization problems. Therefore, it is necessary to stipulate the starting position to reduce the circuit complexity.

[0043] Divide the Q true delay units into 4 groups according to the quadrant, and stipulate that the starting index point of the backward serial can only be selected from the last unit of each group Among them, the starting point of the corresponding forward serial is The starting point of the backward serial corresponding to the kth user is

[0044]

[0045] The corresponding forward serial index set and the backward serial index set are:

[0046]

[0047] The delay value corresponding to the qth serial true delay unit of the kth user

[0048]

[0049] Introduce compensation true delay units τ front and τ back , used to compensate for the delay error caused by q start ≠q min , and initialize

[0050] The cumulative delay obtained by the q-th true delay unit at this time can be accumulated according to the serial index set to obtain:

[0051]

[0052] where is the true delay unit index corresponding to the x-th element in the forward serial set of the k-th user, is the y-th element index in the backward serial set of the k-th user,

[0053] Furthermore, step 4 specifically includes:

[0054] With the goal of maximizing the array gain G, obtain the unconstrained optimal analog beamforming phase matrix

[0055]

[0056] When Q = N and P = 1, equation (15) takes the equal sign, and the unconstrained optimal analog beamforming phase matrix can be obtained

[0057] Model the analog beamforming precoder as an optimization problem:

[0058]

[0059] where W k PS is the phase shift matrix of the phase shifters of the radio frequency chain corresponding to the k-th user, is the phase shift matrix realized by the true delay unit of the radio frequency chain corresponding to the k-th user;

[0060] Use the alternating optimization algorithm to solve the optimization problem in equation (16), and alternately update the phase shift matrix, serial true delay unit, and compensation delay.

[0061] First, fix the compensation delay and the delay of the serial true delay unit, and update the phase shift matrix:

[0062]

[0063] where is the cumulative delay with compensation in equation (14), and f m is the frequency of the m-th subcarrier.

[0064] Secondly, fix the phase shift matrix and the compensation delay, and update the delays of the true delay units in the forward and backward serials:

[0065]

[0066] where In formula (13), The corresponding true time delay unit should also be set to 0 during the time delay update to avoid local optimal solutions.

[0067] The above problem is a simple one-dimensional search problem. By performing a one-dimensional search within [0, t max , the ideal value can be obtained, where t max is the maximum time delay value achievable by the true time delay unit.

[0068] Finally, fix the time delay of the serial true time delay unit and the phase shift matrix, and perform a one-dimensional search on the compensation time delay to obtain

[0069]

[0070] where is the sum of the front / back serial parts obtained by one-dimensional search accumulated according to the corresponding index set, and differs in that the compensation time delay is not added, and the search range of is related to the number Q of true time delay units and is

[0071] Alternately update the phase matrix, the serial time delay value, and the compensation time delay value until where γ is the convergence threshold.

[0072] Equivalent channel where H m is the near-field channel function, and the digital beamforming matrix can be obtained according to the zero-forcing algorithm

[0073] Beneficial effects:

[0074] 1. The present invention uses a serial-parallel hybrid architecture. By serially accumulating the time delay, the maximum time delay required by the true time delay unit is greatly reduced, thereby reducing the circuit complexity and the requirement for the driving clock frequency, achieving the goal of suppressing beam defocusing and ensuring the array gain, and improving the beamforming performance.

[0075] 2. The present invention enables the overall architecture to still obtain relatively ideal beam focusing and signal transmission effects under the limitation of the true time delay unit. Description of the drawings

[0076] Figure 1 is the flowchart of the hybrid beamforming method of the present invention.

[0077] Figure 2 (a) is the ideal circuit architecture diagram in step 2 of the present invention, Figure 2 (b) is the architecture diagram of fixing the starting position and adding the compensation time delay in step 3 of the present invention.

[0078] Figure 3 This is the variation diagram of the normalized array gain of the present invention with the azimuth angle of the user.

[0079] Figure 4 This is the single-user spectral efficiency diagram of the present invention under the condition that the maximum delay of the true delay unit is limited.

