CN101764577A - Baseband pre-distortion power amplifier linearization method based on one-way feedback and non-iterative technique - Google Patents

一种基于单路反馈和非迭代技术的基带预失真功放线性化方法，属于电子技术领域，涉及功率放大器线性化技术。将两路正交的幅度线性增长的训练序列先后输入到功放，然后通过功放输出的I路反馈信号序列构建改进型的查找表；整个功放系统的输入信号先通过幅度计算、查找对应的查找表值，再分别对I、Q两路基带输入信号作相应的预示真处理，预失真处理后的输入信号经D/A转换、滤波和上变频后输到功放，即可得到与整个功放系统基带输入信号呈线性放大的输出信号。整个预失真处理过程可在数字处理芯片(包括DSP或FPGA)中实现，与现有技术相比，本发明无需反正切或者是反余弦运算，无需进行坐标变换，从而降低了计算量，提高了计算效率，其实现系统结构更为简单。

The invention relates to a baseband predistortion power amplifier linearization method based on single-channel feedback and non-iterative technology, which belongs to the field of electronic technology and relates to power amplifier linearization technology. Input the two-way orthogonal amplitude linearly increasing training sequence to the power amplifier successively, and then build an improved look-up table through the I-way feedback signal sequence output by the power amplifier; the input signal of the entire power amplifier system first passes through the amplitude calculation and finds the corresponding look-up table value, and then perform corresponding pre-distortion processing on the I and Q baseband input signals respectively. The input signal after pre-distortion processing is transmitted to the power amplifier after D/A conversion, filtering and up-conversion, and the baseband of the entire power amplifier system can be obtained. The input signal is linearly amplified to the output signal. The whole pre-distortion process can be realized in the digital processing chip (including DSP or FPGA). Compared with the prior art, the present invention does not need arctangent or arccosine operation, and coordinate transformation is not required, thereby reducing the amount of calculation and improving the Computational efficiency, its implementation system structure is simpler.

基于单路反馈和非迭代技术的基带预失真功放线性化方法 Linearization Method of Baseband Predistortion Power Amplifier Based on Single Feedback and Non-Iterative Technique

技术领域technical field

本发明属于电子技术领域，涉及功率放大器线性化技术，具体涉及一种基于数字基带预失真技术的宽带射频功率放大器线性化方法。The invention belongs to the field of electronic technology, and relates to a linearization technology of a power amplifier, in particular to a linearization method of a broadband radio frequency power amplifier based on a digital baseband predistortion technology.

技术背景technical background

理想化的功率放大器应该是一个线性器件，即对于一个输入信号x(t)＝r·e^jθ(r是输入信号的幅度，θ是输入信号的相位)，输出信号y(t)＝K·r·cos(ωt+θ)(K是理想线性功率放大器的增益，ω载波角频率)。然而，实际的功率放大器是一个非线性器件，对于一个输入信号x(t)＝re^jθ，其输出信号为y(t)＝B(r)·cos(ωt+θ+Φ(r))，其中B(r)是非线性功率放大器输出信号的幅度函数，Φ(r)是非线性功率放大器的附加相位函数(研究表明，非线性功率放大器的增益和附加相位均与输入信号的幅度r相关)。因此，非线性特性是功率放大器的一个固有特性。The idealized power amplifier should be a linear device, that is, for an input signal x(t)=r e ^jθ (r is the amplitude of the input signal, θ is the phase of the input signal), the output signal y(t)=K· r · cos (ωt + θ) (K is the gain of the ideal linear power amplifier, ω carrier angular frequency). However, the actual power amplifier is a nonlinear device. For an input signal x(t)=re ^jθ , its output signal is y(t)=B(r)·cos(ωt+θ+Φ(r)), Where B(r) is the amplitude function of the output signal of the nonlinear power amplifier, and Φ(r) is the additional phase function of the nonlinear power amplifier (research shows that both the gain and the additional phase of the nonlinear power amplifier are related to the amplitude r of the input signal). Therefore, nonlinearity is an inherent characteristic of power amplifiers.

