CN112803972A - Digital predistortion method and system suitable for MIMO transmitter full loop distortion compensation - Google Patents

Abstract

The invention belongs to the technical field of wireless communication, and discloses a digital predistortion method and a digital predistortion system suitable for full loop distortion compensation of an MIMO transmitter, wherein the digital predistortion method comprises the following steps: switching on the short-circuit channel to make the N power amplifiers invalid, calculating the parameter value of the combined pre-eliminator according to the input baseband signal and the feedback signal, and compensating the error of the quadrature modulator, the error of the quadrature demodulator and the crosstalk effect; and (3) disconnecting the short-circuit channels to enable the N power amplifiers to work normally, calculating the parameter value of the predistorter according to the input baseband signal and the feedback signal, and performing distortion compensation on the power amplifiers. The method can eliminate the error of the orthogonal modulator, the error of the orthogonal demodulator and the crosstalk effect in a pre-combined manner, then carry out the distortion compensation of the power amplifier which is mutually independent, greatly improve the modeling precision and the compensation performance of a compensation model, effectively inhibit in-band distortion and out-of-band spectrum expansion, and meet the requirements of a broadband MIMO wireless communication system on the performance.

Description

Digital predistortion method and system suitable for MIMO transmitter full loop distortion compensation

Technical Field

The invention belongs to the technical field of wireless communication, and particularly relates to a digital predistortion method and a digital predistortion system suitable for full-loop distortion compensation of a MIMO transmitter.

Background

In practical application scenarios, due to factors such as imperfect devices, various errors inevitably exist in a loop of a multi-antenna MIMO transmitter system, and the improvement of system performance and efficiency is limited. The MIMO transmitter full loop distortion comprises quadrature modulator errors, quadrature demodulator errors, crosstalk effects among multiple branches and power amplifier distortion. The quadrature modulator error and the quadrature demodulator error mainly comprise that the amplitudes of an I path and a Q path are not strictly equal, the phases do not meet the strict quadrature relation, and direct current offset errors exist. The crosstalk effect among the branches mainly comprises a nonlinear crosstalk effect and a linear crosstalk effect. Nonlinear crosstalk occurs before the power amplifier and linear crosstalk occurs after the power amplifier. The power amplifier distortion mainly comprises signal distortion caused by nonlinear distortion and memory effect.

At present, the multi-antenna MIMO predistortion method has become the most promising technique in many error compensation schemes of the wireless communication system. In the MIMO predistortion method for jointly compensating the quadrature modulator error, the quadrature demodulator error, the crosstalk effect and the power amplifier distortion, in 2013, digital predistortion DPD method CP-DPD based on conjugate polynomial CP is proposed in "Compensation of I/Q amplification and nonlinear distortion in MIMO wireless transmitters" by D.Saffar et al, however, the CP-DPD method does not consider the influence of DC offset, so after adding DC offset Compensation DCC in the CP-DPD method, the CP-DPD-DCC method is obtained; in 2015, Milica Bozic et al proposed a Two-Box Model predistortion method in "Joint Compensation of I/Q Impatientins, Power Amplifier nonlinear and Cross in MIMO Transmitters using Two-Box Model"; in 2017, Zain Ahmed Khan et al put forward an ALCC-DPD method based on an augmented linear complex conjugate ALCC in "Digital prediction for Joint simulation of I/Q Implanace and MIMO Power Amplifier simulation"; in 2018, Praven Jaraut et al proposed a Composite Neural Network Predistortion method in Composite Neural Network Digital Predistortion Model for Joint simulation of Cross, I/Q Impalance, nonlinear in MIMO Transmitters.

The MIMO predistortion method mainly carries out combined compensation on quadrature modulator errors, quadrature demodulator errors, nonlinear and linear crosstalk effects among multiple branches and power amplifier distortion. However, in the compensation model of the MIMO predistortion method described above, a conjugate signal term and other branch signal terms are inevitably included to compensate for the quadrature modulator error and crosstalk; therefore, in an actual MIMO transmitter system, as the number of MIMO branches, the memory depth of the power amplifier, and the nonlinear order increase, the number of parameters of the compensation model in the MIMO predistortion method increases exponentially, and the defects of the MIMO predistortion method that the number of model parameters is large, the computational complexity is high, and the data storage amount of the operation link is large, which are not favorable for practical application, are further revealed. Meanwhile, with the enhancement of errors, the performance of the method in the aspects of inhibiting in-band distortion and out-of-band spectrum expansion is poor, and the practical application and popularization of the method are further limited.