[0080] Figure 5 This is the variation diagram of the multi-user spectral efficiency of the present invention with the maximum delay of the true delay unit. Detailed implementation manners

[0081] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0082] As Figure 1 shown, the present invention provides a serial-parallel hybrid true delay architecture and beamforming method for a near-field uniform circular antenna array. The method includes the following steps:

[0083] Step 1: In the near-field downlink broadband transmission scenario, the transmitting end is equipped with a uniform circular antenna array. The orthogonal frequency division multiplexing technology of 10 subcarriers is used to serve 4 single-antenna users. The frequency f of the m-th subcarrier m is:

[0084]

[0085] where f c The system center frequency is 28 GHz, and the total system bandwidth B is 3 GHz.

[0086] The antenna array includes a total of 256 antenna elements. The antenna element spacing d is set to be half a wavelength. According to the geometric structure of the uniform circular array, the radius of the uniform circular array is 0.2183 meters, and the near-field Rayleigh distance is 35.57 meters.

[0087] The polar coordinate model is used to characterize the channel model between the antenna array elements and the user. Let the center of the uniform circle be the pole. The position of the n-th transmitting antenna is expressed as p n =(R, ψ n , 0). For the position of the k-th user, it is represented by the polar coordinates u k =(r k , θ k , φ k ). r k represents the distance from the user to the center of the circle, and θ k ∈[0, π] and φ k ∈[0, 2π] respectively represent the elevation angle and azimuth angle of the user to the center of the uniform circular array.

[0088] The near-field channel \(h\) between user \(k\) and the \(n\)th antenna element at the \(m\)th subcarrier m,k,n can be modeled as:

[0089]

[0090] where \(\beta\) m,k is the complex channel gain of the los path, satisfying represents the free-space gain, \(G\) r and \(G\) k represent the antenna gains at the transmitter and receiver, respectively.

[0091] \(\alpha\) m \((r,\theta,\varphi)\) is the beam steering vector between the base station and user \(k\) at the \(m\)th subcarrier:

[0092]

[0093] where \(c\) is the speed of light, represents the distance from the \(n\)th antenna element to the \(k\)th user.

[0094]

[0095] Step 2: Adopting a fully connected architecture with true time delay-phase shifters, 256 antenna elements are divided into \(Q = 16\) subarrays, each subarray having 16 antenna elements (\(P = N / Q\)). Assuming the number of RF chains is 4, satisfying \(N\) RF \(= K \ll N\), each RF chain drives 16 true time delay units, and each true time delay unit is connected to 16 phase shifters, thus forming an analog beamformer architecture based on TTD-PS:

[0096] \(W\) m \(= [W\) 1,m ,…,W\) k,m (5)

[0097]

[0098] where is the analog precoding matrix of the \(k\)th user's corresponding RF chain at the \(m\)th subcarrier, is the phase shift matrix of the phase shifters of the \(k\)th user's corresponding RF chain, is the phase shift matrix implemented by the true time delay unit of the \(k\)th user's corresponding RF chain at the \(m\)th subcarrier, is the time delay value of the true time delay unit of the \(k\)th user's corresponding RF chain.

[0099] At the \(q\)th subarray, the phase shift value that the analog beamformer needs to implement is:

[0100]

[0101] wherein represents the azimuth angle of the center point of the q-th subarray.

[0102] In Equation (7), the frequency-independent part is implemented by a phase shifter, and the frequency-dependent part is implemented by a true time delay unit.

[0103] The phase shift matrix [W k PS q of the k-th user at the phase shifter of the q-th subarray is:

[0104]

[0105] where a c (r k , θ k , φ k ) is the beam steering vector between the base station and user k at the center frequency, f c is the system center frequency, c is the speed of light, r k represents the distance from the user to the center of the circle in polar coordinates, θ k ∈ [0, π] and φ k ∈ [0, 2π] respectively represent the elevation angle and azimuth angle of user k to the center of the uniform circular array, R is the radius of the uniform circular antenna array, represents the azimuth angle from the center point of the q-th subarray to the center of the circle, N is the number of antenna elements, Q is the number of true time delay units, and P is the number of phase shifters connected to each true time delay unit.