随着移动通信技术迅速发展，无线通信频段变得越来越拥挤，频带资源越来越紧张，为了在有限的频谱范围内容纳更多的通信信道，人们提出了一些宽带数字传输技术(OFDM、WCDMA等)和高频谱利用率的调制方式(M-QAM等)。采用这些技术所传输的信号具有非恒定包络、宽频带和高峰平比等特点，当调制信号通过非线性的功率放大器后将产生带内和带外失真，输出信号频谱扩展，干扰邻近信道，增大通信系统误码率。因此，现代通信系统对射频功率放大器的线性度提出了很高的要求，功率放大器线性化技术已成为下一代无线通信系统的关键技术之一，成为了当今世界研究的一大热点。With the rapid development of mobile communication technology, wireless communication frequency bands are becoming more and more crowded, and frequency band resources are becoming increasingly tight. In order to accommodate more communication channels within a limited spectrum range, some broadband digital transmission technologies (OFDM, WCDMA, etc.) and modulation methods with high spectrum utilization (M-QAM, etc.). The signals transmitted by these technologies have the characteristics of non-constant envelope, wide frequency band, and peak-to-average ratio. When the modulated signal passes through a nonlinear power amplifier, in-band and out-of-band distortion will occur, and the output signal spectrum will expand and interfere with adjacent channels. Increase the bit error rate of the communication system. Therefore, modern communication systems put forward very high requirements on the linearity of RF power amplifiers. The linearization technology of power amplifiers has become one of the key technologies in the next generation of wireless communication systems, and has become a hot research topic in the world today.

功率放大器线性化技术很多，常用的有功率回退、前馈、负反馈和预失真等。功率回退法是传统而有效的一种方法，然而，工作点的回退降低了功放的电源利用效率并导致很高的热耗散；前馈能够做到宽带放大和高线性度，是目前比较成熟的一种线性化技术，但其缺点是效率低且复杂；负反馈是用牺牲放大器的增益来换取失真抑制的方法，其窄带宽和稳定性问题限制了其应用；预失真分为模拟预失真和数字预失真，数字预失真技术具有稳定、高效、宽带宽与自适应等优势，能达到中等程度的线性化，是比较有前途的一种线性化技术。There are many linearization techniques for power amplifiers, commonly used are power back-off, feedforward, negative feedback, and pre-distortion. The power back-off method is a traditional and effective method. However, the back-off of the operating point reduces the power utilization efficiency of the power amplifier and leads to high heat dissipation; feed-forward can achieve broadband amplification and high linearity, which is currently the A relatively mature linearization technology, but its disadvantage is low efficiency and complexity; negative feedback is a method of sacrificing the gain of the amplifier in exchange for distortion suppression, and its narrow bandwidth and stability problems limit its application; predistortion is divided into analog Predistortion and digital predistortion. Digital predistortion technology has the advantages of stability, high efficiency, wide bandwidth and self-adaptation, and can achieve a moderate degree of linearization. It is a promising linearization technology.

数字预失真技术是在功放之前加入一个与功放非线性特性相逆的预失真器，预失真器在数字域中对IQ基带信号进行预处理，经过预处理后的基带信号经D/A变换、滤波、上变频后送入功率放大器，功率放大器将信号放大后发射出去，由于在基带的预失真器中对信号进行了非线性预处理，且其非线性特性与功放的非线性特性相逆，则在功放输出端可得到线性放大的射频信号。通常，为了自适应的获得预失真器的查找表参数或者是多项式的系数，需要将功放输出信号的一部分经衰减后反馈回来，反馈信号经下变频、滤波、A/D变换后供自适应学习算法使用，通过自适应算法获取或调整预失真器的参数(查找表或者是多项式的系数)，使预失真器的非线性特性与功放的非线性特性相逆，从而达到线性化的目的。Digital predistortion technology is to add a predistorter inverse to the nonlinear characteristics of the power amplifier before the power amplifier. The predistorter preprocesses the IQ baseband signal in the digital domain, and the preprocessed baseband signal is converted by D/A. After filtering and up-converting, it is sent to the power amplifier, and the power amplifier amplifies the signal and transmits it out. Since the signal is pre-processed nonlinearly in the baseband predistorter, and its nonlinear characteristics are opposite to those of the power amplifier, Then a linearly amplified radio frequency signal can be obtained at the output end of the power amplifier. Usually, in order to adaptively obtain the look-up table parameters or polynomial coefficients of the predistorter, it is necessary to feed back a part of the output signal of the power amplifier after attenuation, and the feedback signal is used for adaptive learning after down-conversion, filtering, and A/D conversion The algorithm is used to obtain or adjust the parameters of the pre-distorter (look-up table or polynomial coefficients) through an adaptive algorithm, so that the nonlinear characteristics of the pre-distorter and the nonlinear characteristics of the power amplifier are reversed, so as to achieve the purpose of linearization.