Through the above analysis, the problems and defects of the prior art are as follows:

(1) in an actual MIMO transmitter system, with the increase of MIMO branch number, memory depth of a power amplifier and nonlinear order, the number of parameters of a compensation model in the MIMO predistortion method increases exponentially, which reveals the defects of many model parameters, high calculation complexity and the like in the MIMO predistortion method, and is difficult to apply in an actual system.

(2) With the enhancement of errors, the performance of the existing MIMO predistortion method in the aspects of inhibiting in-band distortion and out-of-band spectrum expansion is poor, and the performance requirements of the existing standard cannot be met.

The difficulty in solving the above problems and defects is: how to reduce the redundant items in the compensation model, thereby simplifying the compensation model, reducing the computational complexity and simultaneously keeping the system at higher performance.

The significance of solving the problems and the defects is as follows: on the premise of not losing performance, the calculation complexity of the MIMO predistortion method is reduced, namely the compensation efficiency of the MIMO predistortion method is improved, and the method has important significance for improving the working efficiency of actual devices and reducing the power consumption of a wireless communication system.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a digital predistortion method and a digital predistortion system which are suitable for the full-loop distortion compensation of an MIMO transmitter.

The invention is realized in this way, a digital predistortion method and system suitable for the compensation of the full loop distortion of the MIMO transmitter, the digital predistortion method suitable for the compensation of the full loop distortion of the MIMO transmitter includes the following steps:

step one, turning on a short-circuit channel to make N power amplifiers invalid, and inputting a baseband signal s at the nth moment_i(n) passing s through predistorter in through state and joint canceller in through state in sequence_i(n) as received output u_i(n); wherein, i ═ 1.·, N, denotes the ith MIMO branch; the through state represents that the output signal of the module is equal to the input signal;

step two, combining the output signal u of the pre-canceller_i(n) outputting x after being interfered by the quadrature modulator error module_i(n), quadrature modulator error interference signal x_i(n) output z after being interfered by nonlinear crosstalk module_i(n)；

Step three, nonlinear crosstalk interference signal z_i(n) passing through the short-circuit path and outputting as it is as y_i(n)，y_i(n) entering a feedback loop and obtaining a signal f after being interfered by the linear crosstalk module_i(n), interference signal f of linear crosstalk_i(n) obtaining a signal g after being disturbed by quadrature demodulator errors_i(n)；

Step four, interfering the error of the quadrature demodulator with a signal g_i(n) and an input baseband signal s_i(n) sending the parameters into a parameter calculation module of the joint pre-eliminator, copying the calculated parameter values into the joint pre-eliminator, and enabling the joint pre-eliminator to enter a normal operation state;

step five, disconnecting the short-circuit channel to enable the N power amplifiers to work normally, and inputting a baseband signal s_i(n) passing s through a predistorter in the through state_i(n) outputting as it isIs w_i(n)；

Step six, an output signal w of the predistorter_i(n) pre-compensating for quadrature modulator errors, non-linear crosstalk, linear crosstalk and quadrature demodulator errors by a joint pre-canceller in a normal operating state to obtain a pre-processed signal u_i(n)；

Step seven, repeating step two, the output signal z of the nonlinear crosstalk module_i(n) output y affected by non-linear distortion and memory effects in the power amplifier_i(n)；

Step eight, the power amplifier outputs a signal y_i(n) entering a feedback loop, and obtaining a signal g after being sequentially interfered by linear crosstalk and quadrature demodulator errors_i(n) converting the signal g_i(n) and an input baseband signal s_i(N) sending the parameters into a predistorter parameter calculation module, and updating the calculated parameter values into N predistorters to enable the predistorters to enter a normal operation state;

step nine, inputting a baseband signal s_i(n) obtaining a signal w after inverse processing of the nonlinear distortion and the memory effect of the power amplifier through a predistorter in a normal operation state_iAnd (n) repeating the sixth step and the seventh step.

