TWI846624B - Digital pre-distortion system and method for beamforming signals - Google Patents

Abstract

The present invention relates to a digital pre-distortion system and method for beamforming signals. The system includes a first digital pre-distortion processing unit that applies digital pre-distortion parameters to an original signal to generate a digitally pre-distorted signal. The digitally pre-distorted signal passes through a beamforming unit and a reception unit to generate a beam-orient reception signal. A filtering unit samples the beam-orient reception signal with bandwidth limitation to produce a beam-orient limited-band reception signal. The beam-orient limited-band reception signal, after computation by an estimation unit and a second digital pre-distortion processing unit for a predetermined number of iterations, employs the final iteration result of the second digital pre-distortion processing unit as the updated digital pre-distortion parameters. These updated parameters are transmitted back to the first digital pre-distortion processing unit, thereby serving as the digital pre-distortion parameters. The digital pre-distortion processing method follows the processes. The system and method effectively address multiple nonlinear effects within the beamforming unit.

Description

Digital predistortion processing system and method for beamforming signal

The present invention relates to digital pre-distortion processing technology, and in particular to digital pre-distortion technology applied to a multi-input multi-output antenna communication system to solve the non-linear effects of multiple power amplifiers.

As the communication environment develops towards high frequencies, in order to overcome the challenges brought by signal transmission attenuation, the importance of related technologies such as massive multi-input multi-output (Massive MIMO) and array antennas will continue to increase. However, related technologies such as multi-input multi-output and array antennas are also accompanied by some challenges, one of which is nonlinear distortion.

In high-frequency communication environments, nonlinear distortion will occur as signals pass through complex channels and multiple antenna paths, thus affecting signal quality. Digital Pre-Distortion (DPD) technology becomes crucial in this situation because it can effectively reduce signal distortion and improve system performance and reliability.

However, as the number of antennas in the communication environment increases, the application of DPD technology has become more complex. Since each antenna path may have different nonlinear characteristics, an independent DPD compensation process is required for each antenna, such as US Patent Publication No. US20210399697 and US Patent Publication No. 20230137470. This will lead to the following problems:

1. Increased costs: Executing an independent DPD process requires more computing resources and hardware equipment, which in turn increases the cost of the system.

2. Increased system complexity: The independent DPD process requires more processing steps and control mechanisms, which increases the overall complexity of the system.

3. Increased design difficulty: Individual DPD design and adjustment is required for each antenna, which will increase the difficulty of system design.

In addition, Chinese Patent No. CN109981505B (patent name: Power-scalable beam-directed digital predistortion device and method, transceiver system) uses an indirect learning structure to calculate the coefficients of the beam splitting gain module based on the original input signal and the main beam signal, and generates a predistortion signal, and linearizes the main beam signal with the generated predistortion signal. However, the predistortion signal processing of this Chinese patent does not use the received signal of the receiving antenna to solve the nonlinear distortion problem, which will cause the received signal to still have serious nonlinear distortion problems.

Furthermore, Taiwan Patent No. I700888 (Patent Name: Digital Predistortion Circuit and Digital Predistortion Method) sequentially performs predistortion processing, digital-to-analog conversion, and amplification processing on an input signal to generate an output signal. The first bandwidth of the input signal after predistortion processing is greater than the second bandwidth of the input signal after digital-to-analog conversion. The output signal is filtered to filter signals outside the second bandwidth to generate a second output signal. The third output signal after amplification processing of the predistortion signal is estimated based on the predistortion signal of the input signal after predistortion processing and the second output signal. The third bandwidth of this third output signal is greater than the second bandwidth. The parameters of the predistortion processing are determined based on the third output signal and the predistortion signal. However, this process of repeated amplification and filtering, and adding different signals for processing is too complicated and will increase the design cost. Moreover, the problem to be solved by this Taiwan patent is not the in-band pre-distortion processing of the output signal. The parameters of the pre-distortion processing are determined to correct the distortion outside the second bandwidth.

Based on the above problems, nonlinear distortion compensation in future high-frequency communication environments will be an important topic. In other words, optimizing digital pre-distortion compensation technology to reduce costs, complexity and design difficulty while ensuring system performance is an urgent problem to be solved.

In view of the problems of the prior art, the purpose of the present invention is to use a simple structure to perform digital pre-distortion processing on the output signal, and to use a method with lower computational complexity to process the digital pre-distortion problem. Moreover, during the processing, the digital pre-distortion processing of the nonlinear effect of the power amplifier can be solved without changing the phase of the array antenna or adjusting the gain of the power amplifier, so as to facilitate the combination and application of the multi-input and multi-output antenna transceiver system to suppress the nonlinear problem of multiple antennas.

According to the purpose of the present invention, a digital predistortion processing system for beamforming signals is provided, comprising a first digital predistortion processing unit, a second digital predistortion processing unit, a beamforming unit, a receiving unit, a filtering unit and an estimating unit. The first digital predistortion processing unit receives an original signal and adds a digital predistortion parameter to the original signal to generate a digital predistortion signal. The beamforming unit is connected to the digital predistortion processing unit to receive the digital predistortion signal and forms a beam-steering output signal through the beamforming unit. The receiving unit receives the beam-steering output signal to form a beam-steering received signal. The filtering unit is connected to the receiving unit, the estimating unit and the second digital predistortion processing unit. The filtering unit receives the beam-steering received signal and performs bandwidth-limited sampling on the beam-steering received signal to generate a beam-steering band-limited received signal with a bandwidth-limited factor. The estimation unit and the second digital pre-distortion processing unit each receive the beam-steering band-limited received signal. The estimation unit calculates the beam-steering band-limited received signal and the error signal to generate an estimated pre-distortion parameter. The second digital pre-distortion processing unit generates an estimated pre-distortion signal from the estimated pre-distortion parameter and the beam-steering band-limited received signal. After the estimation unit and the second digital pre-distortion processing unit perform a predetermined number of iterations in this way, the estimated pre-distortion parameters calculated by the second digital pre-distortion processing unit in the last iteration are used as new digital pre-distortion parameters, and the new digital pre-distortion parameters are transmitted to the first digital pre-distortion processing unit as digital pre-distortion parameters.