在数字基带预失真中，常采用的自适应算法有LMS(最小均方)、RLS(递归最小二乘)等迭代算法，这些算法都需要经过学习和迭代，算法能得到最优预失真器参数的前提是要保证算法能够收敛，且基于迭代的自适应算法通常都需要经过大量的计算才能获得满意的结果。In digital baseband predistortion, commonly used adaptive algorithms include iterative algorithms such as LMS (least mean square) and RLS (recursive least squares). These algorithms need to be learned and iterated, and the algorithm can obtain the optimal predistorter parameters. The premise of the algorithm is to ensure that the algorithm can converge, and iterative-based adaptive algorithms usually require a lot of calculations to obtain satisfactory results.

针对迭代法存在的缺点，人们提出了基于非迭代的预失真方法，已有的非迭代预失真方法原理如下：Aiming at the shortcomings of the iterative method, people proposed a non-iterative pre-distortion method. The principle of the existing non-iterative pre-distortion method is as follows:

首先采用训练序列发生器产生幅度线性增长的一组训练序列，设输入信号的最大幅度为归一化电平1，训练序列的初始相位为

则第i个时刻的输入信号为

(其中i＝1，2，...，N)，无记忆非线性功放系统表现为AM/AM变换和AM/PM变换，该信号经过功放后反馈回来的信号为(将IQ信号写成复信号的形式)：First, the training sequence generator is used to generate a set of training sequences whose amplitude increases linearly. The maximum amplitude of the input signal is normalized level 1, and the initial phase of the training sequence is

Then the input signal at the i-th moment is

(where i=1, 2,..., N), the non-memory nonlinear power amplifier system is manifested as AM/AM conversion and AM/PM conversion, and the signal fed back after the signal passes through the power amplifier is (write the IQ signal as a complex signal form):

== ythe y II (( ii )) ++ jyjy QQ (( ii ))

其中，y_I(i)和y_Q(i)为i时刻反馈回来的I路信号和Q路信号，

和

分别表示幅度为

的信号经过功放后反馈回来的信号的幅度值和相位的改变量，令 r i = A ( i N ) , θ i = Φ ( i N ) 则可得到反馈回来的i时刻的IQ两路信号分别为：Among them, y _I (i) and y _Q (i) are the I channel signal and the Q channel signal fed back at time i,

and

Respectively, the magnitude of

After the signal passes through the power amplifier, the amplitude value and phase change of the feedback signal, so that r i = A ( i N ) , θ i = Φ ( i N ) Then the two IQ signals at time i that are fed back can be obtained as follows:

根据以上两式，将反馈回的IQ两路信号经过以下运算就可得到r_i和θ_i的值According to the above two formulas, the values of r _i and θ _i can be obtained by performing the following operations on the feedback IQ two-way signals

rr ii == ythe y II (( ii )) 22 ++ ythe y QQ (( ii )) 22 ,, (( ii == 1,21,2 ,, .. .. .. ,, NN )) -- -- -- (( 44 ))

为了简化运算，可以将训练序列的初始相位

设定为0。对于i时刻输入幅度为

且初始相位为0的信号，经过功放非线性放大后其幅度变成了r_i，相位在原信号的基础上改变了θ_i。基于间接学习结构的原理(附图1)，可用后失真方法得到的功放逆模型参数作为预失真器的参数，为了使经过后失真处理以后的信号与输入功放前的信号相等，则对于输入幅度为r_i的信号，经过后失真处理后其幅度应该变成i/N，且相位的改变量为-θ_i，如果求得了关于信号幅度r_i和相位改变量-θ_i的序列，则根据后失真原理可得到预失真器的查找表参数，相应的预失真器的查找表为：In order to simplify the operation, the initial phase of the training sequence can be

Set to 0. For time i, the input amplitude is

And the signal whose initial phase is 0, after being amplified nonlinearly by the power amplifier, its amplitude becomes r _i , and the phase changes θ _i on the basis of the original signal. Based on the principle of indirect learning structure (accompanying drawing 1), the parameters of the inverse model of the power amplifier obtained by the post-distortion method can be used as the parameters of the pre-distorter. In order to make the signal after post-distortion processing equal to the signal before the input power amplifier, the For the signal r _i , its amplitude should become i/N after post-distortion processing, and the phase change amount is -θ _i , if the sequence about the signal amplitude r _i and phase change amount -θ _i is obtained, then according to The post-distortion principle can obtain the look-up table parameters of the pre-distorter, and the corresponding look-up table of the pre-distorter is:

表1Table 1

按照以上方法，先用幅度线性增长的训练序列通过功放系统，并将反馈回的IQ两路信号采集回来，根据(4)(5)两式，经过计算就可以获得预失真器中查找表的参数r_i和-θ_i，从而可以构建表格1中的预失真器查找表。According to the above method, first use the training sequence with linearly increasing amplitude to pass through the power amplifier system, and collect the feedback IQ two-way signal. According to the two formulas (4) and (5), after calculation, you can get the value of the lookup table in the predistorter. Parameters r _i and -θ _i , so that the predistorter lookup table in Table 1 can be constructed.

