Predistortion system and method for analog fully connected hybrid beamforming system

Abstract

The application discloses a predistortion system and method for an analog full-connection hybrid beamforming system, and relates to the technical field of mobile communication, and comprises a beamforming coefficient network, an input end of the beamforming coefficient network receives input signals of each channel; a nonlinear base function generation module, the nonlinear base function generation module comprises a plurality of one-dimensional nonlinear base functions, an input end of the one-dimensional nonlinear base function is connected with an output end of the beamforming coefficient network, and receives input signals of equivalent power amplifiers of each channel; a plurality of space domain response modules, an input end of the space domain response module is connected with an output end of the one-dimensional nonlinear base function; a plurality of multipliers, weighted linear superposition signals output by the space domain response module are multiplied with respective coefficients by using different multipliers to obtain different results; and an adder, different results are added by using the adder to obtain a predistortion signal. The application can reduce the technical complexity of DPD in the analog full-connection hybrid beamforming system.

Description

Technical Field

[0001] This invention belongs to the field of mobile communication technology, specifically relating to a predistortion system and method for analog fully connected hybrid beamforming systems. Background Technology

[0002] Modern wireless communication systems widely employ Multiple-Input Multiple-Output (MIMO) technology to support the rapidly increasing number of users. In the upcoming 5G era, the number of transmitter radio frequency links will further increase. Correspondingly, Massive MIMO (mMIMO) technology has become a crucial technology in 5G transmitters, promising to significantly improve network capacity and data transmission rates, as well as communication reliability. Massive MIMO systems are often combined with beamforming (BF) technology to improve spectral efficiency. There are two typical architectures for massive MIMO hybrid beamforming systems: subarray architecture (SA) and fully-connected architecture (FC). The hybrid beamforming (HBF) architecture strikes a balance between the flexibility and complexity of beamforming, making it a promising beamforming solution for massive MIMO systems.

[0003] Power amplifiers (PAs) are among the most power-consuming components in transmitters, exhibiting significant nonlinear distortion when operating in the high-efficiency region. To meet transmitter linearity and efficiency requirements, digital predistortion (DPD) technology is widely used in transmitters. Currently, research on DPD linearization schemes for HBF arrays mainly focuses on subarray-connected architectures, with little research on HBF array linearization under analog fully connected architectures. Compared to subarray-connected architectures, in analog fully connected hybrid beamforming arrays, the transmitted signals from all transmit links undergo phase-shifting and combining through an analog beamforming network located at the amplifier front end. This mixed signal generates intermodulation distortion between different transmitted signals after amplification, making the DPD linearization problem in the system more challenging. Therefore, it is desirable for DPD schemes to not only eliminate the independent nonlinear distortion of each signal in the transmit link but also to eliminate intermodulation distortion between different transmitted signals. In this case, the conventional single-input single-output (SISO) DPD modeling scheme suitable for subarray architectures can no longer linearize analog fully connected HBF arrays, and DPD schemes under analog fully connected architectures require further research. Summary of the Invention

[0004] To address the aforementioned problems in the prior art, this invention provides a predistortion system and method for a simulated fully connected hybrid beamforming system. The technical problem to be solved by this invention is achieved through the following technical solution:

[0005] In a first aspect, the present invention provides a predistortion system for an analog fully connected hybrid beamforming system, comprising:

[0006] Beamforming coefficient network: The input terminal of the beamforming coefficient network receives the input signals of each channel and is used to generate the equivalent input signals of each channel power amplifier.

[0007] The nonlinear basis function generation module includes multiple one-dimensional nonlinear basis functions. The input end of the one-dimensional nonlinear basis function is connected to the output end of the beamforming coefficient network, receives the input signals of the equivalent power amplifiers of each channel, and performs nonlinear operations on the input signals of the equivalent power amplifiers of each channel.

[0008] The spatial response module is equipped with multiple spatial response modules. The input of the spatial response module is connected to the output of the one-dimensional nonlinear basis function, which is used to perform weighted linear superposition of the output signal processed by the nonlinear basis function generation module.

