A four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding and its power allocation method

Abstract

This invention discloses a four-dimensional antenna array multi-beam transmission system and its power allocation method based on zero-forcing precoding. Addressing the problem of severe interference between high-power beam sidelobes and low-power beam main lobes in existing four-dimensional antenna array multi-beam systems with non-uniform power allocation, this invention determines the power allocation coefficients of each beam based on the user distance, constructs an inter-beam coupling matrix, and inverses it to obtain a zero-forcing precoding matrix. At the baseband side, the scaled beam symbols are precoded and pre-modulated, then superimposed into a single common baseband signal, which is up-converted and transmitted by a time-modulated array. The receiver can directly demodulate without precoding prior information. This invention effectively suppresses inter-beam interference while compensating for path loss differences, significantly improves the bit error rate performance of low-power beams, and reduces the overall system power consumption. Furthermore, the receiver is completely transparent, with low hardware complexity, making it suitable for various multi-beam communication systems.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication and antenna array technology, specifically relating to a power allocation and interference suppression method for a four-dimensional antenna array multi-beam system, and particularly a transmitter power control scheme based on zero-forcing precoding. Background Technology

[0002] Multi-beam technology is one of the key technologies for improving the spectral efficiency, coverage, and user access capacity of wireless communication systems. Traditional multi-beam implementation methods are mainly divided into Digital Beamforming (DBF), Analog Beamforming (ABF), and a hybrid architecture of the two (Hybrid Beamforming (HBF). DBF is equipped with a complete RF link after each antenna element, and performs amplitude and phase weighting in the digital domain, which has extremely high flexibility and accuracy. It can generate multiple independently controllable beams simultaneously and supports complex multi-user MIMO processing. However, as the number of antenna elements increases, the hardware cost, power consumption, and signal processing complexity of DBF increase sharply, especially in millimeter-wave massive MIMO systems. ABF uses only a single RF link and applies phase shifts to each array element through a phase shifter network. It has low hardware cost and low power consumption, but can only generate one beam at a time, and cannot support multi-beam transmission at the same frequency, thus limiting spectral efficiency. HBF combines the advantages of DBF and ABF, using a small number of RF links to connect the analog phase-shifting network, achieving a trade-off between hardware complexity and system performance. However, its number of multiple beams is limited by the number of RF links, and the analog part still has limitations such as weak quantization phase and amplitude control capabilities.

[0003] Unlike traditional phased arrays, four-dimensional antenna arrays introduce time as a fourth dimension and utilize switching modulation to achieve amplitude weighting, significantly reducing hardware complexity compared to fully digital beamforming architectures. The Hybrid Digital-Analog Modulation (HDAM) framework enables simultaneous transmission of multiple beams at the same frequency by combining baseband modulation and time modulation. Specifically, by rationally designing the timing of the switching of each array element, the four-dimensional antenna array can superimpose different modulation sidebands onto the baseband signal, thus requiring only one RF front-end to achieve multi-beam transmission. Patent CN118590109A proposes a method and system for multi-beamforming of a four-dimensional antenna array based on spatiotemporal joint baseband modulation.

[0004] However, existing HDAM frameworks typically allocate equal transmit power to all beams in multi-beam transmission. In real-world communication scenarios, the distance differences between different users and the base station result in varying path losses for each beam. Continuing with equal power allocation leads to wasted transmit power for near-end users, who don't require excessively high transmit power; conversely, far-end users receive significantly weaker signals, degrading communication quality and potentially preventing connection establishment. To compensate for these path loss differences, an intuitive improvement is to proportionally adjust the transmit power of the corresponding beams based on user distance (lower for near-end, higher for far-end). In HDAM frameworks, this corresponds to scaling the modulation signal amplitude of each beam at the baseband side. For example, Kejin Chen et al.'s paper, "Joint Beamforming and Waveform Optimization in 4-DAntenna Arrays for Integrated Sensing and Communication Systems," proposes using optimization algorithms to achieve arbitrary beamform patterns for amplitude scaling. However, this method has high optimization complexity and is not suitable for real-time computation.