[0106] The parallel delay t (q,k) of the k-th user at the true time delay unit of the q-th subarray is:

[0107]

[0108] According to Equation (9), it can be obtained that t (q,k) is positively correlated with the distance from the center point of the q-th subarray of the antenna to the user. Based on the symmetry of the uniform circular array, the 16 true time delay units are evenly divided into two groups. Serial connection is used within the group, and parallel connection is used between the two groups.

[0109] Let the index of the true time delay unit with the lowest delay value of the k-th user be Define the index set of the subarray as the forward serial set, and the starting index of the forward serial is while is the backward serial set, and the starting index of the backward serial is

[0110] In the ideal circuit scenario, the serial starting position can be any true delay unit. Based on the user's position, the starting indices of the forward / backward serials are obtained, and a selection circuit is used to select the appropriate starting true delay unit, which can fully utilize the advantage of the cumulative delay. and is the starting unit index in the ideal circuit scenario. As Figure 2 (a) shows, the RF chain can freely select any true delay unit as the starting unit through the selection circuit.

[0111] Step 3: It may occur at any true delay unit. A 1-of-16 selection circuit is required between the RF chain and the true delay network to obtain Meanwhile, as the user's position changes, the index sets of the forward and backward serials will also change frequently, resulting in the need for a large number of dynamic switches in the circuit. The change in the series-parallel structure may also lead to synchronization problems. Therefore, it is necessary to specify the starting position to reduce the circuit complexity.

[0112] The Q true delay units are divided into 4 groups according to the quadrants. As Figure 2 (b) shows, the true delay units within each group are serially fixed. It is stipulated that the starting index point of the backward serial can only be selected from the last unit of each group Among them, the starting point of the corresponding forward serial is The starting point of the backward serial corresponding to the kth user is

[0113]

[0114] The corresponding forward serial index set and the backward serial index set are:

[0115]

[0116] The delay value corresponding to the qth serial true delay unit of the kth user

[0117]

[0118] Compensation true delay units τ front and τ back are introduced before the selection circuit to compensate for the start delay error caused by q min ≠q

[0119] At this time, the cumulative delay obtained by the qth true delay unit can be accumulated according to the serial index set to obtain:

[0120]

[0121] where is the true delay unit index corresponding to the x-th element in the k-th user's forward serial set, is the index of the y-th element in the k-th user's backward serial set,

[0122] Step 4: With the goal of maximizing the array gain G, obtain the unconstrained optimal analog beamforming phase matrix

[0123]

[0124] When Q = 16 and P = 1, equation (15) takes the equality, and the unconstrained optimal analog beamforming phase matrix can be obtained

[0125] Model the analog beamforming precoder as an optimization problem:

[0126]

[0127] where W k PS is the phase shift matrix of the phase shifters of the RF chain corresponding to the k-th user, is the phase shift matrix implemented by the true delay units of the RF chain corresponding to the k-th user;

[0128] Use the alternating optimization algorithm to solve the optimization problem in equation (16), and alternately update the phase shift matrix, serial true delay units, and compensation delay.

[0129] First, fix the compensation delay and the delay of the serial true delay units, and update the phase shift matrix:

[0130]

[0131] where is the cumulative delay with compensation in equation (14), and f m is the frequency of the m-th subcarrier.

[0132] Secondly, fix the phase shift matrix and the compensation delay, and update the delays of the forward and backward serial true delay units:

[0133]

[0134] where in equation (13) The corresponding true delay unit should also be set to 0 in the delay update to avoid local optimal solutions.

[0135] The above problem is a simple one-dimensional search problem. By performing a one-dimensional search in [0, t max , the ideal value can be obtained, where t max is the maximum time delay value that can be achieved by the true time delay unit.