该非迭代方法的主要缺点是预失真处理时需要将IQ信号转换成极坐标的形式，根据输入信号的幅度分别对幅度和相位进行修正，并需要将修正后的极坐标形式的信号再转换成IQ信号的形式，因此，预失真处理需要两次坐标转换来完成，并且在获得查找表参数θ_i的过程中需要反正切运算，运算较复杂。The main disadvantage of this non-iterative method is that the IQ signal needs to be converted into polar coordinates during pre-distortion processing, and the amplitude and phase are respectively corrected according to the amplitude of the input signal, and the corrected polar coordinates signal needs to be converted into The form of the IQ signal, therefore, the pre-distortion process needs two coordinate conversions to complete, and the arctangent operation is required in the process of obtaining the look-up table parameter θ _i , and the operation is more complicated.

发明内容Contents of the invention

本发明提供一种基于单路反馈和非迭代技术的基带预失真功放线性化方法，该方法基于单路反馈和非迭代技术，可大大简化运算过程、提高效率，并可简化实现电路和降低系统成本。The invention provides a baseband predistortion power amplifier linearization method based on single-channel feedback and non-iterative technology. The method is based on single-channel feedback and non-iterative technology, which can greatly simplify the operation process, improve efficiency, and simplify the implementation circuit and reduce system cost.

本发明的目的是通过以下技术方案实现的：The purpose of the present invention is achieved through the following technical solutions:

一种基于单路反馈和非迭代技术的基带预失真功放线性化方法，如图2所示，包括以下步骤：A baseband predistortion power amplifier linearization method based on single feedback and non-iterative technology, as shown in Figure 2, comprises the following steps:

步骤1：将一组幅度线性增长且相位为0的训练序列

其中i＝1，2，…，i，…N，作为输入信号经D/A转换、滤波和上变频后输入到功放；然后采集I路反馈信号，即将功放输出信号经耦合、下变频、滤波和A/D转换；设功放允许的最大输入信号的幅度为归一化电平1，则第i个时刻功放的输入信号为

第i个时刻I路反馈信号为I₁(i)；则对应于训练序列

的反馈序列为{I₁(i)}。训练序列

可用序列产生器产生。Step 1: A set of training sequences whose amplitude is linearly increased and whose phase is 0

Among them, i=1, 2, ..., i, ... N, as the input signal is input to the power amplifier after D/A conversion, filtering and up-conversion; and A/D conversion; if the amplitude of the maximum input signal allowed by the power amplifier is normalized level 1, then the input signal of the power amplifier at the i-th moment is

The i-th moment I channel feedback signal is I ₁ (i); then it corresponds to the training sequence

The feedback sequence of is {I ₁ (i)}. training sequence

Can be generated using a sequence generator.

步骤2：将另一组幅度线性增长且相位为90度的训练序列

其中i＝1，2，…，i，…N，作为输入信号经D/A转换、滤波和上变频后输入到功放；然后采集I路反馈信号，即将功放输出信号经耦合、下变频、滤波和A/D转换；设功放允许的最大输入信号的幅度为归一化电平1，则第i个时刻功放的输入信号为

第i个时刻I路反馈信号为I₂(i)；则对应于训练序列

的反馈序列为{I₂(i)}。训练序列

可用序列产生器产生。Step 2: Add another set of training sequences whose amplitude is linearly increased and whose phase is 90 degrees

Among them, i=1, 2, ..., i, ... N, as the input signal is input to the power amplifier after D/A conversion, filtering and up-conversion; and A/D conversion; if the amplitude of the maximum input signal allowed by the power amplifier is normalized level 1, then the input signal of the power amplifier at the i-th moment is

The feedback signal of channel I at the i-th moment is I ₂ (i); then it corresponds to the training sequence

The feedback sequence of is {I ₂ (i)}. training sequence

Can be generated using a sequence generator.

步骤3：构建初始预失真处理查找表。Step 3: Construct the initial predistortion processing lookup table.

预失真处理查找表包括索引项和查找表项组成，其中索引项为输入信号幅度r_i，查找表项包括查找表LUT1和查找表LUT2。The pre-distortion processing lookup table consists of an index item and a lookup table item, wherein the index item is the input signal amplitude r _i , and the lookup table item includes a lookup table LUT1 and a lookup table LUT2.