Among them, the connection relationship between the first digital pre-distortion processing unit, the second digital pre-distortion processing unit, the beamforming unit, the receiving unit, the filtering unit and the estimation unit forms an indirect learning framework.

上述公式使用記憶多項式(Memory Polynomial,MP)來做為數學模型。其中，

表示估算預失真參數，

表示估算預失真訊號，ω(w)表示帶限函數，以帶限函數作為過濾單元，W表示帶限函數長度，K為多項式階數，P為記憶深度，n為第幾次的迭代，且n由0開始，迭代次數為N，第一次迭代的估算預失真訊號表示為

，而最後一次迭代的估算預失真訊號表示為

，y _BO 為波束導向接收訊號。 The second digital pre-distortion processing unit calculates an estimated pre-distortion processing formula for estimating the pre-distortion signal in each iteration, as shown below:

The above formula uses memory polynomial (MP) as a mathematical model.

represents the estimated predistortion parameters,

represents the estimated predistortion signal, ω ( w ) represents the band-limited function, the band-limited function is used as the filtering unit, W represents the length of the band-limited function, K is the polynomial order, P is the memory depth, n is the number of iterations, and n starts from 0, the number of iterations is N, and the estimated predistortion signal of the first iteration is expressed as

, and the estimated predistortion signal of the last iteration is expressed as

, y _BO is the beam-steering receiving signal.

其中

為估算預失真訊號矩陣。[Y _{BO_BL} ]為以帶限記憶多項式模型為基底之波束導向帶限接收訊號基函數矩陣。

則為每一次迭代循環中的最後 一次的估算預失真參數的矩陣。又，估算預失真訊號矩陣由各次迭代的估算預失真訊號的轉置矩陣，可表示如下公式：

其中N為總迭代次數，由於(0)~(N-1)的數量等於N，

~

為各次迭代後由數位預失真處理單元在各次迭代運算出估算預失真訊號，又波束導向帶限接收訊號基函數矩陣[Y _{BO_BL} ]，如下公式所示：[Y _{BO_BL} ]=[[y _{10,BO_BL} ],…,[y _{k0,BO_BL} ],…,[y _{1p,BO_BL} ],…,[y _{kp,BO_BL} ]] ^T 其中[y _{kp,BO_BL} ]為各次迭代的波束導向帶限接收訊號矩陣，如下公式所示： Among them, the estimated pre-distortion processing formula can be rewritten into a matrix form, as shown in the following formula:

in

is the estimated predistortion signal matrix. [ Y _{BO_BL} ] is the beam-steering band-limited received signal basis function matrix based on the band-limited memory polynomial model.

Then is the matrix of the last estimated predistortion parameters in each iteration cycle. In addition, the estimated predistortion signal matrix is the transposed matrix of the estimated predistortion signal of each iteration, and can be expressed as follows:

Where N is the total number of iterations. Since the number of (0)~(N-1) is equal to N,

~

The estimated predistortion signal is calculated by the digital predistortion processing unit after each iteration, and the beam-steering band-limited received signal basis function matrix [ Y _{BO_BL} ] is as shown in the following formula: [ Y _{BO_BL} ]=[[ y10 _{,BO_BL} ] , … , [ yk0 _{, BO_BL} ] , … , [ y1p _{,BO_BL} ] , … , [ ykp _{,BO_BL} ]] ^Twhere [ ykp _{,BO_BL} ] is the beam-steering band-limited received signal matrix of each iteration, as shown in the following formula:

又，每一次迭代循環中最後一次的估算預失真參數矩陣

，如下公式所示：

其中，

為每一次迭代循環中最後一次的估算預失真參數。 [ y _{kp,BO_BL} ]=[ y _{kp,BO_BL} (0) , ............ ,y _{kp,BO_BL} ( N -1)] ^T where y _{kp,BO _ BL} ( n ) represents the estimated beam-steering received signal after the estimated pre-distortion parameters of the beam-steering band-limited received signal are proposed, as shown in the following formula:

Again, the last estimated predistortion parameter matrix in each iteration cycle is

, as shown in the following formula:

in,

is the last estimated predistortion parameter in each iteration cycle.

其中μ為0~1的任意數，[e _{BO_BL} ]為誤差訊號矩陣，如下公式所示：

[z _{BO_BL} ]為數位預失真訊號矩陣，如下公式所示： [z _{BO_BL} ]=[z _{BO_BL} (0),............,z _{BO_BL} (N-1)] T

為估算單元運算出的在n次迭代循環中的各次迭代運算出的估算預失真參數矩陣，如下公式所示：

其中，

為在n次迭代循環中各次迭代的估算預失真參數。 The estimation unit estimates the pre-distortion parameter vector information using a damped Newton algorithm, as shown in the following formula:

Where μ is an arbitrary number between 0 and 1, and [ e _{BO_BL} ] is the error signal matrix, as shown in the following formula:

[ z _{BO_BL} ] is the digital pre-distortion signal matrix, as shown in the following formula: [ z _{BO_BL} ]=[ z _{BO_BL} (0) , ............ ,z _{BO_BL} ( N -1)] T

The estimated predistortion parameter matrix calculated by the estimation unit in each iteration of the n iteration loops is shown in the following formula:

in,

is the estimated predistortion parameter for each iteration in the n iteration loops.