[0005] Further research revealed that the aforementioned simple amplitude scaling method introduces new technical problems: since the radiation pattern of any practical antenna array has non-zero sidelobes, when the sidelobes of a high-power beam (corresponding to a far-end user) point towards the main lobe direction of a low-power beam (corresponding to a near-end user), the sidelobe signal of the high-power beam will strongly interfere with the useful signal of the low-power beam, causing the user to be unable to demodulate correctly. Therefore, how to effectively suppress inter-beam interference exacerbated by non-uniform power distribution at the transmitting end without introducing complex receiver processing has become a pressing technical problem to be solved in this field. Summary of the Invention

[0006] The technical problem this invention aims to solve is that existing four-dimensional antenna array-based systems face a dilemma in multi-beam transmission: if equal power allocation is used, it cannot compensate for the path loss differences caused by distance variations among users, resulting in excessively low received power for distant users; if simple baseband amplitude scaling is used to achieve non-uniform power allocation, the presence of non-zero sidelobes in the actual array pattern causes severe interference between the sidelobes of high-power beams and the main lobes of low-power beams, leading to a decrease in the received signal-to-interference ratio (SIR) or even demodulation failure for low-power beam users. Therefore, this invention provides a power allocation method based on zero-forcing precoding, aiming to simultaneously achieve adaptive power allocation based on user distance and suppress inter-beam interference at the transmitter in a low-complexity manner, improving system bit error rate performance and energy efficiency while maintaining transparency at the receiver.

[0007] To achieve the above-mentioned objectives, the present invention adopts the following technical solution. Consider a four-dimensional antenna array with N elements of equal amplitude and in-phase excitation; the expression for its far-field time-domain radiation field is:

[0008]

[0009] In the formula, f0 is the carrier frequency, N represents the number of antenna elements, and U k (t) represents the switching function, β represents the wavenumber in free space (β = 2π / λ, λ represents the wavelength in free space), d represents the distance between two adjacent units, θ represents the angular direction, t represents time, e represents the natural base, j represents the imaginary unit, and f p For carrier time modulation frequency, a mk U represents the equivalent complex excitation of the k-th element in the m-th order sideband. k (t) is the timing function of the k-th unit, which is the on / off switch here; U k The modulation period of (t) is T p =1 / f p That is, satisfying U k (t)=U k (t+nT p ), n=1,2,3…

[0010] After the harmonic components are filtered out by the bandpass filter, only the fundamental component (m=0) is retained. At this time, the equivalent complex excitation of the k-th element is a. 0k By using a specific timing design that ensures "only one switch is on at any given time," it is possible to guarantee... The system needs to generate Q beams simultaneously, pointing at angles θ1, θ2, ..., θQ respectively. Each beam p carries an independent baseband data symbol s. p (t). Let L be the path attenuation factor between user p and the base station. p To compensate for propagation differences, a scaling factor α is introduced. p , such that (α p ) 1 / 2 L p It is a constant. In practice, L p It can be obtained through uplink channel probing, user feedback, or distance estimation. This coefficient normalizes the amplitude of the received signal for each user.

[0011] The common baseband signal is:

[0012]

[0013] After time modulation and bandpass filtering, the signal received in direction θp can be expressed as:

[0014]

[0015] The coupling matrix G is defined as follows:

[0016]

[0017] That is, the coupling matrix G∈C Q×Q G(p,q) is defined as the normalized pattern gain in direction θp when the array forms a beam at angle θq. Specifically, when the main beam of the array points to θq, G(p,q) is the complex response of that pattern at θp. Although |G(p,q)| is small, when the power difference is significant, the interference term ∑ q≠p G(p,q)α q s q The amplitude of (t) may reach or even exceed that of the useful signal α. p s p The amplitude of (t) causes low-power beam users to be unable to demodulate correctly. Therefore, this invention introduces zero-forcing precoding to suppress this interference. The scaled symbol vector is defined as:

[0018]

[0019] Calculate the zero-forcing precoding matrix W=G −1 The pre-encoded symbol vector is:

[0020]

[0021] The pre-encoded symbol vector in formula (6) The p-th element in is defined as p (t). After introducing precoding, the common baseband signal is corrected as follows:

[0022]

[0023] After undergoing the same time modulation, filtering, and spatial propagation, the signal received by user p is:

[0024]

[0025] This demonstrates that inter-beam coupling is significantly reduced. The receiver can directly demodulate the signal without any pre-coding a priori information. p (t). Original baseband signal s p (t) satisfies the following condition: zero mean, i.e. [s p] = 0; unit amplitude, i.e. [|s p | 2 ]=1; the means are independent, that is, for i≠j, [s* is j ]=0. Due to the reasonable design of the time series, the array has low sidelobe characteristics, therefore the off-diagonal elements satisfy:

[0026]

[0027] Here, ε is a positive number and satisfies ε≪1 in the case of low sidelobes (e.g., ε≤0.01 corresponds to a 40dB sidelob level). The diagonal elements satisfy G(p,p)=1. The time series design can guarantee (Q-1)ε<1, therefore G −1 This can be expanded into a Neumann series:

[0028]

[0029] The diagonal elements of Δ are all 0, and the off-diagonal elements |Δ pq |≤ε(p≠q). Defining the baseband average power and swapping the order of expectation and integration, we obtain the baseband average power as:

[0030]

[0031] We introduced the time-averaged cross-correlation coefficient between different beams:

[0032]

[0033] It is obvious that ρ is satisfied pp =1, and from the Cauchy-Schwarz inequality, we can obtain |ρ pq |≤1. The covariance matrix is:

[0034]

[0035] Using the Neumann series expansion (10), we can obtain G −1 =I-Δ+O(ε 2 Substituting the first-order terms into equation (13), we get:

[0036]

[0037] Substituting equation (14) into equation (11) for baseband average power, we can separate the main term and the cross term:

[0038]

[0039] Take the absolute value of the cross terms and use |Δ pq |≤ε and|ρ pq If |≤1, the worst upper bound of the absolute value of the cross term can be obtained:

[0040]

[0041] It should be noted that the upper bound in equation (16) is based on the worst-case assumption that |ρ pq The value |ρ is derived from the condition that |ρ=1, but this only occurs when the two beams point in the exact same direction. In practical multi-beam systems, the beams point in different directions, therefore |ρ=1. pq | is usually much smaller than 1. Therefore, the contribution of the cross term is negligible, leading to a concise approximate expression:

[0042]

[0043] For the equal power allocation scheme, P can be obtained. bb ≈Q.

[0044] The peak value of the transmitted signal is determined by the baseband envelope |s(t)|, since |h p Given (t)|=1, we can obtain:

[0045]

[0046] The worst-case peak scenario in all | p | This occurs when the maximum value is reached simultaneously. From formula (6), we can obtain:

[0047]

[0048] Among them ||·|| ∞ This represents the infinity norm. Therefore, compared to the uncoded case, the amplification factor of the peak amplitude does not exceed ||G|. −1 || ∞ Using the Neumann series (10) and the row sum norm inequality ||Δ|| ∞ ≤Q(1-ε), we can obtain:

[0049]

[0050] The corresponding peak power amplification factor is . To maintain linearity, the power amplifier must operate with a power back-off amount proportional to the peak power. Therefore, compared with the equal power allocation scheme, the additional power back-off amount required by the scheme proposed in the present invention is:

[0051]

[0052] Substituting Equation (20) into Equation (21), the upper bound of the additional power back-off amount can be obtained as:

[0053]

[0054] In a four-dimensional antenna array, with the help of the baseband signal power in Equation (17), the ratio of the output power of the power amplifier in the proposed power allocation scheme to that in the conventional scheme is approximately:

[0055]

[0056] wherein G PA is the gain of the power amplifier. Due to the difference in spatial loss between users, ∑α p < Q; therefore, the proposed power allocation scheme reduces the output power of the power amplifier. The DC power consumption of the power amplifier is:

[0057]

[0058] wherein η is the efficiency of the power amplifier. Although η decreases as power back-off increases, the DC power reduction caused by the reduced output power is more significant. Taking a class B power amplifier as an example, the efficiency η satisfies the following proportional relationship:

[0059]

[0060] Substituting Equations (22), (23) and (25) into Equation (24), when the power amplifier is a class B amplifier, the ratio of the DC power consumption of the power amplifier in the two schemes is:

[0061]

[0062] If the power amplifier used is a high back-off efficiency power amplifier (such as Doherty power amplifier), the decrease of efficiency η with power back-off will be more limited, so the DC power consumption will decrease more significantly.