其中，输入信号幅度 r i = I 1 ( i ) 2 + I 2 ( i ) 2 , 相应输入信号幅度r_i下查找表LUT1的查找值为iI₁(i)/r_i ²N，相应输入信号幅度r_i下查找表LUT2的查找值为iI₂(i)/r_i ²N。where the input signal amplitude r i = I 1 ( i ) 2 + I 2 ( i ) 2 , The lookup value of the lookup table LUT1 under the corresponding input signal amplitude r _i is iI ₁ (i)/r _i ² N, and the lookup value of the lookup table LUT2 under the corresponding input signal amplitude r _i is iI ₂ (i)/r _i ² N.

具体初始预失真处理查找表如表2所示：The specific initial predistortion processing lookup table is shown in Table 2:

表2Table 2

步骤4：对步骤3所得的初始预失真处理查找表进行插值处理，得到最终的预失真处理查找表。Step 4: Perform interpolation processing on the initial predistortion processing lookup table obtained in step 3 to obtain a final predistortion processing lookup table.

步骤3所得的初始预失真处理查找表的索引项是输入信号幅度序列r_i，该序列不是等间隔的，为了得到等间隔的其它查找表项，这里可采用两种方法来获得其它幅度值对应的查找表项参数：一是用线性插值法获取其它输入信号幅度所对应的查找表LUT1和查找表LUT2的值；另一种是利用曲线拟合的方法用多项式来拟合幅度和相位特性曲线，然后再根据拟合的多项式通过计算后填充相应查找表LUT1和LUT2的值。The index item of the initial predistortion processing lookup table obtained in step 3 is the input signal amplitude sequence r _i , which is not equally spaced. In order to obtain other equally spaced lookup table items, two methods can be used to obtain other amplitude values corresponding to The parameters of the lookup table items: one is to use linear interpolation to obtain the values of the lookup table LUT1 and the lookup table LUT2 corresponding to the amplitude of other input signals; the other is to use a curve fitting method to use polynomials to fit the amplitude and phase characteristic curves , and then fill in the values of the corresponding lookup tables LUT1 and LUT2 after calculation according to the fitted polynomial.

步骤5：对于整个功放系统的输入信号x(n)＝re^jθ＝x_I(n)+jx_Q(n)，其中n代表信号输入时刻，r为输入信号幅度，θ为输入信号相位；采用如图3所示的预失真器进行处理，具体过程为：首先根据输入信号x(n)的幅度，在步骤4所得的最终的预失真处理查找表中找到相应输入信号幅度下的查找表LUT1和查找表LUT2的值LTU1和LTU2；然后分别计算x_I(n)·LTU1、x_I(n)·LTU2、x_Q(n)·LTU1和x_Q(n)·LTU2；再计算(x_I(n)·LTU1-x_Q(n)·LTU2)和(x_Q(n)·LTU1+x_I(n)·LTU2)，将(x_I(n)·LTU1-x_Q(n)·LTU2)作为预失真器的I路输出信号pr_I(n)，将(x_Q(n)·LTU1+x_I(n)·LTU2)作为预失真器的Q路输出信号pr_Q(n)。Step 5: For the input signal x(n)=re ^jθ =x _I (n)+jx _Q (n) of the entire power amplifier system, where n represents the signal input moment, r is the input signal amplitude, and θ is the input signal phase; The predistorter as shown in Figure 3 performs processing, and the specific process is: first, according to the amplitude of the input signal x(n), find the lookup table LUT1 under the corresponding input signal amplitude in the final predistortion processing lookup table obtained in step 4 and the values LTU1 and LTU2 of the lookup table LUT2; then calculate x _I (n) LTU1, x _I (n) LTU2, x _Q (n) LTU1 and x _Q (n) LTU2 respectively; then calculate (x _I (n) LTU1-x _Q (n) LTU2) and (x _Q (n) LTU1+x _I (n) LTU2), the (x _I (n) LTU1-x _Q (n) LTU2 ) is used as the I-channel output signal pr _I (n) of the predistorter, and (x _Q (n)·LTU1+x _I (n)·LTU2) is used as the Q-channel output signal pr _Q (n) of the predistorter.

步骤6：将步骤5所得的预失真器的I路输出信号pr_I(n)和Q路输出信号pr_Q(n)经D/A转换、滤波和上变频后输入到功放，从功放输出端即可得到线性化的输出信号。Step 6: Input the I-channel output signal pr _I (n) and the Q-channel output signal pr _Q (n) of the predistorter obtained in step 5 to the power amplifier after D/A conversion, filtering and up-conversion, and from the output terminal of the power amplifier A linearized output signal can be obtained.