其中預失真訊號矩陣可改寫為，如下公式所示：[z _{BO_BL} ]=[z _{BO_BL} (0),............,z _{BO_BL} (N-1)] ^T 其中[X _{BO_BL} ]為以記憶多項式的數學模型為基底之原始訊號矩陣的基函數轉置矩陣，如下公式所示：[X _{BO_BL} ]=[[X _{10,BO_BL} ],…,[X _{K0,BO_BL} ],…,[X _{1P,BO_BL} ],…,[X _{KP,BO_BL} ]] ^T 而原始訊號矩陣[X _{kp,BO_BL} ]，可以由下列公式表示：[X _{kp,BO_BL} ]=[x _{kp,BO_BL} (0),............,x _{kp,BO_BL} (N-1)] ^T 又，x _kp,BOBL (n)表示預失真訊號矩陣提出預失真參數，可以由下列公式表示：

The first digital pre-distortion processing unit uses the pre-distortion parameter matrix to simulate the pre-distortion signal matrix after the pre-distortion parameter processing, as shown in the following formula:

The predistorted signal matrix can be rewritten as shown in the following formula: [ z _{BO_BL} ]=[ z _{BO_BL} (0) , ............ ,z _{BO_BL} ( N -1)] ^Twhere [ X _{BO_BL} ] is the basis function transpose matrix of the original signal matrix based on the mathematical model of memory polynomials, as shown in the following formula: [ X _{BO_BL} ]=[[ X _{10 ,BO_BL} ] , … , [ X _{K 0 ,BO_BL} ] , … , [ X _{1 P,BO_BL} ] , … , [ X _{KP,BO_BL} ]] ^Tand the original signal matrix [ X _{kp,BO_BL} ] can be expressed by the following formula: [ X _{kp,BO_BL} ]=[ x _{kp,BO_BL} (0) , ............ ,x _{kp,BO_BL} ( N -1)] ^Tand , x _kp,BOBL ( n ) represents the predistortion signal matrix. The predistortion parameters can be expressed by the following formula:

The beamforming unit includes a plurality of phase shifters, a plurality of power amplifiers and a plurality of transmitting antennas. Each power amplifier is connected to one of the plurality of phase shifters and one of the plurality of transmitting antennas. The beamforming unit receives the predistortion signal from each phase shifter, and changes the phase angle of the predistortion signal with its own phase angle, and outputs a predistortion signal with a phase difference. The predistortion signal with a phase difference is defined as a phase difference signal. Each power amplifier receives its own phase difference signal, and outputs a power-amplified phase difference signal after power amplification. The power-amplified phase difference signal is defined as a power-amplified phase difference signal. Each power amplifier transmits its own power-amplified phase difference signal to its own connected transmitting antenna. Each transmitting antenna radiates its own power-amplified phase difference signal into the air as an electromagnetic wave, and forms a beam-guided output signal in the air.

表示，每一發射天線將各自的功率放大相位差訊號以y _i (n)表示，各次的波束導向接收訊號則以y _BO (n)表示，並可進一步以下列公表示彼此的關係： y = z ． a ． a ^H

y =[y _BO (0),............,y _BO (N-1)] ^T z 為進入到各路相移器的原始訊號

轉置矩陣，矩陣維度為N×1。 y 為波束導向接收訊號的轉置矩陣，矩陣維度為N×1，與原訊號維度相同。而 a 矩陣維度為1×I， a ^H 矩陣維度為I×1， a ^H 表示追蹤向量的共軛轉置，且 a ^H =

，同時他也表示了各發射天線與接收單元之相對相位， a =[W ₁ W ₂…W _i ]則表示各相移器所需改變之相位。 The original signal is represented by x , the original signal of each iteration cycle is represented by x ( n ), the digital predistortion signal generated by the original signal of each iteration cycle after passing through the first digital predistortion processing unit is represented by zBO_BL (n ) , the number of transmitting antennas is represented by i , _and the digital predistortion signal entering each phase shifter is represented by

Each transmitting antenna amplifies its own power phase difference signal as y _i ( n ), and each beam-steering receiving signal is represented by y _BO ( n ), and the relationship between them can be further expressed by the following formula: y = z ． a ． a ^H

y =[ y _BO (0) , ............ ,y _BO ( N -1)] ^T z is the original signal entering each phase shifter

Transpose matrix, the matrix dimension is N ×1. y is the transpose matrix of the beam-steering received signal, the matrix dimension is N ×1, the same as the original signal dimension. The a matrix dimension is 1× I , and the a ^H matrix dimension is I ×1. a ^H represents the conjugate transpose of the tracking vector, and a ^H =

At the same time, he also expressed the relative phase of each transmitting antenna and the receiving unit, _and a = [ W ₁ W ₂ … Wi ] represents the phase that each phase shifter needs to change.

According to the purpose of the present invention, a digital predistortion processing method for a beamforming signal is provided, which is applied to the digital predistortion processing system for a beamforming signal, and includes the following steps: a first digital predistortion processing unit receives an original signal and adds a digital predistortion parameter to the original signal to generate a digital predistortion signal; a beamforming unit receives the digital predistortion signal and forms a beam-steering output signal through the beamforming unit; a receiving unit receives the beam-steering output signal to form a beam-steering band-limited received signal; a filtering unit receives the beam-steering band-limited received signal and performs bandwidth-limited sampling on the beam-steering band-limited received signal to generate a beam-steering band-limited received signal with a bandwidth-limited factor; an estimating unit and a second digital predistortion processing unit receive the digital predistortion signal and form a beam-steering output signal through the beamforming unit; a receiving unit receives the beam-steering output signal to form a beam-steering band-limited received signal; a filtering unit receives the beam-steering band-limited received signal and performs bandwidth-limited sampling on the beam-steering band-limited received signal to generate a beam-steering band-limited received signal with a bandwidth-limited factor; and a filtering unit receives the beam-steering band-limited received signal and performs bandwidth-limited sampling on the beam-steering band-limited received signal. The first embodiment of the present invention is to receive the beam-steering band-limited received signal and perform an indirect learning step. In the indirect learning step, the estimation unit generates an estimated pre-distortion parameter by calculating the beam-steering band-limited received signal and the error signal. The second digital pre-distortion processing unit generates an estimated pre-distortion signal from the estimated pre-distortion parameter and the beam-steering band-limited received signal. The estimation unit and the second digital pre-distortion processing unit cycle in this manner. After calculating the predetermined number of iterations, the second digital pre-distortion processing unit performs the last iteration as the new digital pre-distortion parameter, completing the indirect learning step. The second digital pre-distortion processing unit transmits the new digital pre-distortion parameter to the first digital pre-distortion processing unit as the digital pre-distortion parameter, and the first digital pre-distortion processing unit then processes it according to the steps from receiving the original signal.