[0009] Multiple multipliers are used to multiply the weighted linear superposition signal output by the spatial response module with their respective coefficients to obtain different results;

[0010] An adder is used to sum different results to obtain a predistorted signal.

[0011] Secondly, the present invention also provides a predistortion method for a simulated fully connected hybrid beamforming system, comprising:

[0012] Input signals from each channel are fed into the predistortion system;

[0013] The beamforming coefficient network responds to the input signal of each channel and outputs the equivalent input signal of each channel power amplifier;

[0014] The nonlinear basis function generation module responds to the input signals of the equivalent power amplifier channels by performing nonlinear operations on the input signals of the equivalent power amplifier channels.

[0015] The spatial response module responds to the output signal processed by the nonlinear basis function generation module, and performs weighted linear superposition to output a weighted linear superposition signal;

[0016] The weighted linear superposition signal output by the spatial response module is multiplied by its respective coefficients to obtain different results. The different results are then added together to obtain the predistorted signal.

[0017] Thirdly, the present invention also provides a simulated fully connected hybrid beamforming system, comprising: a predistortion system for reducing complexity.

[0018] The beneficial effects of this invention are:

[0019] This invention provides a predistortion system and method for a simulated fully connected hybrid beamforming system, comprising a beamforming coefficient network, a nonlinear basis function generation module, and a spatial response module. By introducing beamforming and spatial response modules, the model basis function is reduced from multidimensional to one-dimensional, while significantly reducing the number of model coefficients in the predistorter, thereby reducing the complexity of DPD technology in the simulated fully connected hybrid beamforming system.

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of a digital predistortion architecture on which the predistorter model provided in the prior art embodiments is based;

[0022] Figure 2 This is a schematic diagram of a predistorter model architecture provided in a prior art embodiment;

[0023] Figure 3 This is a schematic diagram of a predistortion system for a simulated fully connected hybrid beamforming system provided in an embodiment of the present invention;

[0024] Figure 4 This is a flowchart of a predistortion method for a simulated fully connected hybrid beamforming system provided in an embodiment of the present invention;

[0025] Figure 5 This is a schematic diagram comparing the number of structural coefficients and model accuracy of the traditional multidimensional correction model and the predistorter model proposed in this invention for different signal stream numbers Q.

[0026] Figure 6 This is a schematic diagram of the simulation experiment results provided in an embodiment of the present invention. Detailed Implementation

[0027] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0028] For related technologies, please refer to Figure 1 , Figure 1 This is a schematic diagram of a digital predistortion architecture based on a predistorter model provided in a prior art embodiment, used for linearizing a hybrid massive MIMO transmitter with a fully connected analog architecture. This digital predistortion architecture includes Q predistorters, each capable of linearizing the corresponding beam signal. When a predistorter needs updating, the output signals y1,…,y of each power amplifier are updated. p After filtering, down-conversion, and analog-to-digital conversion, the signals are combined with the input signals s1,…,s of the digitally synthesized equivalent power amplifier.P Alignment; then, the multi-beam estimation module synthesizes equivalent multi-beam signals in the digital domain based on the beamforming conditions of the aligned power amplifier output signal. These estimated multi-beam signals are transmitted to the predistorter training module and used together with the transmitted signal for training and updating the DPD model.

[0029] Please see Figure 2 , Figure 2 This is a schematic diagram of a predistorter model architecture provided in the prior art embodiments. Since there is intermodulation distortion from each transmission channel signal in the array beam signal, it is necessary to use a multivariate model based on Q input to describe and compensate for this nonlinear distortion, as shown in formula (1). The predistorter model architecture is extremely complex, and the required predistorter model coefficients increase exponentially with the increase of the number of signal streams and the model order, resulting in huge computational resource overhead. In addition, considering that the predistortion algorithm needs to be implemented in real time in the system, this predistorter model faces great application difficulties.

[0030]

[0031] In view of this, the present invention provides a predistortion system for analog fully connected hybrid beamforming systems. By introducing prior information on beamforming coefficients and spatial response, the system avoids directly constructing intermodulation terms of multi-stream signals in the predistorter, thereby significantly reducing the number and complexity of predistorter model coefficients and solving the problem of digital predistortion linearization in analog fully connected architecture hybrid beamforming large-scale MIMO systems.