本发明采用单路反馈和非迭代技术对功放系统基带输入信号进行数字预失真处理，将两路正交的幅度线性增长的训练序列先后输入到功放，然后通过功放输出的I路反馈信号序列构建改进型的查找表；整个功放系统的输入信号先通过幅度计算、查找对应的查找表值，再分别对I、Q两路基带输入信号作相应的预示真处理，预失真处理后的输入信号经D/A转换、滤波和上变频后输到功放，即可得到与整个功放系统基带输入信号呈线性放大的输出信号。整个预失真处理过程可在数字处理芯片(包括DSP或FPGA)中实现，与现有技术相比，本发明无需反正切或者是反余弦运算，无需进行坐标变换，从而降低了计算量，提高了计算效率，其实现系统结构更为简单。The present invention adopts single-channel feedback and non-iterative technology to carry out digital pre-distortion processing on the baseband input signal of the power amplifier system, and successively input two-way orthogonal training sequences with amplitude linear growth into the power amplifier, and then construct through the I-way feedback signal sequence output by the power amplifier Improved look-up table; the input signal of the entire power amplifier system is firstly calculated by the amplitude, and the corresponding look-up table value is searched, and then the I and Q two-way baseband input signals are respectively subjected to corresponding predictive processing, and the input signal after pre-distortion processing is passed After D/A conversion, filtering and up-conversion, it is output to the power amplifier, and the output signal that is linearly amplified with the baseband input signal of the entire power amplifier system can be obtained. The whole pre-distortion process can be realized in the digital processing chip (including DSP or FPGA). Compared with the prior art, the present invention does not need arctangent or arccosine operation, and coordinate transformation is not required, thereby reducing the amount of calculation and improving the Computational efficiency, its implementation system structure is simpler.

附图说明Description of drawings

图1间接学习结构原理框图。Figure 1. Block diagram of indirect learning structure.

图2是本发明所采用的预失真线性化方法原理框图。Fig. 2 is a functional block diagram of the predistortion linearization method adopted in the present invention.

图3是本发明所采用的查找表预失真器的结构原理框图Fig. 3 is the structural principle block diagram of the look-up table predistorter that the present invention adopts

具体实施方式Detailed ways

下面结合附图来详细说明本发明所述的基于单路反馈和非迭代技术的基带预失真功放线性化方法。The linearization method of the baseband predistortion power amplifier based on single-channel feedback and non-iterative technology according to the present invention will be described in detail below in conjunction with the accompanying drawings.

一种基于单路反馈和非迭代技术的基带预失真功放线性化方法，如图2所示，包括以下步骤：A baseband predistortion power amplifier linearization method based on single feedback and non-iterative technology, as shown in Figure 2, comprises the following steps:

步骤1：将一组幅度线性增长且相位为0的训练序列其中i＝1，2，…，i，…N，作为输入信号经D/A转换、滤波和上变频后输入到功放；然后采集I路反馈信号，即将功放输出信号经耦合、下变频、滤波和A/D转换；设功放允许的最大输入信号的幅度为归一化电平1，则第i个时刻功放的输入信号为第i个时刻I路反馈信号为 I 1 ( i ) = A ( i N ) cos ( Φ ( i N ) ) , 其中幅度函数，为功放附加相位；则对应于训练序列

的反馈序列为{I₁(i)}。训练序列

可用序列产生器产生。Step 1: A set of training sequences whose amplitude is linearly increased and whose phase is 0 Among them, i=1, 2, ..., i, ... N, as the input signal is input to the power amplifier after D/A conversion, filtering and up-conversion; and A/D conversion; if the amplitude of the maximum input signal allowed by the power amplifier is normalized level 1, then the input signal of the power amplifier at the i-th moment is The feedback signal of channel I at the i-th moment is I 1 ( i ) = A ( i N ) cos ( Φ ( i N ) ) , in magnitude function, Additional phase for the PA; then corresponds to the training sequence

The feedback sequence of is {I ₁ (i)}. training sequence

Can be generated using a sequence generator.

