CN104580060A

CN104580060A - Digital pre-distortion correcting device and method against IQ unbalance distortion

Info

Publication number: CN104580060A
Application number: CN201510027992.1A
Authority: CN
Inventors: 唐宏; 李玉妍; 李兆玉; 黄锐; 田燕; 夏小夏; 粟根花; 余瑶
Original assignee: Chongqing University of Post and Telecommunications
Current assignee: CERTUSNET CORP
Priority date: 2015-01-20
Filing date: 2015-01-20
Publication date: 2015-04-29
Anticipated expiration: 2035-01-20
Also published as: CN104580060B

Abstract

The invention proposes an IQ imbalance correction technology based on a digital predistortion system, which belongs to the field of wireless communication. The quadrature modulated signal enters the digital predistortion system in two ways. First, the channel estimation is performed to obtain the corresponding estimation parameters and the data is initialized; then the in-phase signal and quadrature signal are calculated according to the input in-phase signal and quadrature signal. DC offset factor and DC compensation; calculate the gain imbalance factor and phase imbalance factor of the in-phase signal and quadrature signal according to the input in-phase signal and quadrature signal; perform mirror correction by substituting corresponding parameters according to the mirror correction model; Finally, the compensated gain and phase factors are updated to complete the IQ imbalance correction process. The method of the invention can reduce the complexity and cost of system hardware, and at the same time improve the accuracy of error estimation and compensation.

Description

A digital predistortion correction device and method for IQ unbalanced distortion

技术领域technical field

本发明涉及无线通信领域，特别涉及多载波无线通信系统中的同相(I)和正交相(Q)不平衡校正(IQ不平衡校正)装置和方法。The invention relates to the field of wireless communication, in particular to an in-phase (I) and quadrature-phase (Q) imbalance correction (IQ imbalance correction) device and method in a multi-carrier wireless communication system.

背景技术Background technique

无线通信系统及相关的标准正在不断使用更加复杂的调制技术(例如64QAM和OFDM(正交频分复用))来增加通信信道的吞吐量。这些更复杂的调制技术增加了对于低成本的直接转换接收器和发射器的同相(I)和正交相(Q)路径之间的不平衡性。基带同相和正交相IQ信号经过正交调制后输出两路信号的增益通常存在着一定差异，该差异就造成了增益不平衡失真，I和Q信道之间的增益不平衡是由混频器、滤波器和数模转换器(ADC)以及信道之间的不一致性导致的。正交调制器中有一个载波信号分离器，用于将本振信号分离成两路相位相差90度的载波信号，而实际硬件电路中要达到完全精确的90度几乎是不可能的，总会存在一个相位误差，当同相和正交相信道的本地振荡器信号之间的相位差不是精确的90度时会发生相位不平衡，该相位不平衡产生的误差就是相位不平衡失真。Wireless communication systems and related standards are continuously using more complex modulation techniques such as 64QAM and OFDM (Orthogonal Frequency Division Multiplexing) to increase the throughput of communication channels. These more complex modulation techniques increase the imbalance between the in-phase (I) and quadrature-phase (Q) paths for low-cost direct conversion receivers and transmitters. The baseband in-phase and quadrature-phase IQ signals usually have a certain difference in the gain of the two output signals after quadrature modulation, which causes gain imbalance and distortion. The gain imbalance between the I and Q channels is determined by the mixer , filters and digital-to-analog converters (ADCs), and inconsistencies between channels. There is a carrier signal separator in the quadrature modulator, which is used to separate the local oscillator signal into two carrier signals with a phase difference of 90 degrees. However, it is almost impossible to achieve a completely accurate 90 degrees in the actual hardware circuit. There is a phase error, and phase imbalance occurs when the phase difference between the local oscillator signals of the in-phase and quadrature-phase channels is not exactly 90 degrees, and the error produced by this phase imbalance is phase imbalance distortion.

考虑正交调制本振的非线性失真，当输入信号的同相分量I(t)和正交分量Q(t)分别表示为：Considering the nonlinear distortion of the quadrature modulated local oscillator, when the in-phase component I(t) and quadrature component Q(t) of the input signal are expressed as:

I(t)＝Gcos(ωt+φ)+D (1)I(t)＝Gcos(ωt+φ)+D (1)

Q(t)＝sinωt (2)Q(t)＝sinωt (2)

其中，G表示增益，ω为信号角频率，t为时间，φ为信号相位，D为直流分量。当调制器的幅度误差ε和相位误差θ，则经过正交调制后输出的信号为：Among them, G represents the gain, ω is the angular frequency of the signal, t is the time, φ is the signal phase, and D is the DC component. When the amplitude error ε and phase error θ of the modulator, the output signal after quadrature modulation is:

$\begin{matrix} {P P}_{out out} = = ((G G cos cos ((ωt ωt + + φ φ)) + + D D.)) ϵ ϵ cos cos {ω ω}_{c c} t t - - sin sin {ω ω}_{c c} t t sin sin (({ω ω}_{c c} t t + + θ θ)) \\ = = Gϵ Gϵ cos cos ((ωt ωt + + φ φ)) cos cos {ω ω}_{c c} t t + + Dϵ Dϵ cos cos {ω ω}_{c c} t t - - cos cos ((ωt ωt - - \frac{π π}{22})) sin sin (({ω ω}_{c c} t t + + θ θ)) \\ = = ((\frac{11}{22} Gϵ Gϵ cos cos φ φ - - \frac{11}{22} cos cos θ θ)) cos cos ((ωt ωt - - {ω ω}_{c c} t t)) + + ((\frac{11}{22} Gϵ Gϵ sin sin φ φ + + \frac{11}{22} sin sin θ θ)) sin sin (({ω ω}_{c c} t t - - ωt ωt)) + + \\ ((\frac{11}{22} Gϵ Gϵ cos cos φ φ + + \frac{11}{22} cos cos θ θ)) cos cos ((ωt ωt + + {ω ω}_{c c} t t)) - - \frac{11}{22} ((Gϵ Gϵ sin sin φ φ + + \frac{11}{22} sin sin θ θ)) sin sin ((ωt ωt + + {ω ω}_{c c} t t)) + + Dϵ Dϵ cos cos {ω ω}_{c c} t t \end{matrix} - - - - - - ((33))$

从这个式子可以看出经过正交调制后的信号包含了3个部分，一个是在有用频率ω_ct+ωt上的信号，一种是由于本振引起的镜像干扰频率ω_ct-ωt上的信号，第三个是本振频率上ω_ct的信号。对我们来说有用信号只有ω_ct+ωt上的信号，而其他信号都是干扰信号。从表达式中也可看出直流偏移D过大对输入信号来说也就增加了本振的振幅，这样可能引起本振泄露，所以我们通常会说直流偏移过大会引起本振泄露。当原始信号经过调制后会引起IQ增益，相位不平衡和直流偏移等非线性失真，我们就需要在数字域上对上述的不平衡进行补偿。From this formula, it can be seen that the signal after quadrature modulation contains three parts, one is the signal on the useful frequency ω _c t+ωt, and the other is the image interference frequency ω _c t-ωt caused by the local oscillator The signal at ω _c t and the third is the signal at ω c t at the local oscillator frequency. For us, the useful signal is only the signal on ω _c t+ωt, while other signals are interference signals. It can also be seen from the expression that if the DC offset D is too large, it will increase the amplitude of the local oscillator for the input signal, which may cause leakage of the local oscillator. Therefore, we usually say that the leakage of the local oscillator is caused by excessive DC offset. When the original signal is modulated, it will cause nonlinear distortion such as IQ gain, phase imbalance and DC offset, and we need to compensate the above imbalance in the digital domain.