[0063] Based on the above principles, the four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding-assisted power allocation described in this invention will be used as an example for illustration. Figure 1 As shown, the system consists of a precoding module, a premodulation module, a baseband signal source, a timing control system, a local oscillator, a mixer, a power divider, an on / off RF switch array, a bandpass filter, a power amplifier, an antenna unit, and RF cables connecting the various parts. The precoding module is connected to the premodulation module, and the precoding module uses a pre-calculated zero-forcing precoding matrix W=G. −1 For the scaled beam symbol vectors s scaled A linear transformation is performed, and then the pre-modulated common baseband signal is obtained by the pre-modulation module. The baseband signal source outputs this common baseband signal to the mixer, while the timing control system is connected to the on / off RF switch array of each channel. The local oscillator signal generated by the local oscillator source and the baseband signal are up-converted by the mixer to generate an RF signal, which is then evenly distributed to N transmission channels (N is the number of antenna elements) by a power divider. Each channel has an on / off RF switch connected in series, and the state of the switch is controlled by the timing control system according to a pre-optimized switching sequence. Each on / off switch has only two states: state "1" (switch on, RF signal passes) and state "0" (switch off, no signal output). The timing control system independently generates a periodic switching control sequence for each array element. The signal modulated by the on / off switches passes sequentially through a bandpass filter and a power amplifier. The passband of the bandpass filter is set to [f0-f p / 2,f0+f p [ / 2] is used to filter out high-order harmonic components generated by switching modulation. The filtered signal is amplified by a power amplifier to the required transmit power and is finally radiated into free space by the corresponding antenna element. Through the above architecture, the present invention achieves zero-forcing pre-cancellation of multi-beam coupling interference at the transmitting end, while adaptively allocating the power of each beam according to the user distance, and the receiving end can directly demodulate without knowing the transmitter scaling ratio in advance.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] Suppressing inter-beam interference: By using zero-forcing precoding to cancel the coupling effect between beams in advance at the transmitter, even if non-uniform power distribution leads to large power differences, the signal-to-interference ratio of low-power beam users can still be significantly improved.

[0066] Transparent at the receiver: All preprocessing is completed at the transmitter. The user receiver does not need to know the precoding matrix or channel state information and can directly use the standard demodulator, which reduces terminal complexity and signaling overhead.

[0067] Adaptive power allocation: The transmit power of each beam is flexibly adjusted according to the user distance or path loss coefficient, which compensates for the difference in path loss and solves the communication quality problem for long-distance users.

[0068] Improved energy efficiency: On the one hand, it reduces unnecessary high-power transmission waste; on the other hand, it significantly reduces DC power consumption by combining high-efficiency power amplifiers and reasonable backoff design.

[0069] Low hardware complexity: Compared with the existing multi-beamforming framework based on space-time joint baseband modulation four-dimensional antenna array, it only adds one baseband matrix inversion and multiplication operation, and the radio frequency part still maintains the single-link, time-modulated switching structure, without increasing the number of radio frequency channels.

[0070] Wide range of applications: It can be extended to a large number of beams, is compatible with various low sidelobe synthesis methods (Taylor, Chebyshev, etc.), and can work in conjunction with a variety of high-efficiency power amplifiers. Attached Figure Description

[0071] Figure 1 This is a schematic diagram of a four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding-assisted power distribution. The system consists of a premodulation module, a precoding module, a baseband signal source, a timing control system, a local oscillator, a mixer, a power divider, an on / off RF switch, a bandpass filter, a power amplifier, antenna elements, and cables connecting the various parts.

[0072] Figure 2 In order to achieve the requirements of beam pointing angle, beam power level and sidelobe level in the implementation plan, the turn-on timing diagram of each unit is shown. Black indicates turn-on and white indicates turn-off.

[0073] Figure 3 According to Figure 2 Power distribution of the multi-beam array formed by timing. Five beams are implemented pointing to -40°, -28°, 0°, 17°, and 38° respectively, with power distribution levels of 1, 0.21, 0.28, 0.01, and 0.58. The solid-line circle symbol represents the beam pointing to -40° (beam 1), the solid-line square symbol represents the beam pointing to -28° (beam 2), the solid-line upper triangle symbol represents the beam pointing to 0° (beam 3), the solid-line diamond symbol represents the beam pointing to 17° (beam 4), and the solid-line lower triangle symbol represents the beam pointing to 38° (beam 5).

[0074] Figure 4The image shows the bit error rate curves for each beam under an equal power allocation (no precoding) scheme. Solid lines indicate the following: a circle (beam 1) pointing to -40°; a square (beam 2) pointing to -28°; an upper triangle (beam 3) pointing to 0°; a diamond (beam 4) pointing to 17°; and a lower triangle (beam 5) pointing to 38°.