[0136] Finally, by fixing the time delay of the serial true time delay unit and the phase shift matrix, a one-dimensional search for the compensation time delay can be obtained

[0137]

[0138] where is the sum of the front / back serial parts obtained by one-dimensional search and accumulated according to the corresponding index set. Different from , there is no compensation time delay added. The search range of and is related to the number Q of true time delay units and is This invention considers the case of 16 true time delay units. At the position where the array gain is the lowest, the required compensation time delay value is the largest, about 0.208 ns. It can be seen that in the case of 16 true time delay units, the compensation time delay can be achieved by using only a single true time delay unit of the same type. Figure 3 Alternately update the phase matrix, the serial time delay value, and the compensation time delay value until

[0139] where γ is the convergence threshold. Equivalent channel

[0140] where H is the near-field channel function, and the digital beamforming matrix can be obtained according to the zero-forcing algorithm m

[0141] In the simulation experiment of this invention, when using 16 true time delay units, the single-user distance is 10 meters, the maximum time delay value of the true time delay unit is 0.2 ns, and the maximum time delay of the compensation time delay is 0.3 ns, the normalized array gain in different azimuth angles is as Figure 3 shown. It can be seen from the figure that in the case where the maximum time delay of the true time delay unit is limited, the performance of the traditional phase shifter method and the parallel true time delay method is severely lost, and the beam defocus phenomenon occurs. Although there is a loss of array gain in some angles compared with the ideal circuit situation for the serial-parallel hybrid true time delay architecture proposed in this invention, the beam defocus has been suppressed, and the effect of ensuring the array gain of each sub-band has been achieved.

[0142] In this invention, when the single user is located at a distance of 10 meters, the azimuth angle elevation angle and the maximum time delay of the true time delay unit is small, the variation of the single-user spectral efficiency with the maximum time delay of the true time delay unit is as Figure 4As shown, when the maximum time delay value of the true time delay unit is limited by hardware, the performance of the series-parallel hybrid true time delay architecture provided by the present invention is significantly better than that of the traditional phase shifter method and the parallel connection method.

[0143] In the case of multiple users, the variation of the spectral efficiency of the present invention with the maximum time delay of the true time delay unit is as Figure 5 shown. It can be seen from the attached drawings that the architecture of the present invention has reached the maximum value when the maximum time delay value of the true time delay unit is 0.2 ns, and the convergence speed is significantly better than that of the parallel architecture. Thus, it can be seen that the hardware requirements for the true time delay of the architecture provided by the present invention are much lower than those of the pure parallel architecture.

[0144] In summary, the series-parallel hybrid true time delay architecture and beamforming method of the near-field uniform circular antenna array proposed by the present invention, while suppressing beam defocusing and ensuring array gain, greatly reduce the maximum time delay value required by the true time delay unit, reduce the hardware requirements for the true time delay unit in the high-frequency band, and enable the overall architecture to still obtain an ideal beam focusing and signal transmission effect when the true time delay unit is limited.

Claims

1. A serial-parallel hybrid true time delay architecture and beamforming method for a near-field uniform circular antenna array, characterized in that, It includes the following steps: Step 1: In the near-field downlink broadband transmission scenario, establish a spherical coordinate model of a uniform circular antenna array, and determine the channel expression from each unit of the antenna array to the user; Step 2: Initialize the phase shift matrix of the phase shifter and the time delay value in the case of parallel true time delay units to obtain the optimal starting unit index under the ideal circuit; Step 3: According to the serial-parallel hybrid true time delay architecture, determine the starting unit of the serial part, initialize the serial true time delay unit and the time delay value of the compensating true time delay unit, and calculate the cumulative time delay at each true time delay unit under the serial-parallel hybrid architecture; Step 4: Use the alternating optimization algorithm to iteratively optimize the phase shift matrix, the time delay value of the true time delay unit, and the compensating time delay value to approximate the ideal beamforming effect, and design the digital beamforming matrix according to the equivalent channel.