[0032] Specifically, the analysis of the beam signal includes:

[0033] First, the input signal of the p-th power amplifier in the array can be expressed as:

[0034]

[0035] Where, x q (n) represents the complex envelope signal from the q-th transmit link, s p (n) represents the complex envelope of the input signal of the p-th power amplifier, which can be regarded as a weighted sum of the transmitted signals, and its beamforming weight is w. p1 ,...,w pQ w pQ Let |w| represent the phase shift of the q-th transmitted signal in the input signal of the p-th power amplifier, satisfying |w|. pq |=1.

[0036] The nonlinear behavior of the array power amplifier can be represented by a memory polynomial model, then the output signal of the p-th power amplifier is expressed as:

[0037]

[0038] in, Let M represent the coefficients of the p-th power amplifier model, K represent the memory depth, and M represent the nonlinear order. For ease of analysis, this invention assumes that the behavioral models of all power amplifiers in the array have the same nonlinear order and memory depth. Substituting formula (2) into formula (3), the output of the p-th power amplifier can be expressed as:

[0039]

[0040] Considering beamforming and spatial phase superposition effects, the q-th beam signal can be expressed as:

[0041]

[0042] Among them, h pq Let r represent the channel response from the p-th transmit antenna to the q-th user receiver. q (n) represents the beam signal of the q-th user; if the RF power amplifiers in the array all have similar nonlinear characteristics, i.e. but:

[0043]

[0044] The total number of model coefficients in this digital predistortion architecture is (M+1)×K, which significantly reduces the number of coefficients in the proposed structural model compared to traditional multi-input correction models.

[0045] Please see Figure 3 As shown, Figure 3 This is a schematic diagram of a predistortion system for a simulated fully connected hybrid beamforming system provided by an embodiment of the present invention. The predistortion system for a simulated fully connected hybrid beamforming system provided by the present invention includes:

[0046] A beamforming coefficient network, wherein the input terminal of the beamforming coefficient network receives the input signals of each channel and is used to generate the equivalent input signals of each channel power amplifier;

[0047] The nonlinear basis function generation module includes multiple one-dimensional nonlinear basis functions. The input end of the one-dimensional nonlinear basis function is connected to the output end of the beamforming coefficient network, receives the input signals of the equivalent power amplifiers of each channel, and performs nonlinear operations on the input signals of the equivalent power amplifiers of each channel.

[0048] The spatial response module is equipped with multiple spatial response modules. The input of the spatial response module is connected to the output of the one-dimensional nonlinear basis function, which is used to perform weighted linear superposition of the output signal processed by the nonlinear basis function generation module.

[0049] Multiple multipliers are used to multiply the weighted linear superposition signal output by the spatial response module with their respective coefficients to obtain different results;

[0050] An adder is used to sum the different results to obtain a predistorted signal.

[0051] Specifically, the predistortion system provided in this embodiment includes a beamforming coefficient network, a nonlinear basis function generation module, and a spatial response module. Compared with the predistortion system in the prior art, the addition of the beamforming coefficient network and the spatial response module reduces the model basis function from multidimensional to one-dimensional by introducing beamforming and spatial response modules. At the same time, it significantly reduces the number of model coefficients in the predistorter, thereby reducing the complexity of DPD technology in the simulated fully connected hybrid beamforming system.

[0052] It should be noted that the nonlinear basis function generation module includes multiple one-dimensional nonlinear basis functions, each of which is a single-input, single-output function.

[0053] In an optional embodiment of the invention, the beamforming coefficient network is the same as the beamforming network in the analog domain, but implemented in the digital domain; the coefficients of the beamforming coefficient network are [W]. Q×P .

[0054] In an optional embodiment of the present invention, the channel response coefficient of the spatial response module is [H]. P×Q .