[0075] Figure 5 The bit error rate curves for each beam under a non-uniform power distribution but no zero-forcing precoding scheme are shown. Solid line circles represent beams pointing to -40° (beam 1), solid line squares represent beams pointing to -28° (beam 2), solid line upper triangles represent beams pointing to 0° (beam 3), solid line diamonds represent beams pointing to 17° (beam 4), and solid line lower triangles represent beams pointing to 38° (beam 5).

[0076] Figure 6 The diagram shows the bit error rate curves for each beam in the proposed scheme. Solid line circles represent beams pointing to -40° (beam 1), solid line squares represent beams pointing to -28° (beam 2), solid line upper triangles represent beams pointing to 0° (beam 3), solid line diamonds represent beams pointing to 17° (beam 4), and solid line lower triangles represent beams pointing to 38° (beam 5).

[0077] Figure 7 The curves represent the total power consumption as a function of maximum beam power in a traditional switched series amplifier architecture. Solid circular symbols represent equal power distribution schemes, and solid square symbols represent the scheme proposed in this invention.

[0078] Figure 8 This is a schematic diagram of a direct-level control architecture for a power amplifier. The system consists of a baseband signal source, a timing control system, a pre-modulation module, a pre-coding module, a local oscillator, a mixer, a power divider, a bandpass filter, a power amplifier, an antenna unit, and RF cables connecting the various parts.

[0079] Figure 9 The curve shows the total power consumption as a function of the maximum beam power under the direct level control architecture of the amplifier. The solid circle symbol represents the equal power allocation scheme, and the solid square symbol represents the scheme proposed in this invention. Detailed Implementation Plan

[0080] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0081] The following example uses a uniform linear array of 16 antenna elements, with an element spacing of half a wavelength, to illustrate the multi-beam transmission effect based on zero-forcing precoding-assisted power allocation, while simultaneously transmitting five 16-QAM signals as a specific application. Assume the five target directions are θ1=-40°, θ2=-28°, θ3=0°, θ4=17°, and θ5=38°, and the signal modulation frequency f... p =10MHz, the symbol rate of all 5 16-QAM signals is 2Msym / s, and the signal-to-noise ratio at the receiver is set to 18dB. The distance between each user and the base station is different, and the corresponding path loss difference is compensated by the power allocation coefficient α=[1,0.21,0.28,0.01,0.58], with the highest and lowest beam power differing by 20dB. Figure 1 A schematic diagram of the four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding-assisted power allocation proposed in this invention is given.

[0082] Five beams are generated according to the steps proposed in this invention. First, the switching duty cycle of the 16 array elements is designed using Taylor weighting (sidelobe level -30 dB), and the phase distribution of the switching timing is optimized so that the array pattern forms the main beams in the five target directions. The resulting switching timing is as follows: Figure 2 As shown, the corresponding multibeam radiation pattern is as follows: Figure 3 As shown. At this point, since the sidelobe level is approximately -30dB, the coupling coefficient between beams ε ≤ 0.032. Based on this, the coupling matrix G∈C is constructed. 5×5 Its off-diagonal elements represent the gain of each beam sidelobe in the directions of the main lobe of other beams. Calculate the zero-forcing precoding matrix W=G. −1 The scaled symbol vector is pre-encoded to obtain a common baseband signal. This signal is then up-converted and distributed to 16 channels via a power divider, with each channel operating according to... Figure 2 The timing-controlled on / off switch, after being filtered by a bandpass filter to remove harmonics and amplified by a power amplifier, is radiated out by the antenna unit.

[0083] The received signal was demodulated in five directions: θ1=-40°, θ2=-28°, θ3=0°, θ4=17°, and θ5=38°. For comparison, equal power allocation (α) is also given. p Two schemes are available: one with all values ​​set to 1 (no precoding) and the other with non-uniform power distribution but no precoding. Figure 4 The bit error rate is the bit error rate of each beam under the equal power allocation scheme. It can be seen that although the bit error rate of each beam is low, in actual systems, equal power allocation will result in insufficient receiving power for remote users, making it difficult to achieve this bit error rate level. Figure 5 For non-uniform power distribution without a precoding scheme, the bit error rate of the low-power beam (pointing at 17°) is as high as 0.1 or higher, making normal communication impossible; Figure 6To achieve the desired bit error rate for the proposed solution, the bit error rate of all beams in their target direction is reduced to 10. -4 As can be clearly seen below, this invention effectively suppresses the interference of high-power beam sidelobes on low-power beams, thus achieving reliable communication.