As mentioned above, the digital pre-distortion technology of the present invention overcomes the nonlinear effects of multiple power amplifiers and is applied in the MIMO system to solve the problem that the application of pre-distortion technology in the existing MIMO communication system causes too many compensation processing times, resulting in the gradual increase in system complexity. In addition, the present invention fully considers the improvement of future signal specifications and provides a solution to reduce costs and complexity for pre-distortion technology, thereby significantly enhancing the pre-distortion processing technology in multi-antenna systems. Compared with the active array antenna pre-distortion technology proposed in the previous patents, the unique pre-distortion architecture of the present invention is significantly different from the existing technology, especially in the context of the ever-changing communication bandwidth in the future, showing better results.

1: The first digital pre-distortion processing unit

2: Second digital pre-distortion processing unit

3: Beamforming unit

4: Receiving unit

5: Filter unit

6: Estimation unit

30: Phase shifter

32: Power amplifier

34: Transmitting antenna

S101~S107, S1051~S1052: Process steps

Figure 1 is a schematic diagram of the architecture of the digital pre-distortion processing system for beamforming signals of the present invention.

Figure 2 is a schematic diagram of the process of the digital pre-distortion processing method of the beamforming signal of the present invention.

FIG3 is a signal simulation diagram of the digital pre-distortion processing of the beamforming signal of the present invention.

The embodiments of the present invention will be further explained below with the help of the relevant drawings. As far as possible, the same reference numerals in the drawings and the specification represent the same or similar components. In the drawings, the shapes and thicknesses may be exaggerated for the sake of simplicity and convenience. It is understood that the components not specifically shown in the drawings or described in the specification have the form known to the ordinary technicians in the relevant technical field. The ordinary technicians in this field can make various changes and modifications based on the content of the present invention.

As shown in FIG1 , the present invention is a digital predistortion processing system for beamforming signals, comprising a first digital predistortion processing unit 1, a second digital predistortion processing unit 2, a beamforming unit 3, a receiving unit 4, a filtering unit 5, and an estimating unit 6. The connection relationship between the first digital predistortion processing unit 1, the second digital predistortion processing unit 2, the beamforming unit 3, the receiving unit 4, the filtering unit 5, and the estimating unit 6 forms an indirect learning architecture.

In the present invention, the first digital predistortion processing unit 1 receives the original signal and adds the digital predistortion parameter to the original signal to generate a digital predistortion signal. The beamforming unit 3 is connected to the digital predistortion processing unit to receive the digital predistortion signal and form a beam-guided output signal through the beamforming unit 3. The receiving unit 4 receives the beam-guided output signal to form a beam-guided receiving signal.

表示，每一個發射天線34將各自的功率放大相位差訊號以y _i (n)表示，各次的波束導向帶限接收訊號則以y _BO (n)表示，並可進一步以下列公表示彼此的關係： y = z ． a ． a ^H

y =[y _BO (0),............,y _BO (N-1)] ^T z 為進入到各路相移器30的原始訊號

轉置矩陣，矩陣維度為N×1。 y 為波束導向帶限接收訊號轉置矩陣，矩陣維度為N×1，與原訊號維度相同。而 a 矩陣維度為1×I， a ^H 矩陣維度為I×1， a ^H 表示追蹤向量的共軛轉置且 a ^H =

，同時他也表示了各發射天線34與接收單元4之相對相位， a =[W ₁ W ₂…W _i ]則表示各相移器30所需改變之相位。 In the present invention, the beamforming unit 3 includes a plurality of phase shifters 30, a plurality of power amplifiers 32 and a plurality of transmitting antennas 34. Each power amplifier 32 is connected to one of the plurality of phase shifters 30 and one of the plurality of transmitting antennas 34. The original signal is represented by x , and each original signal is represented by x ( n ). The digital predistortion signal generated by the original signal x ( n ) after passing through the first digital predistortion processing unit 1 is represented _{by zBO_BL} ( n ). The number of transmitting antennas 34 of the beamforming unit 3 is represented by i . The digital predistortion signal entering each phase shifter 30 is represented by

Each transmitting antenna 34 amplifies its own power phase difference signal as y _i ( n ), and each beam-steering band-limited receiving signal is represented by y _BO ( n ), and the relationship between them can be further represented by the following formula: y = z ． a ． a ^H

y =[ y _BO (0) , ............ ,y _BO ( N -1)] ^T z is the original signal entering each phase shifter 30

Transpose matrix, the matrix dimension is N ×1. y is the transpose matrix of the beam-steering band-limited received signal, the matrix dimension is N ×1, the same as the original signal dimension. The a matrix dimension is 1× I , and the a ^H matrix dimension is I ×1, a ^H represents the conjugate transpose of the tracking vector and a ^H =

At the same time, it also represents the relative phase between each transmitting antenna 34 and the receiving unit 4, _and a = [ W ₁ W ₂ … Wi ] represents the phase that each phase shifter 30 needs to change.

，而是在第二次以後的原始訊號x(1)~x(N-1)，第一數位預失真處理單元1才會有數位預失真參數，但是為了 便於說明，在本案中各次的原始訊號x(n)經過數位預失真處理單元後，數位預失真處理單元的輸出統一稱為預失真訊號

。 It should be noted that when the original signal x (0) is input to the first digital pre-distortion processing unit 1, the first digital pre-distortion processing unit 1 does not have digital pre-distortion parameters. Therefore, each phase shifter 30 receives

, but after the second original signal x (1) ~ x ( N -1), the first digital pre-distortion processing unit 1 will have digital pre-distortion parameters. However, for the sake of convenience, in this case, after the original signal x ( n ) passes through the digital pre-distortion processing unit, the output of the digital pre-distortion processing unit is uniformly called the pre-distortion signal

.