[0055] Based on the same inventive concept, please refer to Figure 4 , Figure 4 This is a flowchart of a predistortion method for a simulated fully connected hybrid beamforming system provided in an embodiment of the present invention. It is used to implement the predistortion system for the simulated fully connected hybrid beamforming system provided in the above embodiment. Please refer to the above for embodiments of this predistortion system; repeated details will not be repeated. The predistortion method includes:

[0056] S101. Input the input signals of each channel to the predistortion system.

[0057] Specifically, in this embodiment, the input signals x1,…,x from each channel are... Q Input to the predistortion system.

[0058] S102, the beamforming coefficient network responds to the input signals of each channel and outputs the equivalent input signals of each channel power amplifier.

[0059] Specifically, in this embodiment, the expression for the input signal of the equivalent power amplifier for each channel is:

[0060]

[0061] Among them, w pq x represents the phase shift of the q-th transmit signal in the input signal of the p-th power amplifier, where p represents the index of each power amplifier (p = 1, 2, ..., P), and Q represents the number of predistortion systems. q (n) represents the complex envelope signal of the q-th transmit link, and n represents the time series, i.e. the nth signal sample point.

[0062] S103, the nonlinear basis function generation module responds to the input signals of the equivalent power amplifiers of each channel and performs nonlinear operations on the input signals of the equivalent power amplifiers of each channel.

[0063] Specifically, in this embodiment, the expression for the output signal processed by the nonlinear basis function generation module is as follows:

[0064]

[0065] Among them, G t (·) represents the t-th nonlinear basis function, t = 1, 2, ..., N2, where N2 represents the total number of nonlinear basis functions; in this embodiment, the memory polynomial model is used as an example, then N2 = (M+1)K, m represents the index of the memory depth, m = 1, 2, ..., M, and k represents the index of the nonlinear order, k = 1, 2, ..., K.

[0066] S104. The spatial response module responds to the output signal processed by the nonlinear basis function generation module and performs weighted linear superposition to output a weighted linear superposition signal.

[0067] Specifically, in this embodiment, the expression for the weighted linear superposition signal output by the spatial response module is:

[0068]

[0069] Among them, h qp This represents the channel response from the p-th transmit antenna to the q-th user receiver.

[0070] It should be noted that there is a one-to-one correspondence between the transmitting antenna and the power amplifier, and a one-to-one correspondence between the transmitting link and the user receiver.

[0071] S105. Multiply the weighted linear superposition signal output by the spatial response module by its respective coefficients to obtain different results, and then add the different results to obtain the predistortion signal.

[0072] Specifically, in this embodiment, the expression for the predistortion signal is:

[0073]

[0074] Among them, a mk These represent the coefficients of the nonlinear basis functions.

[0075] In an optional embodiment of the present invention, please refer to Figure 5 and Figure 6 , Figure 5 This is a schematic diagram comparing the number of structural coefficients and model accuracy of the traditional multidimensional correction model and the predistorter model proposed in this invention for different signal stream numbers Q. Figure 6 This is a schematic diagram of the simulation experiment results provided in an embodiment of the present invention. Figure 5 The figures show the comparison of the number of coefficients and model accuracy of the traditional multidimensional correction model and the predistorter model proposed in this embodiment when the signal stream number Q is 2, 4, 6 and 8 respectively. Figure 6 The comparisons of NMSE before and after correction for four different azimuth and elevation angles are presented respectively. The predistorter model structure proposed in this embodiment has been verified by simulation of a fully connected hybrid beamforming array using a 4-stream 64-element model. Figure 5 It is evident that the model structure proposed in this invention can significantly reduce the number of model coefficients required while achieving similar accuracy to traditional models. Furthermore, Figure 6 Simulation results show that the proposed model structure can significantly compensate for the nonlinear distortion of the beam direction signal. All the above simulation results demonstrate that the proposed model structure has advantages in simulating fully connected hybrid beamforming systems.

[0076] Based on the same inventive concept, the present invention also provides a simulated fully connected hybrid beamforming system, including: a predistortion system for reducing complexity.

[0077] Specifically, in this embodiment, the predistortion system provided in the above embodiments is applied to an analog fully connected hybrid beamforming system to reduce the technical complexity of the DPD in the analog fully connected hybrid beamforming system.