[0084] To comprehensively evaluate the energy efficiency advantages of the present invention, two different hardware implementation architectures for the transmission channel are considered: the first is a traditional switched series amplifier architecture (see schematic diagram). Figure 1 The second type is the amplifier direct-level control architecture (which can achieve zero power consumption in the off-state and no switching insertion loss, as shown in the schematic diagram). Figure 8 In both architectures, the DC power consumption of the equal power allocation scheme (without precoding) and the scheme of this invention (zero-forcing precoding + non-uniform power allocation) are compared. It should be noted that although the non-uniform power allocation scheme without zero-forcing precoding has a lower total output power, its bit error rate performance is extremely poor (e.g., ...). Figure 5 As shown, the low-power beam has a bit error rate as high as 0.1% or higher, which cannot guarantee normal communication, and therefore is not included in the energy consumption comparison. The simulation uses a CGH40006S GaN HEMT power amplifier model with 5 beams Q, power allocation coefficient α=[1,0.21,0.28,0.01,0.58], and the maximum beam power variation range is 0~30dBm.

[0085] (1) Traditional switch-series amplifier architecture

[0086] In this architecture, the RF switch of each array element channel is cascaded with a power amplifier. Insertion loss occurs when the switch is on, and a small amount of leakage current loss remains when it is off. Unwanted harmonics generated by time modulation are filtered out by a bandpass filter. The architecture diagram is shown below. Figure 1 The efficiency of the power amplifier is referenced from the datasheet provided in the chip datasheet; efficiency decreases as output power decreases. Figure 7 The figure shows the curves of total power consumption as a function of maximum beam power for the two schemes under this architecture. The equal power allocation scheme (square-marked curve) has the highest total power consumption at each power point. The scheme of this invention (triangle-marked curve) maintains the power consumption reduction brought about by non-uniform power allocation while reducing inter-beam interference through zero-forcing precoding, thus ensuring communication reliability (e.g., Figure 6 As shown, the bit error rate of all beams is reduced to 10. -4 (The following is an example.) Taking a maximum beam power of 30dBm as an example, the total power consumption of the equal power allocation scheme is 55.3W, while the total power consumption of the present invention is 49.08W, saving approximately 11.2%. Across the entire power range, the total power consumption of the present invention is significantly lower than that of the equal power allocation scheme.

[0087] (2) Amplifier direct level control architecture

[0088] In this architecture, the power amplifier is directly driven by timing control signals: when it needs to be turned off, a corresponding voltage is applied to completely cut off the amplifier, the drain current is almost zero, the power consumption in the off-state is negligible, and the series RF switch is eliminated, resulting in no insertion loss. A schematic diagram of the architecture is shown below. Figure 8 . Figure 9 The figure shows the curves of total power consumption as a function of maximum beam power for the two schemes under this architecture. Taking a maximum beam power of 30dBm as an example, the total power consumption of the equal power allocation scheme is 12.51W, while the total power consumption of the scheme of this invention is 7.476W, saving approximately 40.2%. The power consumption comparison trend across the entire power range is consistent with the traditional architecture, and the scheme of this invention consistently demonstrates a significant energy-saving advantage.

[0089] It is important to emphasize that regardless of the hardware architecture used, the solution of this invention can achieve a significant reduction in energy consumption while ensuring communication quality. Although non-uniform power allocation can reduce the total output power, without zero-forcing precoding, the resulting inter-beam interference will prevent the system from functioning properly, rendering the power consumption advantage meaningless. Therefore, the zero-forcing precoding-assisted power allocation strategy proposed in this invention is key to achieving low-power, high-reliability multi-beam transmission, and it can work well with power amplifier architectures of different efficiency characteristics.

[0090] In summary, the multi-beam transmission method for four-dimensional antenna arrays based on zero-forcing precoding-assisted power allocation proposed in this invention can achieve adaptive beam power allocation according to user distance while maintaining the low hardware complexity of the four-dimensional antenna array. At the same time, it can effectively suppress inter-beam interference exacerbated by non-uniform power allocation, significantly improve the bit error rate performance of low-power beams, and reduce the total power consumption of the system. It has obvious technical advantages and practical value.