步骤2：将另一组幅度线性增长且相位为90度的训练序列

其中i＝1，2，…，i，…N，作为输入信号经D/A转换、滤波和上变频后输入到功放；然后采集I路反馈信号，即将功放输出信号经耦合、下变频、滤波和A/D转换；设功放允许的最大输入信号的幅度为归一化电平1，则第i个时刻功放的输入信号为

第i个时刻I路反馈信号为 I 2 ( i ) = A ( i N ) cos ( π 2 + Φ ( i N ) ) = - A ( i N ) sin ( Φ ( i N ) ) ; 则对应于训练序列

的反馈序列为{I₂(i)}。训练序列

可用序列产生器产生。Step 2: Add another set of training sequences whose amplitude is linearly increased and whose phase is 90 degrees

Among them, i=1, 2, ..., i, ... N, as the input signal is input to the power amplifier after D/A conversion, filtering and up-conversion; and A/D conversion; if the amplitude of the maximum input signal allowed by the power amplifier is normalized level 1, then the input signal of the power amplifier at the i-th moment is

The feedback signal of channel I at the i-th moment is I 2 ( i ) = A ( i N ) cos ( π 2 + Φ ( i N ) ) = - A ( i N ) sin ( Φ ( i N ) ) ; then corresponds to the training sequence

The feedback sequence of is {I ₂ (i)}. training sequence

Can be generated using a sequence generator.

步骤3：构建初始预失真处理查找表。Step 3: Construct the initial predistortion processing lookup table.

预失真处理查找表包括索引项和查找表项组成，其中索引项为输入信号幅度r_i，查找表项包括查找表LUT1和查找表LUT2。The pre-distortion processing lookup table consists of an index item and a lookup table item, wherein the index item is the input signal amplitude r _i , and the lookup table item includes a lookup table LUT1 and a lookup table LUT2.

其中，输入信号幅度 r i = I 1 ( i ) 2 + I 2 ( i ) 2 , 相应输入信号幅度r_i下查找表LUT1的查找值为iI₁(i)/r_i ²N，相应输入信号幅度r_i下查找表LUT2的查找值为iI₂(i)/r_i ²N。where the input signal amplitude r i = I 1 ( i ) 2 + I 2 ( i ) 2 , The lookup value of the lookup table LUT1 under the corresponding input signal amplitude r _i is iI ₁ (i)/r _i ² N, and the lookup value of the lookup table LUT2 under the corresponding input signal amplitude r _i is iI ₂ (i)/r _i ² N.

具体初始预失真处理查找表如下表所示：The specific initial pre-distortion processing lookup table is shown in the following table:

步骤4：对步骤3所得的初始预失真处理查找表进行插值处理，得到最终的预失真处理查找表。Step 4: Perform interpolation processing on the initial predistortion processing lookup table obtained in step 3 to obtain a final predistortion processing lookup table.

步骤3所得的初始预失真处理查找表的索引项是输入信号幅度序列r_i，该序列不是等间隔的，为了得到等间隔的其它查找表项，这里可采用两种方法来获得其它幅度值对应的查找表项参数：一是用线性插值法获取其它输入信号幅度所对应的查找表LUT1和查找表LUT2的值；另一种是利用曲线拟合的方法用多项式来拟合幅度和相位特性曲线，然后再根据拟合的多项式通过计算后填充相应查找表LUT1和LUT2的值。The index item of the initial predistortion processing lookup table obtained in step 3 is the input signal amplitude sequence r _i , which is not equally spaced. In order to obtain other equally spaced lookup table items, two methods can be used to obtain other amplitude values corresponding to The parameters of the lookup table items: one is to use linear interpolation to obtain the values of the lookup table LUT1 and the lookup table LUT2 corresponding to the amplitude of other input signals; the other is to use a curve fitting method to use polynomials to fit the amplitude and phase characteristic curves , and then fill in the values of the corresponding lookup tables LUT1 and LUT2 after calculation according to the fitted polynomial.

步骤5：对于整个功放系统的输入信号x(n)＝re^jθ＝x_I(n)+jx_Q(n)，其中n代表信号输入时刻，r为输入信号幅度，θ为输入信号相位；采用如图3所示的预失真器进行处理，具体过程为：首先根据输入信号x(n)的幅度，在步骤4所得的最终的预失真处理查找表中找到相应输入信号幅度下的查找表LUT1和查找表LUT2的值LTU1和LTU2；然后分别计算x_I(n)·LTU1、x_I(n)·LTU2、x_Q(n)·LTU1和x_Q(n)·LTU2；再计算(x_I(n)·LTU1-x_Q(n)·LTU2)和(x_Q(n)·LTU1+x_I(n)·LTU2)，将(x_I(n)·LTU1-x_Q(n)·LTU2)作为预失真器的I路输出信号pr_I(n)，将(x_Q(n)·LTU1+x_I(n)·LTU2)作为预失真器的Q路输出信号pr_Q(n)。Step 5: For the input signal x(n)=re ^jθ =x _I (n)+jx _Q (n) of the entire power amplifier system, where n represents the signal input moment, r is the input signal amplitude, and θ is the input signal phase; The predistorter as shown in Figure 3 performs processing, and the specific process is: first, according to the amplitude of the input signal x(n), find the lookup table LUT1 under the corresponding input signal amplitude in the final predistortion processing lookup table obtained in step 4 and the values LTU1 and LTU2 of the lookup table LUT2; then calculate x _I (n) LTU1, x _I (n) LTU2, x _Q (n) LTU1 and x _Q (n) LTU2 respectively; then calculate (x _I (n) LTU1-x _Q (n) LTU2) and (x _Q (n) LTU1+x _I (n) LTU2), the (x _I (n) LTU1-x _Q (n) LTU2 ) is used as the I-channel output signal pr _I (n) of the predistorter, and (x _Q (n)·LTU1+x _I (n)·LTU2) is used as the Q-channel output signal pr _Q (n) of the predistorter.