IQ不平衡失真主要包括上述增益不平衡、相位不平衡和直流偏移三个部分。其中，镜像信号就是由增益和相位不平衡引起的。镜像会导致发射杂散指标不能达标，如果镜像和有用信号互相混叠，还会恶化有用信号的矢量幅度误差EVM，另一方面镜像频率和有用信号落在彼此带内，使得射频链路滤波器无法抑制镜像。另外正交调制后的信号还存在着载波泄露，本振信号会泄露到射频输出端，从而造成直流偏移误差。这些IQ不平衡失真会使原信号产生失真分量，且使通信系统的误码率增大。IQ unbalanced distortion mainly includes the above-mentioned three parts of gain unbalance, phase unbalance and DC offset. Among them, the image signal is caused by gain and phase imbalance. The image will cause the emission spurious index to fail to meet the standard. If the image and the useful signal alias each other, it will also deteriorate the vector magnitude error EVM of the useful signal. On the other hand, the image frequency and the useful signal fall within each other's band, making the RF link filter Unable to suppress mirroring. In addition, the signal after quadrature modulation still has carrier leakage, and the local oscillator signal will leak to the radio frequency output end, thereby causing a DC offset error. These IQ unbalanced distortions will cause distortion components in the original signal, and increase the bit error rate of the communication system.

最初采用补偿IQ不平衡的方法是手工方式进行调节，但是这样的方法费时且不能进行自动跟踪补偿。之后提出了复合模型方法，将信道因子和IQ不平衡因子结合在一起，并利用特殊的双导频符号进行校正，这种方法需要大量的频率资源且不适合于时变信道。接着，人们又提出时域盲补偿方法，但这种方法假设IQ不平衡的影响集中在其中一路，不适用于实际的系统且实时性很差。随后，人们提出了自适应的IQ不平衡补偿方法，这种方法需要大量的训练符号和迭代运算以获得均衡器系数，且需要附加反馈电路完成实时检测IQ不平衡的变化情况，实现IQ不平衡的自适应补偿。然而，附加反馈电路大大提高了硬件开销和系统复杂度。The initial method of compensating for IQ imbalance is manual adjustment, but this method is time-consuming and cannot be automatically tracked and compensated. Then a composite model method is proposed, which combines the channel factor and the IQ imbalance factor together, and uses special double pilot symbols for correction. This method requires a large number of frequency resources and is not suitable for time-varying channels. Then, people proposed a time-domain blind compensation method, but this method assumes that the influence of IQ imbalance is concentrated in one of the channels, which is not applicable to the actual system and has poor real-time performance. Subsequently, people proposed an adaptive IQ imbalance compensation method, which requires a large number of training symbols and iterative operations to obtain equalizer coefficients, and requires an additional feedback circuit to complete real-time detection of IQ imbalance changes to achieve IQ imbalance adaptive compensation. However, the additional feedback circuit greatly increases hardware overhead and system complexity.

发明内容Contents of the invention

本发明针对现有的IQ不平衡补偿方法存在的上述缺陷，提出一种结构简单、容易实现、实时性好的校正方法。该校正方法基于数字预失真校正系统，不需要附加额外的反馈链路，在降低系统硬件复杂度和成本的同时还提高了误差估计和补偿的精度。Aiming at the above-mentioned defects in the existing IQ imbalance compensation method, the present invention proposes a correction method with simple structure, easy realization and good real-time performance. The correction method is based on a digital predistortion correction system, does not require an additional feedback link, and improves the accuracy of error estimation and compensation while reducing system hardware complexity and cost.

为了达到上述目的，本发明的技术方案是这样实现的。正交调制后的信号分两路进入数字预失真系统，首先进行信道估计获得相应的估计参数并对数据进行初始化；然后根据输入的同相信号和正交信号计算同相信号和正交信号的直流偏移因子并进行直流补偿，计算同相信号和正交信号的增益不平衡因子和相位不平衡因子，根据镜像校正模型对相应的参数进行镜像校正；最后更新补偿后的增益和相位因子完成IQ不平衡校正。具体技术方案如下：In order to achieve the above object, the technical solution of the present invention is achieved in this way. The quadrature-modulated signal enters the digital predistortion system in two ways. Firstly, the channel estimation is performed to obtain the corresponding estimation parameters and the data is initialized; then the in-phase signal and quadrature signal are calculated according to the input in-phase signal and quadrature signal. DC offset factor and DC compensation, calculate the gain imbalance factor and phase imbalance factor of the in-phase signal and quadrature signal, and perform mirror correction on the corresponding parameters according to the mirror correction model; finally update the compensated gain and phase factors to complete IQ imbalance correction. The specific technical scheme is as follows:

一种数字预失真系统中的IQ不平衡校正装置，该装置包括：数据接收模块、模数转换器、数字预失真模块、数模转换器和射频处理模块；数据接收处理模块接收微弱信号送入低噪声放大器放大，由IQ正交解调器下变频到中频模拟信号；模数转换器把中频模拟信号转换为中频数字信号；数字预失真模块校正功率放大器的非线性及校正调制器的本振泄露和镜像；数模转换器把经过数字预失真模块处理的中频数字信号转换成模拟信号；射频处理模块将转换的模拟信号进行正交调制上变频处理。An IQ unbalance correction device in a digital predistortion system, the device includes: a data receiving module, an analog-to-digital converter, a digital pre-distortion module, a digital-to-analog converter, and a radio frequency processing module; the data receiving and processing module receives weak signals and sends them to The low noise amplifier is amplified and down-converted by the IQ quadrature demodulator to an intermediate frequency analog signal; the analog-to-digital converter converts the intermediate frequency analog signal into an intermediate frequency digital signal; the digital predistortion module corrects the nonlinearity of the power amplifier and corrects the local oscillator of the modulator Leakage and mirroring; the digital-to-analog converter converts the intermediate frequency digital signal processed by the digital pre-distortion module into an analog signal; the radio frequency processing module performs quadrature modulation up-conversion processing on the converted analog signal.

所述数字预失真模块包括：预失真器、IQ补偿器、延时模块，预失真器接收正交调制后的同相信号和正交信号，IQ补偿器根据同相信号和正交信号计算各自的增益因子、幅度不平衡因子和相位不平衡因子，预失真器和IQ补偿器去使能，更新增益和相位不平衡因子，获得预设IQ补偿器增益和相位不平衡因子，并进行直流偏移校正；建立镜像校正模型对上述增益和相位不平衡因子进行镜像校正抑制镜像信号。当需消除系统中A频段内B频段的镜像信号时，镜像生成模型公式为：αA_FW+βB_FW＝A_FB，其中，A_FW为A频段的前向输出信号，B_FW为B频段的前向输出信号，A_FB为A频段的前向反馈信号，α表示A频段发射信号对自身影响的增益和相位系数，β表示A频段对B频段影响的增益和相位系数。所述直流偏移校正具体为，根据直流信号C₁和C₂调用直流偏移方程 $\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{21} + C_{c 2} h_{22} = - C_{2} \end{matrix},$ 计算直流偏移因子C_c1和C_c2，建立直流偏移矩阵 $[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{21} & h_{22} \end{matrix}]}^{- 1} [\begin{matrix} - C_{1} \\ - C_{2} \end{matrix}],$ 用直流偏移矩阵在IQ补偿器中补偿，其中，h₁₁，h₁₂，h₂₁，h₂₂分别为信号从预失真器输出进入IQ补偿器的信号的频率f₁₁，f₁₂，f₂₁，f₂₂的响应。The digital pre-distortion module includes: a pre-distorter, an IQ compensator, and a delay module. The pre-distorter receives the in-phase signal and the quadrature signal after quadrature modulation, and the IQ compensator calculates the respective Gain factor, amplitude imbalance factor and phase imbalance factor, predistorter and IQ compensator disable, update gain and phase imbalance factor, obtain preset IQ compensator gain and phase imbalance factor, and perform DC bias shift correction; establish an image correction model to perform image correction on the above gain and phase imbalance factors to suppress the image signal. When it is necessary to eliminate the image signal of the B frequency band in the A frequency band in the system, the image generation model formula is: αA _FW + βB _FW = A _FB , where A _FW is the forward output signal of the A frequency band, and B _FW is the forward output signal of the B frequency band. A _FB is the forward feedback signal of the A frequency band, α represents the gain and phase coefficient of the influence of the transmitted signal of the A frequency band on itself, and β represents the gain and phase coefficient of the influence of the A frequency band on the B frequency band. The DC offset correction specifically includes calling the DC offset equation according to the DC signals _C1 and _C2 $\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{twenty one} + C_{c 2} h_{twenty two} = - C_{2} \end{matrix},$ Calculate the DC offset factors C _c1 and C _c2 to establish the DC offset matrix $[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{twenty one} & h_{twenty two} \end{matrix}]}^{- 1} [\begin{matrix} - C_{1} \\ - C_{2} \end{matrix}],$ Compensate in the IQ compensator with a DC offset matrix, where h ₁₁ , h ₁₂ , h ₂₁ , h ₂₂ are the frequencies f ₁₁ , f ₁₂ , f ₂₁ of the signal entering the IQ compensator from the output of the predistorter, respectively, f ₂₂ 's response.