2. A serial-parallel hybrid true time delay architecture and beamforming method for a near-field uniform circular antenna array according to claim 1, characterized in that, The said Step 2 includes: The phase shift matrix of the k-th user at the q-th subarray phase shifter [W k PS q is as follows:​ where a c (r k , θ k , φ k ) is the beam steering vector between the base station and user k at the center frequency, fc is the system center frequency, c is the speed of light, r k represents the distance from the user to the center of the circle in polar coordinates, θ k ∈ [0, π] and φ k ∈ [0, 2π] respectively represent the elevation angle and azimuth angle of user k to the center of the uniform circular array, R is the radius of the uniform circular antenna array, represents the azimuth angle of the center point of the q-th subarray to the center of the circle, N is the number of antenna elements, Q is the number of true time delay units, and P is the number of phase shifters connected to each true time delay unit; The parallel delay t of the k-th user in the q-th subarray true delay unit (q,k) is as follows: According to Equation (2), t can be obtained (q,k) is positively correlated with the distance from the center point of the q-th subarray of the antenna to the user. Based on the symmetry of the uniform circular array, the Q true time delay units are evenly divided into two groups. Serial connection is adopted within the group, and parallel connection is adopted between the two groups; Let the index of the true delay unit with the lowest delay value for the k-th user be Define the index set of the subarray For the forward serial set, the starting index of the forward serial is while For the backward serial set, the starting index of the backward serial is In the ideal circuit case, the serial starting position can be any true delay unit. Based on the user's position, the starting index of the front / back serial is obtained, and a selection circuit is used to select the appropriate starting true delay unit, which can give full play to the advantage of the cumulative delay. and is the starting unit index in the ideal circuit case.

3. A serial-parallel hybrid true time delay architecture and beamforming method for a near-field uniform circular antenna array according to claim 1, characterized in that, The said Step 3 includes: It may appear at any true delay unit. A Q out of 1 selection circuit is required between the RF chain and the true delay network to obtain Meanwhile, as the user's location changes, the index sets of the forward and backward serials also change frequently, resulting in a large number of dynamic switches required for the circuit. The change in the series-parallel structure may also lead to synchronization problems. Therefore, it is necessary to specify the starting position to reduce the circuit complexity; Divide Q true delay units into 4 groups according to quadrants, and it is stipulated that the starting index point of the backward serial can only be selected from the last unit of each group among them, and the starting point of the corresponding forward serial is the starting point of the backward serial corresponding to the k-th user is The corresponding forward serial index set and the backward serial index set are as follows: The delay value corresponding to the q-th serial true delay unit of the k-th user Introduce a compensation true time delay unit τ before the selection circuit front and τ back to compensate for the time delay error caused by q start ≠q min and initialize The cumulative delay obtained by the q-th true delay unit at this time can be accumulated according to the serial index set to obtain: wherein is the true delay unit index corresponding to the x-th element in the k-th user's forward serial set, is the true delay unit index corresponding to the y-th element in the k-th user's backward serial set, 4. A serial-parallel hybrid true time delay architecture and beamforming method for a near-field uniform circular antenna array according to claim 3, characterized in that, The said Step 4 includes: Aiming at maximizing the array gain G, an unconstrained optimal analog beamforming phase matrix is obtained Equation (8) reaches equality when Q = N and P = 1, and the unconstrained optimal analog beamforming phase matrix can be obtained. Model the analog beamforming precoder as an optimization problem: where W k PS is the phase shift matrix of the k-th user's corresponding RF chain phase shifter, is the phase shift matrix implemented by the true time delay unit of the k-th user's corresponding RF chain; Use the alternating optimization algorithm to solve the optimization problem in Equation (9), and alternately update the phase shift matrix, the serial true time delay unit, and the compensation time delay; First, fix the compensation time delay and the time delay of the serial true time delay unit, and update the phase shift matrix: Among them is the cumulative time delay after adding compensation in Equation (7), f m is the frequency of the m-th subcarrier; Secondly, fix the phase shift matrix and the compensation time delay, and update the time delay of the serial true time delay unit before and after; Among them In formula (6) The corresponding true time delay unit should also be set to 0 during the time delay update to avoid local optimal solutions; The above problem is a simple one-dimensional search problem. By performing a one-dimensional search in the range of [0, t max , the ideal value can be obtained, where t max is the maximum delay value achievable by the true delay unit; Finally, fix the time delay of the serial true time delay unit and the phase shift matrix, and perform a one-dimensional search for the compensation time delay to obtain Among them is the front / back serial part obtained by one-dimensional search Accumulate according to the corresponding index set, and The difference is that no compensation time delay is added and The search range of is related to the number Q of true time delay units as Alternately update the phase matrix, the serial delay value, and the compensation delay value until where γ is the convergence threshold.

Images