步骤6：将步骤5所得的预失真器的I路输出信号pr_I(n)和Q路输出信号pr_Q(n)经D/A转换、滤波和上变频后输入到功放，从功放输出端即可得到线性化的输出信号。Step 6: Input the I-channel output signal pr _I (n) and the Q-channel output signal pr _Q (n) of the predistorter obtained in step 5 to the power amplifier after D/A conversion, filtering and up-conversion, and from the output terminal of the power amplifier A linearized output signal can be obtained.

以下简述本发明预失真处理原理：The following briefly describes the predistortion processing principle of the present invention:

设整个功放系统输入信号为x(n)＝re^jθ＝x_I(n)+jx_Q(n)，则经过预失真器处理后变成：Assuming that the input signal of the entire power amplifier system is x(n)=re ^jθ ＝x _I (n)+jx _Q (n), after being processed by the predistorter, it becomes:

prpr (( nno )) == AA (( rr )) ee jj (( θθ ++ ΦΦ (( rr )) ))

== xx (( nno )) AA (( rr )) rr ee jΦjΦ (( rr ))

== [[ xx II (( nno )) ++ jxjx QQ (( nno )) ]] [[ AA (( rr )) rr coscos (( ΦΦ (( rr )) )) ++ jj AA (( rr )) rr sinsin (( ΦΦ (( rr )) )) ]] -- -- -- (( 66 ))

== [[ xx II (( nno )) AA (( rr )) rr coscos (( ΦΦ (( rr )) )) -- xx QQ (( nno )) AA (( rr )) rr sinsin (( ΦΦ (( rr )) )) ]]

++ jj [[ xx QQ (( nno )) AA (( rr )) rr coscos (( ΦΦ (( rr )) )) ++ xx II (( nno )) AA (( rr )) rr sinsin (( ΦΦ (( rr )) )) ]]

根据上式用两个查找表LUT1和LUT2来分别存储

和

的值，然后根据输入信号的幅度，在查找表LUT1和LUT2中索引到对应的值，然后进行乘法和加法运算就可直接得到IQ两路输出信号。According to the above formula, two lookup tables LUT1 and LUT2 are used to store them separately

and

value, and then according to the amplitude of the input signal, index the corresponding value in the lookup table LUT1 and LUT2, and then perform multiplication and addition operations to directly obtain the IQ two-way output signal.

由步骤1、2，有：From steps 1 and 2, there are:

rr ii == ii NN -- -- -- (( 77 ))

II 11 (( ii )) == AA (( ii NN )) coscos (( ΦΦ (( ii NN )) )) -- -- -- (( 88 ))

II 22 (( ii )) == AA (( ii NN )) coscos (( ππ 22 ++ ΦΦ (( ii NN )) )) == -- AA (( ii NN )) sinsin (( ΦΦ (( ii NN )) )) -- -- -- (( 99 ))

结合(6)、(7)、(8)和(9)式，可得：Combining equations (6), (7), (8) and (9), we can get:

对应于幅度为r_i的输入信号，查找表LUT1和LUT2中的值为Corresponding to an input signal with amplitude r _i , the values in the lookup tables LUT1 and LUT2 are

LUTLUTs 11 (( rr ii )) == AA (( rr ii )) rr ii coscos (( ΦΦ (( rr ii )) )) == iIi 11 (( ii )) rr ii 22 NN ,, (( ii == 1,21,2 ,, .. .. .. ,, NN ))

LUTLUTs 22 (( rr ii )) == AA (( rr ii )) rr ii sinsin (( ΦΦ (( rr ii )) )) == iIi 22 (( ii )) rr ii 22 NN ,, (( ii == 1,21,2 ,, .. .. .. ,, NN ))