本发明还提出一种IQ不平衡校正的方法，所述方法包括如下步骤，根据输入的同相信号和正交信号计算同相信号和正交信号的直流偏移因子、增益不平衡因子和相位不平衡因子，并进行直流补偿；预失真器和IQ补偿器去使能，使得u(n)＝z(n)＝x(n)，其中，u(n)为进入预失真器之前的信号，z(n)为经过预失真器的输出信号，x(n)为IQ补偿器后的信号；去使能功率放大器PA得到反馈信号经过正交解调后的信号y(n)；通过x(n)和y(n)的差值信道估计及更新的增益和相位不平衡因子，获得预设IQ补偿器增益和相位不平衡因子；激活IQ补偿器，使能PA、启动PD训练序列，对输出的经过IQ校正的同相信号及正交信号进行逆平衡处理，得到经过处理的不平衡的同相信号和正交信号，并将其输出。当需消除系统中A频段内B频段的镜像信号时，镜像生成模型公式为：αA_FW+βB_FW＝A_FB，其中，A_FW为A频段的前向输出信号，B_FW为B频段的前向输出信号，A_FB为A频段的前向反馈信号，α表示A频段发射信号对自身影响的增益和相位系数，β表示A频段对B频段影响的增益和相位系数。所述直流偏移校正具体为，根据直流信号C₁和C₂调用直流偏移方程 $\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{21} + C_{c 2} h_{22} = - C_{2} \end{matrix},$ 计算直流偏移因子C_c1和C_c2，建立直流偏移矩阵 $[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{21} & h_{22} \end{matrix}]}^{- 1} [\begin{matrix} - C_{1} \\ - C_{2} \end{matrix}],$ 用直流偏移矩阵在IQ补偿器中补偿，其中，h₁₁，h₁₂，h₂₁，h₂₂分别为信号从预失真器输出进入IQ补偿器的信号的频率f₁₁，f₁₂，f₂₁，f₂₂的响应。A频段前向数据向A频段反馈数据的传递参数矩阵为： $A_{FW} = [\begin{matrix} A_{FW} I (0) & A_{FW} I (1) & A_{FW} I (2) & . . . & A_{FW} I (N - 2) & A_{FW} I (N - 1) \\ A_{FW} Q (0) & A_{FW} Q (1) & A_{FW} Q (2) & . . . & A_{FW} Q (N - 2) & A_{FW} Q (N - 1) \end{matrix}],$ 其中，第一行A_FWI(0),A_FWI(1)…A_FWI(N-1)为A频段前向数据的同相分量，作为实部；第二行A_FWQ(0),A_FWQ(1)…A_FWQ(N-1)为A频段前向数据的正交分量，作为虚部。B频段前向数据向A频段反馈数据的传递参数矩阵为： $B_{FW} = [\begin{matrix} B_{FW} I (0) & B_{FW} I (1) & B_{FW} I (2) & . . . & B_{FW} I (N - 2) & B_{FW} I (N - 1) \\ B_{FW} Q (0) & B_{FW} Q (1) & B_{FW} Q (2) & . . . & B_{FW} Q (N - 2) & B_{FW} Q (N - 1) \end{matrix}],$ 其中，第一行B_FWI(0),B_FWI(1)…B_FWI(N-1)为B频段前向数据的同相分量，作为实部；第二行B_FWQ(0),B_FWQ(1)…B_FWQ(N-1)为B频段前向数据的正交分量，作为虚部。A频段前向反馈数据的传递参数矩阵为： $A_{FB} = [\begin{matrix} A_{FB} I (0) & A_{FB} I (1) & A_{FB} I (2) & . . . & A_{FB} I (N - 2) & A_{FB} I (N - 1) \\ A_{FB} Q (0) & A_{FB} Q (1) & A_{FB} Q (2) & . . . & A_{FB} Q (N - 2) & A_{FB} Q (N - 1) \end{matrix}],$ 其中，第一行A_FBI(0),A_FBI(1)…A_FBI(N-1)为A频段前向反馈数据的同相分量，作为实部；第二行A_FBQ(0),A_FBQ(1)…A_FBQ(N-1)为A频段前向反馈数据的正交分量，作为虚部。根据公式α(A_FW-α^-1βB_FW)+βB_FW＝αA_FW，对发射信号在零中频进行补偿，消除镜像的影响。The present invention also proposes a method for correcting IQ imbalance, said method comprising the steps of calculating the DC offset factor, gain imbalance factor and phase of the in-phase signal and the quadrature signal according to the input in-phase signal and the quadrature signal unbalance factor, and perform DC compensation; the predistorter and the IQ compensator are disabled so that u(n)=z(n)=x(n), where u(n) is the signal before entering the predistorter , z(n) is the output signal of the predistorter, x(n) is the signal after the IQ compensator; the signal y(n) obtained by disabling the power amplifier PA to obtain the feedback signal after quadrature demodulation; through x (n) and y(n) difference channel estimation and updated gain and phase imbalance factor, obtain preset IQ compensator gain and phase imbalance factor; activate IQ compensator, enable PA, start PD training sequence, Inverse balance processing is performed on the output IQ-corrected in-phase signal and quadrature signal, and the processed unbalanced in-phase signal and quadrature signal are obtained and output. When it is necessary to eliminate the image signal of the B frequency band in the A frequency band in the system, the image generation model formula is: αA _FW + βB _FW = A _FB , where A _FW is the forward output signal of the A frequency band, and B _FW is the forward output signal of the B frequency band. A _FB is the forward feedback signal of the A frequency band, α represents the gain and phase coefficient of the influence of the transmitted signal of the A frequency band on itself, and β represents the gain and phase coefficient of the influence of the A frequency band on the B frequency band. The DC offset correction specifically includes calling the DC offset equation according to the DC signals _C1 and _C2 $\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{twenty one} + C_{c 2} h_{twenty two} = - C_{2} \end{matrix},$ Calculate the DC offset factors C _c1 and C _c2 to establish the DC offset matrix $[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{twenty one} & h_{twenty two} \end{matrix}]}^{- 1} [\begin{matrix} - C_{1} \\ - C_{2} \end{matrix}],$ Compensate in the IQ compensator with a DC offset matrix, where h ₁₁ , h ₁₂ , h ₂₁ , h ₂₂ are the frequencies f ₁₁ , f ₁₂ , f ₂₁ of the signal entering the IQ compensator from the output of the predistorter, respectively, f ₂₂ 's response. The transfer parameter matrix of the A-band forward data to the A-band feedback data is: $A_{FW} = [\begin{matrix} A_{FW} I (0) & A_{FW} I (1) & A_{FW} I (2) & . . . & A_{FW} I (N - 2) & A_{FW} I (N - 1) \\ A_{FW} Q (0) & A_{FW} Q (1) & A_{FW} Q (2) & . . . & A_{FW} Q (N - 2) & A_{FW} Q (N - 1) \end{matrix}],$ Among them, the first line A _FW I(0), A _FW I(1)...A _FW I(N-1) is the in-phase component of the forward data of the A frequency band, as the real part; the second line A _FW Q(0) ,A _FW Q(1)...A _FW Q(N-1) is the quadrature component of the forward data in the A frequency band, which is used as the imaginary part. The transfer parameter matrix of B-band forward data to A-band feedback data is: $B_{FW} = [\begin{matrix} B_{FW} I (0) & B_{FW} I (1) & B_{FW} I (2) & . . . & B_{FW} I (N - 2) & B_{FW} I (N - 1) \\ B_{FW} Q (0) & B_{FW} Q (1) & B_{FW} Q (2) & . . . & B_{FW} Q (N - 2) & B_{FW} Q (N - 1) \end{matrix}],$ Among them, the first line B _FW I(0), B _FW I(1)...B _FW I(N-1) is the in-phase component of the B-band forward data, as the real part; the second line B _FW Q(0) , B _FW Q(1)...B _FW Q(N-1) is the quadrature component of the B-band forward data, which is used as the imaginary part. The transfer parameter matrix of the A-band forward feedback data is: $A_{Facebook} = [\begin{matrix} A_{Facebook} I (0) & A_{Facebook} I (1) & A_{Facebook} I (2) & . . . & A_{Facebook} I (N - 2) & A_{Facebook} I (N - 1) \\ A_{Facebook} Q (0) & A_{Facebook} Q (1) & A_{Facebook} Q (2) & . . . & A_{Facebook} Q (N - 2) & A_{Facebook} Q (N - 1) \end{matrix}],$ Among them, the first line A _FB I(0), A _FB I(1)...A _FB I(N-1) is the in-phase component of the A-band forward feedback data, as the real part; the second line A _FB Q(0 ), A _FB Q(1)...A _FB Q(N-1) is the quadrature component of the A-band forward feedback data, as the imaginary part. According to the formula α(A _FW -α ^-1 βB _FW )+βB _FW =αA _FW , the transmitted signal is compensated at zero intermediate frequency to eliminate the influence of the image.