[0091] The above description is intended for those skilled in the art and describes the invention and its embodiments. This description should be considered illustrative rather than limiting. Those skilled in the art can implement the invention in detail according to the ideas in the claims, and can also make certain changes in form and detail without departing from the spirit and scope of the invention. These changes should all be considered within the scope of the invention.

Claims

1. A four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding-assisted power allocation, characterized in that, It includes: a pre-modulation module, a pre-coding module, a baseband signal source, a timing control system, a local oscillator, a mixer, a power divider, on / off RF switches, a bandpass filter, a power amplifier, an antenna unit, and cables connecting the various parts. The pre-coding module is connected to the pre-modulation module, which is connected to the baseband signal source. The baseband signal source is connected to the signal input of the mixer. The timing control system is connected to the control terminals of each on / off RF switch. The local oscillator is connected to the local oscillator input of the mixer. The output of the mixer is connected to the input of the power divider. Multiple outputs of the power divider are connected to the inputs of each on / off RF switch. The output of each on / off RF switch is connected to a bandpass filter and then to the corresponding power amplifier. The output of each power amplifier is connected to the corresponding antenna unit. The pre-coding module is configured to: perform a linear transformation on the symbol vector scaled by the beam power allocation coefficients according to a pre-calculated pre-coding matrix to obtain a pre-coded symbol vector. The pre-modulation module is configured to: add the pre-coded symbols after pre-modulation to generate a common baseband signal.

2. A power allocation method for a four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding-assisted power allocation as described in claim 1, characterized in that... Includes the following steps: S1. Determine the power allocation coefficient α for each beam based on the distance or path loss between each user and the base station. p This normalizes the amplitude of the received signal for each user. S2. Design the turn-on timing of the switching control signals of each array element to ensure that the array pattern meets the low sidelobe condition; construct the inter-beam coupling matrix based on the pattern, and the proposed timing needs to meet the following constraints: (1) There is one and only one switch turned on at any time, i.e. Where N represents the number of antenna elements, U k (t) represents the timing function of the kth unit; (2) The duty cycle of each unit is weighted by Taylor or Chebyshev to keep the sidelobe level low and ensure that the off-diagonal elements of the array coupling matrix G satisfy |G pq |≤ε(p≠q), where ε satisfies (Q-1)ε<1, and Q is the number of beams; S3. Calculate the zero-forcing precoding matrix, and precode and premodulate the scaled symbol vector in the precoding module and premodulation module, and then superimpose them to generate a common baseband signal. S4. The common baseband signal is up-converted and power-distributed before being sent to each channel. It is then time-modulated by the on / off switch controlled by the timing design in step S2, filtered, amplified, and radiated by the antenna unit. The receiver directly demodulates the signal without pre-coding prior information.

3. The power allocation method for a four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding according to claim 2, characterized in that, Improve system energy efficiency through the following mechanisms: (1) Based on the path loss between each user and the base station, the power allocation coefficient of each beam is determined to be a positive real number α. p Where the subscript p represents the beam number, satisfying (α p ) 1 / 2 L p L is a constant. p It is the path decay factor for the p-th user; (2) Through this non-uniform power distribution, the total transmit power of the system is reduced to approximately equal to the transmit power of the equal power distribution scheme. 1 times, where Q is the total number of beams generated simultaneously; (3) Zero-forcing precoding is performed on the scaled symbols in the baseband, and the precoding matrix is ​​W=G −1 Where the coupling matrix G∈C Q×Q G(p,q) is defined as the normalized pattern gain in direction θp when the array forms a beam at angle θq. (4) Zero-forcing precoding increases the peak amplitude of the transmitted signal, with the maximum increase factor being the infinity norm ||W|| of the precoding matrix. ∞ This necessitates an additional back-off of the power amplifier's operating point. The upper bound of this additional backoff amount is ||W|| ∞ Convergence is guaranteed.

4. The power allocation method for a four-dimensional antenna array multi-beam transmission system based on zero-forcing precoding according to claim 2 or 3, characterized in that, The precoding mentioned refers to zero-forcing precoding, specifically, in order to suppress interference between beams, the scaled symbol vector is left-multiplied by a zero-forcing precoding matrix at the transmitter. The precoding process requires only one Q-order matrix inversion operation and one Q-order matrix multiplication operation; the inversion operation refers to calculating the precoding matrix as W=G. −1 Multiplication refers to the calculation of the pre-coded symbol vector. .

Images