[0078] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that an article or device comprising a list of elements includes not only those elements but also other elements not expressly listed. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device comprising said element. Terms such as "connected" or "linked" are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect. The orientations or positional relationships indicated by terms such as "upper," "lower," "left," and "right" are based on the orientations or positional relationships shown in the accompanying drawings and are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention.

[0079] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0080] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A pre-distortion system for an analog fully connected hybrid beamforming system, characterized in that, The pre-distortion system comprises: a beamforming coefficient network, an input end of the beamforming coefficient network receiving each channel input signal, and used for generating equivalent input signals of each channel power amplifier; an expression of the equivalent input signals of each channel power amplifier is: ； in, Indicates the first The input signal of the power amplifier is the first The phase shift coefficient of the transmitted signal. Indicates the index of each power amplifier. , Indicates the number of predistortion systems. Indicates the first The complex envelope signal of the transmission link, Represents a time series, i.e., the first... One signal sample point; a nonlinear basis function generation module, comprising a plurality of one-dimensional nonlinear basis functions, an input end of the one-dimensional nonlinear basis function being connected with an output end of the beamforming coefficient network, and receiving the equivalent input signals of each channel power amplifier, and performing nonlinear operation on the equivalent input signals of each channel power amplifier; an expression of an output signal processed by the nonlinear basis function generation module is: ； wherein, represents the nonlinear basis function, , represents the total number of nonlinear basis functions, represents an index of the memory depth, , represents an index of the nonlinear order, ; a plurality of spatial domain response modules, an input end of the spatial domain response module being connected with an output end of the one-dimensional nonlinear basis function, and used for performing weighted linear superposition on the output signal processed by the nonlinear basis function generation module; a plurality of multipliers, different weighted linear superposition signals output by the spatial domain response module being multiplied with respective coefficients by using different multipliers, to obtain different results; an adder, the different results being added by using the adder, to obtain a pre-distortion signal.

2. The pre-distortion system for an analog fully connected hybrid beamforming system of claim 1, wherein, The beamforming coefficient network is implemented in a digital domain, coefficients of the beamforming coefficient network are .

3. The pre-distortion system for an analog fully connected hybrid beamforming system of claim 1, wherein, The channel response coefficient of the spatial domain response module is .

4. A pre-distortion method for an analog full-connection hybrid beamforming system, characterized in that: each channel input signal is input into a pre-distortion system; a beamforming coefficient network outputs equivalent input signals of each channel power amplifier in response to the each channel input signal; an expression of an output signal processed by a nonlinear basis function generation module is: ； wherein, represents the nonlinear basis function, , represents the total number of nonlinear basis functions, represents an index of the memory depth, , represents an index of the nonlinear order, ; the nonlinear basis function generation module performs nonlinear operation on the equivalent input signals of each channel power amplifier in response to the equivalent input signals of each channel power amplifier; an expression of the equivalent input signals of each channel power amplifier is: ； wherein denotes the phase shift coefficient of the th transmit signal in the input signal of the th power amplifier, denotes the index of the respective power amplifier, , denotes the number of predistortion systems, denotes the complex envelope signal of the th transmit chain, denotes the time sequence, i.e. the th signal sample point; a spatial domain response module performs weighted linear superposition in response to the output signal processed by the nonlinear basis function generation module, and outputs a weighted linear superposition signal; the weighted linear superposition signal output by the spatial domain response module is multiplied with respective coefficients, to obtain different results, and the different results are added, to obtain a pre-distortion signal.

5. The method for predistortion for an analog fully connected hybrid beamforming system of claim 4, wherein, an expression of the weighted linear superposition signal output by the spatial domain response module is: ； wherein, denotes the channel response from the th transmit antenna to the th user receiver.

6. The method for predistortion for an analog fully connected hybrid beamforming system of claim 5, wherein, an expression of the pre-distortion signal is: ； wherein denote coefficients of the non-linear basis functions.

7. A method for analog full-connected hybrid beamforming system, characterized in that, The pre-distortion system comprises: The pre-distortion system of any one of claims 1-3, used for reducing complexity.

Images