进行直流补偿之前可进行数据初始化处理。通过去使能预失真器、IQ补偿器和PA获得预设的幅度和相位不平衡因子，对输出的经过IQ补偿器补偿的同相信号及正交信号进行逆平衡处理，得到不平衡的同相信号和正交信号，使得u(n)＝z(n)＝x(n)。经去使能功率放大器PA得到反馈信号经过正交解调后的信号y(n)，信号y(n)即为基带信号经过数字上变频后的数据；通过x(n)和y(n)的差值信道估计及上一次更新的增益和相位不平衡因子，获得预设IQ补偿器幅度和相位不平衡因子，激活IQ补偿器，使能PA、启动PD训练序列，对经过IQ校正的同相信号及正交信号进行逆平衡处理，得到经过处理的不平衡的同相信号和正交信号，并将其输出。Data initialization can be performed before DC compensation. By disabling the predistorter, IQ compensator and PA, the preset amplitude and phase imbalance factors are obtained, and the output in-phase signal and quadrature signal compensated by the IQ compensator are subjected to inverse balance processing to obtain an unbalanced synchronous signal. Phase signal and quadrature signal such that u(n)=z(n)=x(n). After disabling the power amplifier PA to obtain the signal y(n) of the feedback signal after quadrature demodulation, the signal y(n) is the data of the baseband signal after digital up-conversion; through x(n) and y(n) The difference channel estimate and the gain and phase imbalance factor of the last update, obtain the preset IQ compensator amplitude and phase imbalance factor, activate the IQ compensator, enable the PA, start the PD training sequence, and perform IQ correction on the same The phase signal and the quadrature signal are inversely balanced, and the processed unbalanced in-phase signal and quadrature signal are obtained and output.

其中，直流偏移因子可采用如下方法获得：Among them, the DC offset factor can be obtained by the following method:

假设原始信号复数表示为：Assuming the complex representation of the original signal is:

z(t)＝cosωt+jsinωt (4)z(t)＝cosωt+jsinωt (4)

经过调制后由于IQ增益和相位不平衡以及直流偏移得到失真信号为：After modulation, the distorted signal due to IQ gain and phase imbalance and DC offset is:

y(t)＝[αcos(ωt+ω_ct+φ)+C₁]+j[βsin(ωt+ω_ct+θ)+C₂] (5)y(t)＝[αcos(ωt+ _ωc t+φ)+C ₁ ]+j[βsin(ωt+ _ωc t+θ)+C ₂ ] (5)

其中α，β为增益不平衡系数，φ，θ为相位不平衡系数，C₁,C₂为直流信号，ω，ω_c为原始信号和调制信号的角频率，t为信号时域表示的时间单位。通过IQ补偿器后直流信号C₁,C₂应完全被消除。可采用直流偏移校正。Among them, α, β are gain imbalance coefficients, φ, θ are phase imbalance coefficients, C ₁ , C ₂ are DC signals, ω, ω _c are the angular frequencies of the original signal and modulated signal, and t is the time expressed in the time domain of the signal unit. After passing through the IQ compensator, the DC signals C ₁ and C ₂ should be completely eliminated. DC offset correction is available.

其中，调用如下公式进行所述直流偏移校正：Wherein, the following formula is called to perform the DC offset correction:

C_c1h₁₁+C_c2h₁₂＝-C₁ C _c1 h ₁₁ +C _c2 h ₁₂ ＝-C ₁

C_c1h₂₁+C_c2h₂₂＝-C₂ (6)C _c1 h ₂₁ +C _c2 h ₂₂ ＝-C ₂ (6)

其中，C_c1和C_c2为直流偏移因子，h₁₁，h₁₂，h₂₁，h₂₂分别为信号从预失真器输出进入IQ补偿器的信号频率(f₁₁，f₁₂，f₂₁，f₂₂)的响应。Among them, C _c1 and C _c2 are the DC offset factors, h ₁₁ , h ₁₂ , h ₂₁ , h ₂₂ are the signal frequencies (f ₁₁ , f ₁₂ , f ₂₁ , f ₂₂ ) Response.

求解出C_c1和C_c2的偏移，用矩阵表示为：Solve the offsets of C _c1 and C _c2 , expressed as a matrix:

$[\begin{matrix} {C C}_{c c 11} \\ {C C}_{c c 22} \end{matrix}] = = {[\begin{matrix} {h h}_{1111} & {h h}_{1212} \\ {h h}_{21 twenty one} & {h h}_{22 twenty two} \end{matrix}]}^{- - 11} [\begin{matrix} - - {C C}_{11} \\ - - {C C}_{22} \end{matrix}] - - - - - - ((77))$

其中，根据输入的同相信号和正交信号计算同相信号和正交信号的增益不平衡系数α，β，相位不平衡系数φ，θ。采集各个频段的前向输出信号和反馈信号计算对应的增益和相位不平衡因子。Wherein, according to the input in-phase signal and quadrature signal, the gain imbalance coefficients α, β and phase imbalance coefficients φ, θ of the in-phase signal and quadrature signal are calculated. The forward output signal and feedback signal of each frequency band are collected to calculate the corresponding gain and phase imbalance factors.

采用本发明的方法，基于数字预失真系统，不需要附加额外的反馈链路，结构简单、容易实现、实时性好，在降低系统硬件复杂度和成本的同时还提高了误差估计和补偿的精度。本发明贴近实际，操作简单，能够较好的指导完成IQ不平衡的校正过程，具有较好的实际利用价值和经济效益。Adopting the method of the present invention, based on the digital predistortion system, no additional feedback link is needed, the structure is simple, easy to implement, and the real-time performance is good, while reducing the complexity and cost of system hardware, it also improves the accuracy of error estimation and compensation . The invention is close to reality, has simple operation, can better guide and complete the correction process of IQ imbalance, and has better practical use value and economic benefit.

本发明的其他优点、目标和特征在某种程度上将在随后的说明书中进行阐述，并且在某种程度上，基于对下文的考察研究对本领域技术人员而言将是显而易见的，或者可以从本发明的实践中得到教导。本发明的目标和其他优点可以通过下面的说明书和权利要求书来实现和获得。Other advantages, objects and features of the present invention will be set forth in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the investigation and research below, or can be obtained from It is taught in the practice of the present invention. The objects and other advantages of the invention will be realized and attained by the following description and claims.