，並分別以各自的相位角改變預失真訊號的相位角，而輸出具相位差的預失真訊號，在此將具相位差的預失真訊號定義為相位差訊號，並以

表示，各功率放大器32分別接收各自相位差訊號，並將相位差訊號進行功率放大後輸出功率放大後的相位差訊號，在此將功率放大後的相位差訊號定義為功率放大相位差訊號，並以y _i (n)表示，各功率放大器32將各自的功率放大相位差訊號y _i (n)傳送到各自連接的發射天線34，每一個發射天線34將各自的功率放大相位差訊號y _i (n)以電磁波輻射到空中，並在空中形成波束導向輸出訊號y _BO (n)。 In the present invention, each phase shifter 30 receives a predistortion signal

, and change the phase angle of the predistortion signal with their respective phase angles, and output a predistortion signal with a phase difference. Here, the predistortion signal with a phase difference is defined as a phase difference signal, and

It indicates that each power amplifier 32 receives its own phase difference signal respectively, and outputs the power-amplified phase difference signal after power amplification. The power-amplified phase difference signal is defined as the power-amplified phase difference signal and represented by y _i ( n ). Each power amplifier 32 transmits its own power-amplified phase difference signal y _i ( n ) to its own connected transmitting antenna 34. Each transmitting antenna 34 radiates its own power-amplified phase difference signal y _i ( n ) into the air in the form of electromagnetic waves, and forms a beam-guided output signal y _BO ( n ) in the air.

。 In the present invention, the filtering unit 5 is connected to the receiving unit 4, the estimating unit 6 and the second digital pre-distortion processing unit 2. The filtering unit 5 receives each beam-steering receiving signal y _BO ( n ), and performs bandwidth-limited sampling on each beam-steering receiving signal y _BO ( n ) to generate each beam-steering band-limited receiving signal y _{BO_BL} ( n ) with a bandwidth-limited factor. The estimating unit 6 and the second digital pre-distortion processing unit 2 receive each beam-steering band-limited receiving signal y _{BO_BL} ( n ). The estimating unit 6 calculates the beam-steering band-limited receiving signal y _{BO_BL} ( n ) and the error signal e ( n ) to generate an estimated pre-distortion parameter.

.

及波束導向帶限接收訊號y _{BO_BL} (n)產生估算預失真訊號

。估算單元6及第二數位預失真處理單元2依此循環運算預定迭代次數後，第二數位預失真處理單元2最後一次迭代運算的估算預失真參數

作為新數位預失真參數

，並傳送新數位預失真參數

到第一數位預失真處理單元1，作為數位預失真參數a _{kp,BO_BL} 。接收單元4為接收天線，過濾單元5為帶通濾波 器，但並不限於此，也可以是由一個高通濾波器及一個低通濾波器所組成，而且波束導向輸出訊號y _BO (n)先經過高通濾波器處理在經過低通濾波器處理，相當於帶通濾波器的訊號處理方式。 Furthermore, the second digital predistortion processing unit 2 estimates the predistortion parameters

and the beam-steering band-limited received signal y _{BO_BL} ( n ) to generate an estimated predistortion signal

After the estimation unit 6 and the second digital pre-distortion processing unit 2 perform the predetermined iterations in a loop, the estimated pre-distortion parameters of the last iteration of the second digital pre-distortion processing unit 2 are

As new digital pre-distortion parameters

, and send new digital predistortion parameters

To the first digital predistortion processing unit 1 as a digital predistortion parameter a _{kp,BO_BL} . The receiving unit 4 is a receiving antenna, and the filtering unit 5 is a bandpass filter, but is not limited thereto, and may also be composed of a high-pass filter and a low-pass filter, and the beam steering output signal y _BO ( n ) is first processed by the high-pass filter and then by the low-pass filter, which is equivalent to the signal processing method of the bandpass filter.

上述公式使用記憶多項式(Memory Polynomial,MP)來做為數學模型。其中，

表示估算預失真參數，

表示估算預失真訊號，ω(w)表示帶限函數，以帶限函數作為過濾單元5，w表示帶限函數長度，K為多項式階數，P為記憶深度，n為第幾次的迭代，且n由0開始，迭代次數為N，第一次迭代的估算預失真訊號表示為

，而最後一次迭代的估算預失真訊號表示為

，y _BO 為波束導向接收訊號。 In the present invention, the second digital pre-distortion processing unit 2 calculates an estimated pre-distortion processing formula for estimating the pre-distortion signal in each iteration, as shown below:

The above formula uses memory polynomial (MP) as a mathematical model.

represents the estimated predistortion parameters,

represents the estimated predistortion signal, ω ( w ) represents the band-limited function, the band-limited function is used as the filter unit 5, w represents the length of the band-limited function, K is the polynomial order, P is the memory depth, n is the number of iterations, and n starts from 0, the number of iterations is N, and the estimated predistortion signal of the first iteration is expressed as

, and the estimated predistortion signal of the last iteration is expressed as

, y _BO is the beam-steering receiving signal.

其中

為估算預失真訊號矩陣。[Y _{BO_BL} ]為以記憶多項式模型為基底之波束導向帶限接收訊號的基函數矩陣。

則為每一次迭代循環中最後一次的估算預失真參數矩陣。又，估算預失真訊號矩陣

為各次迭代的估算預失真訊號

的轉置矩陣，可表示如下公式：

其中

~

為各次迭代後由數位預失真處理單元在各次迭代運算出估算預失真訊號

。又波束導向帶限接收訊號基函數矩陣[Y _{BO_BL} ]，如下公式所示：[Y _{BO_BL} ]=[[y _{10,BO_BL} ],…,[y _{k0,BO_BL} ],…,[y _{1p,BO_BL} ],…,[y _{kp,BO_BL} ]] ^T 其中[y _{kp,BO_BL} ]為各次迭代的波束導向帶限接收訊號矩陣，如下公式所示：[y _{kp,BO_BL} ]=[y _{kp,BO_BL} (0),............,y _{kp,BO_BL} (N-1)] ^T 其中y _{kp,BO_BL} (n)表示波束導向帶限接收訊號提出估算預失真參數