附图说明Description of drawings

图1本发明提供的一种IQ不平衡校正的装置结构图。FIG. 1 is a structure diagram of an IQ imbalance correction device provided by the present invention.

图2本发明方法中IQ不平衡校正的流程示意图；The schematic flow chart of IQ imbalance correction in the method of the present invention in Fig. 2;

图3本发明方法中数据初始化的流程示意图；The schematic flow chart of data initialization in the method of the present invention in Fig. 3;

图4正交调制框图；Fig. 4 quadrature modulation block diagram;

图5本发明方法中用于直流补偿的IQ补偿器框图；The IQ compensator block diagram that is used for DC compensation in the method of the present invention in Fig. 5;

图6本发明方法中镜像校正的流程示意图。Fig. 6 is a schematic flow chart of image correction in the method of the present invention.

具体实施方式Detailed ways

为了使本发明的目的，技术方案及优点更加清楚明白，以下结合附图通过具体实施例对本发明作进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本发明，并不用于限定本发明。In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

下面结合具体实施例及附图对本发明提出的IQ不平衡校正方法进一步予以阐述，以便使本发明的精神，技术特征及有益效果更易于了解。The IQ imbalance correction method proposed by the present invention will be further described below in conjunction with specific embodiments and accompanying drawings, so as to make the spirit, technical features and beneficial effects of the present invention easier to understand.

图1所示为本发明提供的一种IQ不平衡校正的数字预失真系统结构图。具体包括：数据接收模块、模数转换器ADC、数字预失真模块、数模转换器DAC和射频处理模块。当数据接收处理模块中的天线接收到微弱信号时，经过低噪声放大器LNA的放大，再由IQ正交解调器下变频到中频模拟信号，再经过一个模数转换器ADC把模拟的中频信号转换成数字的中频信号送入到预失真模块，该模块对信号进行处理实现本发明的所有功能，包括直流补偿、镜像校正等功能，然后数模转换器DAC把经过数字预失真模块处理的数字信号转换成模拟信号再经过射频处理模块进行正交上变频等射频处理，最后经过功率放大器发送出去。FIG. 1 is a structural diagram of a digital predistortion system for IQ imbalance correction provided by the present invention. Specifically include: data receiving module, analog-to-digital converter ADC, digital pre-distortion module, digital-to-analog converter DAC and radio frequency processing module. When the antenna in the data receiving and processing module receives a weak signal, it is amplified by the low noise amplifier LNA, then down-converted to an intermediate frequency analog signal by the IQ quadrature demodulator, and then converted to an analog intermediate frequency signal by an analog-to-digital converter ADC. The intermediate frequency signal converted into digital is sent to the pre-distortion module, which processes the signal to realize all functions of the present invention, including functions such as DC compensation and image correction, and then the digital-to-analog converter DAC converts the digital signal processed by the digital pre-distortion module The signal is converted into an analog signal, then undergoes radio frequency processing such as quadrature up-conversion by the radio frequency processing module, and finally sends it out through the power amplifier.

图2所示为本发明IQ不平衡校正方法流程示意图，包括如下步骤：Fig. 2 is a schematic flow chart of the IQ imbalance correction method of the present invention, including the following steps:

步骤201、数据初始化处理；Step 201, data initialization processing;

步骤202、根据输入的同相信号和正交信号计算同相信号和正交信号的直流偏移因子并进行直流补偿；根据输入的同相信号和正交信号计算同相信号和正交信号的增益不平衡因子和相位不平衡因子；Step 202, calculate the DC offset factor of the in-phase signal and the quadrature signal according to the input in-phase signal and the quadrature signal and perform DC compensation; calculate the in-phase signal and the quadrature signal according to the input in-phase signal and the quadrature signal Gain imbalance factor and phase imbalance factor;

步骤204、建立镜像校正模型根据相应参数进行镜像校正；Step 204, establishing a mirror correction model to perform mirror correction according to corresponding parameters;

步骤205、根据更新补偿后的增益和相位因子完成IQ不平衡校正。Step 205, complete IQ imbalance correction according to the updated and compensated gain and phase factors.

激活IQ补偿器，使能PA、启动PD训练序列，对经过不平衡IQ校正的同相信号及正交信号进行逆平衡处理，得到经过处理的不平衡的同相信号和正交信号，并将其输出。Activate the IQ compensator, enable the PA, start the PD training sequence, perform inverse balance processing on the unbalanced IQ corrected in-phase signal and quadrature signal, obtain the processed unbalanced in-phase signal and quadrature signal, and its output.

可采用如下方法进行初始化，图3所示为本发明实施例中数据初始化流程示意图，包括如下步骤：The following method can be used for initialization. Figure 3 is a schematic diagram of the data initialization process in the embodiment of the present invention, including the following steps:

步骤301，获取预失真器之前的信号u(n)，经过预失真器的输出信号z(n)，IQ补偿器后的信号x(n)，将上述信号去使能预失真器PD和IQ补偿器处理。使得u(n)＝z(n)＝x(n)；Step 301, obtain the signal u(n) before the predistorter, pass the output signal z(n) of the predistorter, and the signal x(n) after the IQ compensator, and disable the above signals to enable the predistorter PD and IQ Compensator processing. so that u(n)=z(n)=x(n);

步骤302，去使能功率放大器PA得到反馈信号经过正交解调后的信号y(n)，该信号即为基带信号经过数字上变频后的数据；Step 302, disabling the power amplifier PA to obtain the signal y(n) of the feedback signal after quadrature demodulation, which is the data of the baseband signal after digital up-conversion;

步骤303、通过x(n)和y(n)的差值信道估计及更新的增益和相位不平衡因子，获得预设IQ补偿器增益和相位不平衡因子。Step 303. Obtain the preset gain and phase imbalance factors of the IQ compensator through the difference channel estimation of x(n) and y(n) and the updated gain and phase imbalance factors.

图4为正交调制框图。假设原始信号复数表示为：Figure 4 is a block diagram of quadrature modulation. Assuming the complex representation of the original signal is:

z(t)＝cosωt+jsinωt (11)z(t)＝cosωt+jsinωt (11)

其中，z(t)为原始信号的复数，cosωt为信号实部，jsinωt为信号虚部，ω为信号角频率，t为时间。经过调制后由于IQ增益和相位不平衡以及直流偏移得到失真信号为：Among them, z(t) is the complex number of the original signal, cosωt is the real part of the signal, jsinωt is the imaginary part of the signal, ω is the angular frequency of the signal, and t is the time. After modulation, the distorted signal due to IQ gain and phase imbalance and DC offset is:

y(t)＝[αcos(ωt+ω_ct+φ)+C₁]+j[βsin(ωt+ω_ct+θ)+C₂] (12)y(t)＝[αcos(ωt+ _ωc t+φ)+C ₁ ]+j[βsin(ωt+ _ωc t+θ)+C ₂ ] (12)

其中α，β增益不平衡，φ，θ为相位不平衡，C₁,C₂为直流信号。通过IQ补偿器后直流信号C₁,C₂应完全被消除。Among them, α and β gain imbalance, φ and θ are phase imbalance, and C ₁ and C ₂ are DC signals. After passing through the IQ compensator, the DC signals C ₁ and C ₂ should be completely eliminated.

图5为本发明方法中用于直流补偿的IQ补偿器框图。通过直流因子C和IQ补偿器完成根据输入的同相信号和正交信号计算同相信号和正交信号的直流偏移因子并进行直流补偿。Fig. 5 is a block diagram of an IQ compensator used for DC compensation in the method of the present invention. The DC offset factor of the in-phase signal and the quadrature signal is calculated according to the input in-phase signal and the quadrature signal through the DC factor C and the IQ compensator, and DC compensation is performed.

图5划分为两个虚框，在此图中略去DAC模块，z_i(n)，z_q(n)分别表示进入补偿器的I路和Q路信号，α，β分别表示两路的增益，为相位。直流偏移校正如下：Figure 5 is divided into two imaginary boxes, the DAC module is omitted in this figure, z _i (n), z _q (n) respectively represent the I-channel and Q-channel signals entering the compensator, α, β respectively represent the gains of the two channels , for the phase. The DC offset correction is as follows:

建立方程 $\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{21} + C_{c 2} h_{22} = - C_{2} \end{matrix} - - - (13)$ 计算直流偏移因子C_c1和C_c2。其中，C₁和C₂为直流信号，h₁₁，h₁₂，h₂₁，h₂₂分别为信号从预失真器输出进入IQ补偿器的信号的频率f₁₁，f₁₂，f₂₁，f₂₂的响应。解出C_c1和C_c2的偏移用矩阵表示为： $[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{21} & h_{22} \end{matrix}]}^{- 1} [\begin{matrix} - C_{1} \\ - C_{2} \end{matrix}] - - - (14)$ build equation $\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{twenty one} + C_{c 2} h_{twenty two} = - C_{2} \end{matrix} - - - (13)$ Calculate the DC offset factors C _c1 and C _c2 . Among them, C ₁ and C ₂ are DC signals, h ₁₁ , h ₁₂ , h ₂₁ , and h ₂₂ are the frequencies f ₁₁ , f ₁₂ , f ₂₁ , and f ₂₂ of the signals output from the predistorter to the IQ compensator, respectively. response. Solving the offsets of C _c1 and C _c2 is expressed as a matrix: $[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{twenty one} & h_{twenty two} \end{matrix}]}^{- 1} [\begin{matrix} - C_{1} \\ - C_{2} \end{matrix}] - - - (14)$

以消除A频段内的镜像信号为例，其镜像生成模型为：αA_FW+βB_FW+δC_FW+…+ηD_FW＝A_FB，将A,B,C,……,D各个频段的前向数据(A_FW,B_FW,C_FW,……D_FW)和A频段的反馈数据A_FB的传递参数以矩阵的形式表示出来，求解各个频段传递参数的系数α,β,δ…η，实现镜像抑制。Taking the elimination of the image signal in the A frequency band as an example, the image generation model is: αA _FW + βB _FW + δC _FW +...+ηD _FW ＝A _FB , the forward The data (A _FW , B _FW , C _FW ,...D _FW ) and the feedback data of the A frequency band A _FB transfer parameters are expressed in the form of a matrix, and the coefficients α, β, δ...η of the transfer parameters of each frequency band are solved to realize Image suppression.

图6为本发明镜像校正的流程示意图。本示范性实施例以两个频段A频段和B频段为例说明镜像校正过程：FIG. 6 is a schematic flow chart of image correction in the present invention. In this exemplary embodiment, two frequency bands A and B are taken as examples to illustrate the image correction process:

步骤601，在第一频段A内发送前向序列，在第二频段B内发送前向序列。In step 601, the forward sequence is transmitted in the first frequency band A, and the forward sequence is transmitted in the second frequency band B.

步骤602，采集所述A频段的前向输出信号A_FW、B频段的前向输出信号B_FW和A频段的前向反馈信号A_FB。Step 602, collect the forward output signal A _FW of the A frequency band, the forward output signal B _FW of the B frequency band and the forward feedback signal A _FB of the A frequency band.

A频段前向数据向A频段反馈数据的传递参数矩阵为：The transfer parameter matrix of the A-band forward data to the A-band feedback data is:

${A A}_{FW FW} = = [\begin{matrix} {A A}_{FW FW} I I ((00)) & {A A}_{FW FW} I I ((11)) & {A A}_{FW FW} I I ((22)) & . . . . . . & {A A}_{FW FW} I I ((N N - - 22)) & {A A}_{FW FW} I I ((N N - - 11)) \\ {A A}_{FW FW} Q Q ((00)) & {A A}_{FW FW} Q Q ((11)) & {A A}_{FW FW} Q Q ((22)) & . . . . . . & {A A}_{FW FW} Q Q ((N N - - 22)) & {A A}_{FW FW} Q Q ((N N - - 11)) \end{matrix}] - - - - - - ((1515))$

其中，第一行A_FWI(0),A_FWI(1)…A_FWI(N-1)为A频段前向数据的同相分量，为实部；第二行A_FWQ(0),A_FWQ(1)…A_FWQ(N-1)为A频段前向数据的正交分量，为虚部。Among them, the first line A _FW I(0), A _FW I(1)...A _FW I(N-1) is the in-phase component of the forward data of the A frequency band, which is the real part; the second line A _FW Q(0) ,A _FW Q(1)...A _FW Q(N-1) is the quadrature component of the forward data in the A frequency band, which is the imaginary part.

B频段前向数据向A频段反馈数据的传递参数矩阵为：The transfer parameter matrix of B-band forward data to A-band feedback data is:

${B B}_{FW FW} = = [\begin{matrix} {B B}_{FW FW} I I ((00)) & {B B}_{FW FW} I I ((11)) & {B B}_{FW FW} I I ((22)) & . . . . . . & {B B}_{FW FW} I I ((N N - - 22)) & {B B}_{FW FW} I I ((N N - - 11)) \\ {B B}_{FW FW} Q Q ((00)) & {B B}_{FW FW} Q Q ((11)) & {B B}_{FW FW} Q Q ((22)) & . . . . . . & {B B}_{FW FW} Q Q ((N N - - 22)) & {B B}_{FW FW} Q Q ((N N - - 11)) \end{matrix}] - - - - - - ((1616))$

其中，第一行B_FWI(0),B_FWI(1)…B_FWI(N-1)为B频段前向数据的同相分量，为实部；第二行B_FWQ(0),B_FWQ(1)…B_FWQ(N-1)为B频段前向数据的正交分量，为虚部。Among them, the first line B _FW I(0), B _FW I(1)...B _FW I(N-1) is the in-phase component of the B-band forward data, which is the real part; the second line B _FW Q(0) ,B _FW Q(1)...B _FW Q(N-1) is the quadrature component of the B-band forward data, which is the imaginary part.

A频段前向反馈数据的传递参数矩阵为：The transfer parameter matrix of the A-band forward feedback data is:

${A A}_{FB Facebook} = = [\begin{matrix} {A A}_{FB Facebook} I I ((00)) & {A A}_{FB Facebook} I I ((11)) & {A A}_{FB Facebook} I I ((22)) & . . . . . . & {A A}_{FB Facebook} I I ((N N - - 22)) & {A A}_{FB Facebook} I I ((N N - - 11)) \\ {A A}_{FB Facebook} Q Q ((00)) & {A A}_{FB Facebook} Q Q ((11)) & {A A}_{FB Facebook} Q Q ((22)) & . . . . . . & {A A}_{FB Facebook} Q Q ((N N - - 22)) & {A A}_{FB Facebook} Q Q ((N N - - 11)) \end{matrix}] - - - - - - ((1717))$

其中，第一行A_FBI(0),A_FBI(1)…A_FBI(N-1)为A频段前向反馈数据的同相分量，为实部；第二行A_FBQ(0),A_FBQ(1)…A_FBQ(N-1)为A频段前向反馈数据的正交分量，为虚部。Among them, the first line A _FB I(0), A _FB I(1)...A _FB I(N-1) is the in-phase component of the A-band forward feedback data, which is the real part; the second line A _FB Q(0 ), A _FB Q(1)...A _FB Q(N-1) is the quadrature component of the A-band forward feedback data, which is the imaginary part.