後之估算波束導向接收訊號，如下公式所示：

又，每一次迭代循環中最後一次的估算預失真參數矩陣

，如下公式所示：

其中，

為各次迭代循環中的最後一次的估算預失真參數。 Among them, the estimated pre-distortion processing formula can be rewritten into a matrix form, as shown in the following formula:

in

is the estimated predistortion signal matrix. [ Y _{BO_BL} ] is the basis function matrix of the beam-steering band-limited received signal based on the memory polynomial model.

is the last estimated predistortion parameter matrix in each iteration cycle.

is the estimated predistortion signal for each iteration

The transposed matrix of can be expressed as follows:

in

~

After each iteration, the digital pre-distortion processing unit calculates the estimated pre-distortion signal

. The beam-steering band-limited received signal basis function matrix [ Y _{BO_BL} ] is as shown in the following formula: [ Y _{BO_BL} ]=[[ y _{10 ,BO_BL} ] , … , [ y _{k0 ,BO_BL} ] , … , [ y _{1 p,BO_BL} ] , … , [ y _{kp,BO_BL} ]] ^Twherein [ y _{kp,BO_BL} ] is the beam-steering band-limited received signal matrix of each iteration, as shown in the following formula: [ y _{kp,BO_BL} ]=[ y _{kp,BO_BL} (0) , ............ ,y _{kp,BO_BL} ( N -1)] ^Twherein y _{kp,BO_BL} ( n ) represents the estimated pre-distortion parameters proposed for the beam-steering band-limited received signal.

The estimated beam-steering received signal is shown in the following formula:

Again, the last estimated predistortion parameter matrix in each iteration cycle is

, as shown in the following formula:

in,

is the estimated predistortion parameter for the last time in each iteration cycle.

其中μ為0~1的任意數，[e _{BO_BL} ]為誤差訊號矩陣，如下公式所示：

[z _{BO_BL} ]為數位預失真訊號矩陣，如下公式所示：[z _{BO_BL} ]=[z _{BO_BL} (0),............,z _{BO_BL} (N-1)] ^T

為在n次迭代循環中的各次運算出的預失真參數的矩陣，如下公式所示：

其中，

為在n次迭代循環中各次迭代的該估算預失真參數。 The estimation unit 6 estimates the pre-distortion parameter vector information using a damped Newton algorithm, as shown in the following formula:

Where μ is an arbitrary number between 0 and 1, and [ e _{BO_BL} ] is the error signal matrix, as shown in the following formula:

[ z _{BO_BL} ] is the digital pre-distortion signal matrix, as shown in the following formula: [ z _{BO_BL} ]=[ z _{BO_BL} (0) , ............ ,z _{BO_BL} ( N -1)] ^T

is the matrix of predistortion parameters calculated in each of the n iteration loops, as shown in the following formula:

in,

is the estimated predistortion parameter for each iteration in the n iteration loops.

The first digital pre-distortion processing unit 1 uses the pre-distortion parameter matrix to simulate the pre-distortion signal matrix after the pre-distortion parameter processing, as shown in the following formula: [ z _{BO_BL} ]=[ X _{BO_BL} ][ a _{BO_BL} ] The pre-distortion signal matrix can be rewritten as shown in the following formula: [ z _{BO_BL} ]=[ z _{BO_BL} (0) , ............ ,z _{BO_BL} ( N -1)] ^T Wherein [ X _{BO_BL} ] is the basis function transposed matrix of the beam-steering band-limited original signal matrix based on the mathematical model of the memory polynomial, as shown in the following formula: [ X _{BO_BL} ]=[[ X _{10 ,BO_BL} ] , … , [ X _{k 0 ,BO_BL} ] , … , [ X _{1 P,BO_BL} ] , … , [ X _{kp,BO_BL} ]] ^TThe beam-steering band-limited original signal matrix [ X _{kp,BO_BL} ] can be expressed by the following formula: [ X _{kp,BO_BL} ] = [ x _{kp,BO_BL} (0) , ............ ,x _{kp,BO_BL} ( N -1)] ^TAnd , x _{kp,BO_BL} ( n ) is the original signal for proposing pre-distortion parameters, which can be expressed by the following formula:

As shown in FIG2 , another digital predistortion processing method for beamforming signals of the present invention is applied to the aforementioned digital predistortion processing system for beamforming signals, and includes the following steps: (S101) a first digital predistortion processing unit 1 receives an original signal and adds a digital predistortion parameter to the original signal to generate a digital predistortion signal; (S102) a beamforming unit 3 receives the digital predistortion signal and forms a beam-guided output signal through the beamforming unit 3; (S103) receiving The unit 4 receives the beam-steering output signal to form a beam-steering band-limited received signal; (S104) the filtering unit 5 receives the beam-steering band-limited received signal and performs bandwidth-limited sampling on the beam-steering band-limited received signal to generate a beam-steering band-limited received signal with a bandwidth-limited factor; (S105) the estimation unit 6 and the second digital pre-distortion processing unit 2 each receive the beam-steering band-limited received signal and perform an indirect learning step, wherein the indirect learning step includes the following steps: (S1051) The estimation unit 6 calculates the beam-steering band-limited received signal and the error signal to generate an estimated pre-distortion parameter; (S1052) the second digital pre-distortion processing unit 2 generates an estimated pre-distortion signal from the estimated pre-distortion parameter and the beam-steering band-limited received signal; the estimation unit 6 and the second digital pre-distortion processing unit 2 calculate a predetermined number of iterations according to (S1051) to (S1052), for example, the predetermined number of iterations is any integer between 2 and 100. In some embodiments of the present invention, The predetermined number of iterations is 10. The second digital predistortion processing unit 2 performs the last iteration operation as the new digital predistortion parameter, which completes the indirect learning step; (S106) the second digital predistortion processing unit 2 transmits the new digital predistortion parameter to the first digital predistortion processing unit 1; (S107) the first digital predistortion processing unit 1 uses the new digital predistortion parameter as the digital predistortion parameter; and then the first digital predistortion processing unit 1 continues to process according to the steps from receiving the original signal.