步骤603，根据A频段的前向输出信号A_FW、B频段的前向输出信号B_FW和A频段的前向反馈信号A_FB，按照镜像生成模型公式αA_FW+βB_FW＝A_FB计算镜像模型传递参数系数α和β。建立α和β的影响系数矩阵， $α = [\begin{matrix} α_{11} & α_{12} \\ α_{21} & α_{22} \end{matrix}],$ 表示正交混频器的不平衡导致A频段发射信号对自身影响的增益系数矩阵，其中，α₁₁,α₁₂,α₂₁,α₂₂为相应的增益系数， $β = [\begin{matrix} β_{11} & β_{12} \\ β_{21} & β_{22} \end{matrix}],$ 表示正交混频器的不平衡导致A频段对B频段影响的增益系数矩阵，β₁₁,β₁₂,β₂₁,β₂₂为相应的增益系数。Step 603, according to the forward output signal A _FW of the A frequency band, the forward output signal B _FW of the B frequency band and the forward feedback signal A _FB of the A frequency band, calculate the mirror image model according to the mirror image generation model formula αA _FW +βB _FW =A _FB Pass parameter coefficients α and β. Establish the influence coefficient matrix of α and β, $α = [\begin{matrix} α_{11} & α_{12} \\ α_{twenty one} & α_{twenty two} \end{matrix}],$ Indicates the gain coefficient matrix of the influence of the A-band transmitted signal on itself due to the imbalance of the quadrature mixer, where α ₁₁ , α ₁₂ , α ₂₁ , and α ₂₂ are the corresponding gain coefficients, $β = [\begin{matrix} β_{11} & β_{12} \\ β_{twenty one} & β_{twenty two} \end{matrix}],$ It represents the gain coefficient matrix of the influence of the A frequency band on the B frequency band due to the imbalance of the quadrature mixer, and β ₁₁ , β ₁₂ , β ₂₁ , and β ₂₂ are the corresponding gain coefficients.

将α和β的影响系数矩阵以及A_FW、B_FW和A_FB的传递参数矩阵分别代入镜像生成模型，公式展开为：The influence coefficient matrices of α and β and the transfer parameter matrices of A _FW , B _FW and A _FB are respectively substituted into the image generation model, and the formula is expanded as:

$\begin{matrix} [\begin{matrix} {α α}_{1111} & {α α}_{1212} & {β β}_{1111} & {β β}_{1212} \\ {α α}_{21 twenty one} & {α α}_{22 twenty two} & {β β}_{21 twenty one} & {β β}_{22 twenty two} \end{matrix}] [\begin{matrix} {A A}_{FW FW} I I ((00)) & {A A}_{FW FW} I I ((11)) & {A A}_{FW FW} I I ((22)) & . . . . . . & {A A}_{FW FW} I I ((N N - - 22)) & {A A}_{FW FW} I I ((N N - - 11)) \\ {A A}_{FW FW} Q Q ((00)) & {A A}_{FW FW} Q Q ((11)) & {A A}_{FW FW} Q Q ((22)) & . . . . . . & {A A}_{FW FW} Q Q ((N N - - 22)) & {A A}_{FW FW} Q Q ((N N - - 11)) \\ {B B}_{FW FW} I I ((00)) & {B B}_{FW FW} I I ((11)) & {B B}_{FW FW} I I ((22)) & . . . . . . & {B B}_{FW FW} Q Q ((N N - - 22)) & {B B}_{FW FW} Q Q ((N N - - 11)) \\ {B B}_{FW FW} Q Q ((00)) & {B B}_{FW FW} Q Q ((11)) & {B B}_{FW FW} Q Q ((22)) & . . . . . . & {B B}_{FW FW} Q Q ((N N - - 22)) & {B B}_{FW FW} Q Q ((N N - - 11)) \end{matrix}] \\ = = [\begin{matrix} {A A}_{FB Facebook} I I ((00)) & {A A}_{FB Facebook} I I ((11)) & {A A}_{FB Facebook} I I ((22)) & . . . . . . & {A A}_{FB Facebook} I I ((N N - - 22)) & {A A}_{FB Facebook} I I ((N N - - 11)) \\ {A A}_{FB Facebook} Q Q ((00)) & {A A}_{FB Facebook} Q Q ((22)) & {A A}_{FB Facebook} Q Q ((22)) & . . . . . . & {A A}_{FB Facebook} Q Q ((N N - - 22)) & {A A}_{FB Facebook} Q Q ((N N - - 11)) \end{matrix}] \end{matrix}$

解上面的矩阵，令Solving the above matrix, let

$c c = = [\begin{matrix} {α α}_{1111} & {α α}_{1212} & {β β}_{1111} & {β β}_{1212} \\ {α α}_{21 twenty one} & {α α}_{22 twenty two} & {β β}_{21 twenty one} & {β β}_{22 twenty two} \end{matrix}]$

$H h = = [\begin{matrix} {A A}_{FW FW} I I ((00)) & {A A}_{FW FW} I I ((11)) & {A A}_{FW FW} I I ((22)) & . . . . . . & {A A}_{FW FW} I I ((N N - - 22)) & {A A}_{FW FW} I I ((N N - - 11)) \\ {A A}_{FW FW} Q Q ((00)) & {A A}_{FW FW} Q Q ((11)) & {A A}_{FW FW} Q Q ((22)) & . . . . . . & {A A}_{FW FW} Q Q ((N N - - 22)) & {A A}_{FW FW} Q Q ((N N - - 11)) \\ {B B}_{FW FW} I I ((00)) & {B B}_{FW FW} I I ((11)) & {B B}_{FW FW} I I ((22)) & . . . . . . & {B B}_{FW FW} I I ((N N - - 22)) & {B B}_{FW FW} I I ((N N - - 11)) \\ {B B}_{FW FW} Q Q ((00)) & {B B}_{FW FW} Q Q ((11)) & {B B}_{FW FW} Q Q ((22)) & . . . . . . & {B B}_{FW FW} Q Q ((N N - - 22)) & {B B}_{FW FW} Q Q ((N N - - 11)) \end{matrix}]$

$P P = = [\begin{matrix} {A A}_{FB Facebook} I I ((00)) & {A A}_{FB Facebook} I I ((11)) & {A A}_{FB Facebook} I I ((22)) & . . . . . . & {A A}_{FB Facebook} I I ((N N - - 22)) & {A A}_{FB Facebook} I I ((N N - - 11)) \\ {A A}_{FB Facebook} Q Q ((00)) & {A A}_{FB Facebook} Q Q ((11)) & {A A}_{FB Facebook} Q Q ((22)) & . . . . . . & {A A}_{FB Facebook} Q Q ((N N - - 22)) & {A A}_{FB Facebook} Q Q ((N N - - 11)) \end{matrix}]$

则公式简化为CH＝P，求解矩阵C。Then the formula is simplified to CH=P, and the matrix C is solved.

CH＝PCH=P

CHH^H＝PH^H CHH ^H =PH ^H

C＝PH^H(HH^H)^-1 C＝PH ^H (HH ^H ) ^-1

其中，H^H为矩阵H的转置矩阵。求解出影响系数α和β后，将A_FW减去α^-1βB_FW，使表达式变为：Among them, H ^H is the transpose matrix of matrix H. After solving the influence coefficients α and β, subtract α ^-1 βB _FW from A _FW , so that the expression becomes:

α(A_FW-α^-1βB_FW)+βB_FW＝αA_FW α(A _FW -α ^-1 βB _FW )+βB _FW ＝αA _FW

步骤604，当需要在A、B频段发射数据时，根据上述公式利用影响系数对发射信号在零中频进行系数补偿，如此便可消除镜像的影响。Step 604, when it is necessary to transmit data in the A and B frequency bands, use the influence coefficient to perform coefficient compensation on the transmission signal at the zero intermediate frequency according to the above formula, so that the influence of the image can be eliminated.

可见，本发明基于数字预失真系统，不需要附加额外的反馈链路，结构简单、容易实现、实时性好，在降低系统硬件复杂度和成本的同时还提高了误差估计和补偿的精度。It can be seen that the present invention is based on a digital predistortion system, does not require an additional feedback link, has a simple structure, is easy to implement, and has good real-time performance. It also improves the accuracy of error estimation and compensation while reducing system hardware complexity and cost.

本发明所列举的实施方式如上所述，但所述的内容只是为了便于理解本发明而采用的一个案例，并非用以限定本发明。在不背离本发明思想以及实质的情况下，熟悉本领域的技术人员可根据本发明在实施的形式上或细节上做出各种相应的修改和变化，但本发明的专利保护范围，仍须以所附的权利要求书所界定的范围为准。The embodiments listed in the present invention are as described above, but the content described is only a case adopted for easy understanding of the present invention, and is not intended to limit the present invention. Without departing from the idea and essence of the present invention, those skilled in the art can make various corresponding modifications and changes in the form or details of the implementation according to the present invention, but the patent protection scope of the present invention still needs to be The scope defined by the appended claims shall prevail.

Claims

1. A device for correcting IQ imbalance in a digital predistortion system, characterized in that the device comprises: a data receiving module, an analog-to-digital converter, a digital predistortion module, a digital-to-analog converter and a radio frequency processing module, and a data receiving module The received weak signal is sent to the low-noise amplifier for amplification, and the frequency is down-converted to the intermediate frequency analog signal by the IQ quadrature demodulator; the analog-to-digital converter converts the intermediate frequency analog signal into an intermediate frequency digital signal; the digital pre-distortion module processes the intermediate frequency digital signal and corrects the power The non-linearity of the amplifier and the local oscillator leakage and image correction of the modulator; the digital-to-analog converter converts the intermediate frequency digital signal processed by the digital pre-distortion module into an analog signal; the radio frequency processing module performs quadrature modulation up-conversion processing on the converted analog signal .

2. The device according to claim 1, wherein the digital pre-distortion module comprises: a pre-distorter, an IQ compensator, and a delay module, and the pre-distorter receives the quadrature-modulated in-phase signal and the quadrature signal, the IQ compensator calculates the respective gain factor, amplitude imbalance factor and phase imbalance factor according to the in-phase signal and the quadrature signal, and the predistorter and the IQ compensator are disabled, so that u(n)=z(n) =x(n), wherein, u(n) is the signal before entering the predistorter, z(n) is the output signal through the predistorter, x(n) is the signal after the IQ compensator; update gain and phase Unbalance factor, obtain the preset IQ compensator gain and phase imbalance factors, and perform DC offset correction; establish an image correction model to perform image correction on the above factors to suppress the image signal; activate the IQ compensator, enable PA, and start PD training The sequence performs inverse balance processing on the in-phase signal and quadrature signal after IQ correction, and outputs the unbalanced in-phase signal and quadrature signal.

3. The device according to claim 2, wherein, when the image signal of the B frequency band in the A frequency band needs to be eliminated in the system, the image correction model formula is: α A _FW + β B _FW = A _FB , wherein A _FW is The forward output signal of the A frequency band, B _FW is the forward output signal of the B frequency band, A _FB is the forward feedback signal of the A frequency band, α represents the gain and phase coefficient of the influence of the transmitted signal of the A frequency band on itself, and β represents the influence of the A frequency band on the Gain and phase coefficients for B-band effects.

4. The device according to claim 2, wherein the DC offset correction is specifically, calling the DC offset equation according to the DC signals _C1 and _C2

\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{twenty one} + C_{c 2} h_{twenty two} = - C_{2} \end{matrix},

Calculate the DC offset factors C _c1 and C _c2 to establish the DC offset matrix

[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{twenty one} & h_{twenty two} \end{matrix}]}^{- 1} [\begin{matrix} {- C}_{1} \\ {- C}_{2} \end{matrix}],

Compensate in the IQ compensator with a DC offset matrix, where h ₁₁ , h ₁₂ , h ₂₁ , h ₂₂ are the frequencies f ₁₁ , f ₁₂ , f ₂₁ of the signal entering the IQ compensator from the output of the predistorter, respectively, f ₂₂ 's response.

5. A method for IQ imbalance correction, characterized in that said method comprises the steps of calculating the DC offset factor, the gain imbalance factor of the in-phase signal and the quadrature signal according to the input in-phase signal and the quadrature signal and phase imbalance factor, and perform DC compensation; the predistorter and the IQ compensator are disabled, so that u(n)=z(n)=x(n), where u(n) is before entering the predistorter The signal, z (n) is the output signal through the predistorter, x (n) is the signal after the IQ compensator; to enable the power amplifier PA to obtain the feedback signal y (n) after quadrature demodulation; Through the difference channel estimation of x(n) and y(n) and updating the gain and phase imbalance factors, the preset IQ compensator gain and phase imbalance factors are obtained, and the DC offset correction is performed; the image correction model is established for the above Gain and phase unbalance factors perform image correction to suppress the image signal; activate the IQ compensator, enable the PA, start the PD training sequence, perform inverse balance processing on the output IQ-corrected in-phase signal and quadrature signal, and output the unbalanced in-phase and quadrature signals.

6. The method according to claim 5, wherein, when the image signal of the B frequency band in the A frequency band needs to be eliminated in the system, the image correction model formula is: αA _FW +βB _FW =A _FB , wherein, A _FW is The forward output signal of the A frequency band, B _FW is the forward output signal of the B frequency band, A _FB is the forward feedback signal of the A frequency band, α represents the gain and phase coefficient of the influence of the transmitted signal of the A frequency band on itself, and β represents the influence of the A frequency band on the Gain and phase coefficients for B-band effects.

7. The method according to claim 5, wherein the DC offset correction is specifically, calling the DC offset equation according to the DC signals _C1 and _C2

\begin{matrix} C_{c 1} h_{11} + C_{c 2} h_{12} = - C_{1} \\ C_{c 1} h_{twenty one} + C_{c 2} h_{twenty two} = - C_{2} \end{matrix},

[\begin{matrix} C_{c 1} \\ C_{c 2} \end{matrix}] = {[\begin{matrix} h_{11} & h_{12} \\ h_{twenty one} & h_{twenty two} \end{matrix}]}^{- 1} [\begin{matrix} {- C}_{1} \\ {- C}_{2} \end{matrix}],

8. The method according to claim 6, characterized in that, according to the formula:

α(A _FW -α ^-1 βB _FW )+βB _FW ＝αA _FW , the transmitted signal is compensated at the zero intermediate frequency, and the image effect is eliminated, wherein, the transfer parameter matrix of the A-band forward data to the A-band feedback data is:

A_{FW} = [\begin{matrix} A_{FW} I (0) & A_{FW} I (1) & A_{FW} I (2) & . . . & A_{FW} I (N - 2) & A_{FW} I (N - 1) \\ A_{FW} Q (0) & A_{FW} Q (1) & A_{FW} Q (2) & . . . & A_{FW} Q (N - 2) & A_{FW} Q (N - 1) \end{matrix}],

The first line is the in-phase component of the A-band forward data, which is used as the real part; the second line is the quadrature component of the A-band forward data, which is used as the imaginary part; the transfer parameter matrix of the B-band forward data to the A-band feedback data is:

B_{FW} = [\begin{matrix} B_{FW} I (0) & B_{FW} I (1) & B_{FW} I (2) & . . . & A_{FW} I (N - 2) & B_{FW} I (N - 1) \\ B_{FW} Q (0) & B_{FW} Q (1) & B_{FW} Q (2) & . . . & A_{FW} Q (N - 2) & B_{FW} Q (N - 1) \end{matrix}],

The first line is the in-phase component of the B-band forward data, as the real part; the second line is the quadrature component of the B-band forward data, as the imaginary part; the transfer parameter matrix of the A-band forward feedback data is:

A_{Facebook} = [\begin{matrix} A_{Facebook} I (0) & A_{Facebook} I (1) & A_{Facebook} I (2) & . . . & A_{Facebook} I (N - 2) & A_{Facebook} I (N - 1) \\ A_{Facebook} Q (0) & A_{Facebook} Q (1) & A_{Facebook} Q (2) & . . . & A_{Facebook} Q (N - 2) & A_{Facebook} Q (N - 1) \end{matrix}],

The first behavior is the in-phase component of the A-band forward feedback data, which is used as the real part, and the second behavior is the quadrature component of the A-band forward feedback data, which is used as the imaginary part.