As shown in FIG. 3 , the original signal is represented by the green line. When the original signal is amplified by the power amplifier 32, a nonlinear distortion signal is formed due to nonlinear distortion, as represented by the red line. However, the predistortion signal generated by the nonlinear distortion signal after the beam steering band-limiting processing of the present invention is represented by the blue line. It can be observed that the present invention has indeed solved the problem of nonlinear distortion. In addition, it should be noted that the present invention can also be applied to multiple beamforming units 3, not limited to a single beamforming unit 3. For example, when the traditional pre-distortion processing method is applied to eight beamforming units 3, and each beamforming unit 3 has eight array antennas, 64 pre-distortion processings are required, but the present invention only requires eight pre-distortion processings. Furthermore, the present invention does not process the output signal multiple times. The present invention simplifies the processing method of the nonlinear distortion problem after filtering the processed received signal, so that the pre-distortion processing speed becomes faster. In addition, the present invention uses the received signal to solve the nonlinear distortion problem, which can truly solve the nonlinear distortion problem.

The present invention uses digital pre-distortion technology to overcome the nonlinear effects of multiple power amplifiers 32. In particular, when applied to MIMO communication systems, it can solve the problem that the compensation processing times in MIMO communication systems are too many, resulting in the pre-distortion processing system architecture being too complicated. In addition, the present invention fully considers future signal specifications and provides a solution to reduce the cost and complexity of pre-distortion technology, thereby significantly enhancing the pre-distortion processing technology in MIMO communication systems. Compared with the pre-distortion technology in the patent case described in the previous technology, the unique pre-distortion architecture of the present invention is significantly different, especially in the context of the ever-changing communication bandwidth in the future, it shows better efficacy.

The above is only an example to illustrate the preferred implementation of the present invention, and is not intended to limit the scope of implementation. All simple substitutions and equivalent changes made according to the scope of the patent application of the present invention and the content of the patent specification are within the scope of the patent application of the present invention.

1: The first digital pre-distortion processing unit

2: Second digital pre-distortion processing unit

3: Beamforming unit

4: Receiving unit

5: Filter unit

6: Estimation unit

30: Phase shifter

32: Power amplifier

34: Transmitting antenna

Claims (8)

A digital predistortion processing system for a beamforming signal comprises: a first digital predistortion processing unit, receiving an original signal and adding a digital predistortion parameter to the original signal to generate a digital predistortion signal; a beamforming unit, connected to the digital predistortion processing unit, receiving the digital predistortion signal and forming a beam-steering output signal through the beamforming unit; a receiving unit, receiving the beam-steering output signal to form a beam-steering received signal; a filtering unit, connected to the receiving unit, receiving the beam-steering received signal and performing bandwidth-limited sampling on the beam-steering received signal to generate a beam-steering band-limited received signal with a bandwidth-limited factor; an estimating unit, connected to the filtering unit, A unit, the estimation unit generates an estimated pre-distortion parameter by operating the beam-steering band-limited received signal and an error signal; and a second digital pre-distortion processing unit, connected to the filtering unit and the estimation unit, the second digital pre-distortion processing unit generates an estimated pre-distortion signal from the estimated pre-distortion parameter and the beam-steering band-limited received signal; wherein, after the estimation unit and the second digital pre-distortion processing unit cyclically calculate the estimated pre-distortion parameter and the estimated pre-distortion signal of each iteration according to a predetermined number of iterations, the estimated pre-distortion parameter calculated by the second digital pre-distortion processing unit in the last iteration is used as a new digital pre-distortion parameter, and the new digital pre-distortion parameter is transmitted to the first digital pre-distortion processing unit as the digital pre-distortion parameter. In the digital predistortion processing system for beamforming signals as claimed in claim 1, the second digital predistortion processing unit calculates the estimated predistortion signal in each iteration using an estimated predistortion processing formula, and the estimated predistortion processing formula is as follows:

The estimated predistortion processing formula uses memory polynomials as a mathematical model, where

represents the estimated predistortion parameters,

represents the estimated predistortion signal, ω ( w ) represents a band-limited function, the band-limited function is used as the filtering unit, W represents the length of the band-limited function, K is the polynomial order, P is the memory depth, n is the number of iterations, and n starts from 0, the number of iterations is N, and the estimated predistortion signal of the first iteration is expressed as

, and the estimated predistortion signal of the last iteration is expressed as

1), y _BO is the beam-guided receiving signal. In the digital predistortion processing system for beamforming signals as described in claim 2, the estimated predistortion processing formula can be rewritten in matrix form as shown in the following formula:

in

is the matrix of the estimated predistortion signal, called the estimated predistortion signal matrix, [ Y _{BO_BL} ] is the basis function matrix of the beam-steering band-limited received signal based on the memory polynomial model, called the beam-steering band-limited received signal basis function matrix,

The matrix of the estimated predistortion parameters at the last time in each iteration cycle is called the estimated predistortion parameter matrix. The estimated predistortion signal matrix is composed of the transposed matrix of the estimated predistortion signal of each iteration, which can be expressed as follows:

in

~

The estimated pre-distortion signal is calculated by the digital pre-distortion processing unit after each iteration, and the beam-steering band-limited received signal basis function matrix [ Y _{BO_BL} ] is as shown in the following formula: [ Y _{BO_BL} ]=[[ y _{10 ,BO_BL} ] , … , [ y _{k0 ,BO_BL} ] , … , [ y _{1 p,BO_BL} ] , … , [ y _{kp,BO_BL} ]] ^T ; wherein [ y _{kp,BO_BL} ] is the matrix of the beam-steering band-limited received signal of each iteration, called the beam-steering band-limited received signal matrix, as shown in the following formula: [ y _{kp,BO_BL} ]=[ y _{kp,BO_BL} (0) , ............ ,y _{kp,BO_BL} ( N -1)] ^T ; wherein y _{kp,BO_BL} ( n ) represents an estimated beam-steering received signal after the estimated pre-distortion parameter is obtained from the beam-steering band-limited received signal, as shown in the following formula:

Again, the estimated predistortion parameter matrix at the end of each iteration cycle is as shown in the following formula:

in,

is the estimated predistortion parameter of the last iteration. In the digital predistortion processing system for beamforming signals as described in claim 3, the estimation unit estimates the vector information of the predistortion parameters using a damped Newton algorithm, as shown in the following formula:

Where μ is an arbitrary number between 0 and 1, [ e _{BO_BL} ] is the matrix of the error signal, called the error signal matrix, as shown in the following formula:

[ z _{BO_BL} ] is the matrix of the digital pre-distortion signal, called the digital pre-distortion signal matrix, as shown in the following formula: [ z _{BO_BL} ]=[ z _BOBL (0) , ............ ,z _BOBL ( N -1)] ^T ;

The matrix of the estimated pre-distortion parameters calculated by the estimation unit in each iteration in the n iteration loops is called the estimated pre-distortion parameter matrix, as shown in the following formula:

in,

is the estimated predistortion parameter for each iteration in the n iteration loops. In the digital predistortion processing system for beamforming signals as claimed in claim 4, the first digital predistortion processing unit uses the estimated predistortion parameter matrix calculated by the estimation unit to simulate a matrix of the predistortion signal after predistortion parameter processing, which is called a predistortion signal matrix, as shown in the following formula:

The predistortion signal matrix can be rewritten as shown in the following formula: [ z _{BO_BL} ]=[ z _{BO_BL} (0) , ............ ,z _{BO_BL} ( N -1)] ^T ; wherein [ X _{BO_BL} ] is a basis function transposed matrix of the beam-steering band-limited original signal matrix based on the mathematical model of the memory polynomial, called the beam-steering band-limited original signal basis function transposed matrix, as shown in the following formula: [ X _{BO_BL} ]=[[ X _{10 ,BO_BL} ] , … , [ X _{K 0 ,BO_BL} ] , … , [ X _{1 P,BO_BL} ] , … , [ X _{KP,BO_BL} ]] ^T ; and the beam-steering band-limited original signal matrix [ X _{kp,BO_BL} ] can be expressed by the following formula: [ X _{kp,BO_BL} ]=[ x _{kp,BO_BL} (0) , ............ ,x _{kp,BO_BL} ( N -1)] ^T ; and, x _{kp,BO_BL} ( n ) is the original signal for proposing the predistortion parameter, which can be expressed by the following formula:

A digital predistortion processing system for beamforming signals as described in claim 1, wherein the beamforming unit comprises: a plurality of phase shifters, each phase shifter receives a predistortion signal and changes the phase angle of the predistortion signal with its own phase angle to output a predistortion signal with a phase difference, wherein the predistortion signal with a phase difference is defined as a phase difference signal; a plurality of power amplifiers, each power amplifier is connected to one of the plurality of phase shifters, and each power amplifier receives a respective A phase difference signal is generated, and the phase difference signal is power-amplified and then outputted as a power-amplified phase difference signal. Here, the power-amplified phase difference signal is defined as a power-amplified phase difference signal; and a plurality of transmitting antennas, each of which is connected to one of the plurality of power amplifiers, and each transmitting antenna receives its own power-amplified phase difference signal. Each transmitting antenna radiates its own power-amplified phase difference signal into the air as an electromagnetic wave, and forms a beam-guided output signal in the air. The digital predistortion processing system for beamforming signals as claimed in claim 6, wherein the number of the transmitting antennas is represented by i , the digital predistortion signal is represented by z _{BO_BL} ( n ), and the digital predistortion signal entering each phase shifter is represented by

In other words, each transmitting antenna represents the power amplified phase difference signal connected to each other as y _i ( n ), and the beam-steering band-limited received signal as y _BO ( n ), and the relationship between the original signal and the beam-steering band-limited received signal can be further represented by the following formula: y = z ． a ． a ^H ;

y =[ y _BO (0) , ............ ,y _BO ( N -1)] ^T ; z is the original signal entering each phase shifter

Transpose matrix, the matrix dimension is N ×1, y is the transpose matrix of the beam-steering band-limited received signal, the matrix dimension is N ×1, which is the same as the original signal dimension, and the a matrix dimension is 1× I , a ^H matrix dimension is I ×1, a ^H represents the conjugate transpose of the tracking vector, and a ^H =

At the same time, it also represents the relative phase of each transmitting antenna and the receiving unit, and a = [ W ₁ W ₂ … W _I ] represents the phase of each phase shifter. A digital predistortion processing method for a beamforming signal is applied to the digital predistortion processing system of any one of claim items 1 to 7, comprising the following steps: the first digital predistortion processing unit receives the original signal and adds the original signal to the digital predistortion parameter to generate the digital predistortion signal; the beamforming unit receives the digital predistortion signal and forms the beam through the beamforming unit. The receiving unit receives the beam-steering output signal to form the beam-steering band-limited receiving signal; the filtering unit receives the beam-steering band-limited receiving signal and performs bandwidth-limited sampling on the beam-steering band-limited receiving signal to generate the beam-steering band-limited receiving signal with a bandwidth-limited factor; the estimation unit and the second digital pre-distortion processing unit each receive the beam-steering band-limited receiving signal. signal, and performs an indirect learning step, wherein the indirect learning step further includes: the estimation unit performs the operation with the beam-steering band-limited received signal and the error signal to generate the estimated pre-distortion parameter; the second digital pre-distortion processing unit generates the estimated pre-distortion signal from the estimated pre-distortion parameter and the beam-steering band-limited received signal; the estimation unit and the second digital pre-distortion processing unit cycle in this manner After calculating the predetermined number of iterations, the second digital pre-distortion processing unit performs the last iteration as the new digital pre-distortion parameter, completing the indirect learning step; the second digital pre-distortion processing unit transmits the new digital pre-distortion parameter to the first digital pre-distortion processing unit as the digital pre-distortion parameter, and the first digital pre-distortion processing unit then processes the signal according to the steps from receiving the original signal.