Further, the error interference process of the quadrature modulator in the second step and the error interference process of the quadrature demodulator in the third step include:

calculating by utilizing unbalanced coefficients alpha and beta and a direct current offset gamma, wherein the unbalanced coefficients alpha and beta are mainly determined by amplitude unbalanced xi and phase unbalanced theta, and the relation is as follows: α ═ 1+ (1+ ε) e^jθ]/2、β＝[1-(1+ε)e^-jθ]/2；

In step two, an input baseband signal u is input_i(n) after the signal is sent to an error module of the quadrature modulator, an error interference signal x of the quadrature modulator is obtained_i(n) according to the following formula:

x_i(n)＝αRe[u_i(n)]+βIm[u_i(n)]j+γ；

wherein Re [. cndot ] is an operation of obtaining a real part of the complex signal, Im [. cndot ] is an operation of obtaining an imaginary part of the complex signal, and j represents an imaginary unit;

in step three, the interference signal f of linear crosstalk_i(n) obtaining a signal g after being disturbed by quadrature demodulator errors_i(n) according to the following formula:

g_i(n)＝αRe[f_i(n)]+βIm[f_i(n)]j+γ。

further, the interference process of the nonlinear crosstalk module in the second step and the interference process of the linear crosstalk module in the third step include:

using the crosstalk coefficient alpha_i，jPerforming a calculation of where_i，jRepresenting the crosstalk value of the ith branch to the jth branch; in step two, the quadrature modulator error interference signal x_i(n) obtaining a nonlinear crosstalk interference signal z after the nonlinear crosstalk interference module is interfered by a nonlinear crosstalk effect_i(n) according to the following formula:

z_i(n)＝α_1，1x₁(n)+α_2，1x₂(n)+...+α_N，1x_N(n)；

in step three, the signal y_i(n) entering a feedback loop and obtaining a signal f after being interfered by the linear crosstalk module_i(n) according to the following formula:

f_i(n)＝α_1，1y₁(n)+α_2，1y₂(n)+...+α_N，1y_N(n)。

further, in step four, the calculation process of the joint pre-canceller parameter calculation module includes:

one-frame input baseband signal s with continuous acquisition frame length L_i(n) forming an input baseband signal matrix S_i＝[s_i(1) s_i(2)...s_i(L)]^T(ii) a Collecting the same frame s_i(n) corresponding feedback signal g_i(n) forming a feedback signal matrix G_i＝[g_i(1) g_i(2)...g_i(L)]^T(ii) a Wherein, the g_i(n) is s_i(n) are sequentially interfered by quadrature modulator error, nonlinear crosstalk effect, linear crosstalk effect and quadrature demodulator interferenceThe latter signal;

by using S_iMatrix of real part S_irAnd an imaginary matrix S_imOf a feedback signal matrix G_iAnd an all-one matrix I ═ 1.. 1 of size 1xL]The coefficient matrix M ═ M of the joint pre-canceller is calculated according to the following formula_i，j}^TWherein i is more than or equal to 1 and less than or equal to N, j is more than or equal to 1 and less than or equal to 3N:

wherein inv (·) represents the inversion of the square matrix, x represents the conjugation of the matrix, T represents the transposition of the matrix, and j represents the unit of imaginary number.

Further, in step six, the pre-canceling the quadrature modulator error, the nonlinear crosstalk effect, the linear crosstalk effect, and the quadrature demodulator error by using the joint pre-canceller includes:

coefficient matrix M ═ M using joint pre-canceller_i，j}^TFor input baseband signal w_i(n) after error inverse processing, obtaining a pre-processing signal u_i(n) according to the following formula:

u_i(n)＝m_i，1 Re[w₁(n)]+m_i，2Im[w₁(n)]j+m_i，3+

m_i，4Re[w₂(n)]+m_i，5Im[w₂(n)]j+m_i，6+

…

m_i，3N-2Re[w_N(n)]+m_i，3N-1 Im[w_N(n)]j+m_i，3N，，

where Re [. cndot. ] is the operation of calculating the real part of the complex signal, Im [. cndot. ] is the operation of calculating the imaginary part of the complex signal, and j represents the imaginary unit.

Further, in step seven, the crosstalk interference signal z_i(n) are affected by power amplifier nonlinear distortion and memory effects, including:

modeling a power amplifier using a memory polynomial having a coefficient h_i，k，qK and Q are respectively the nonlinear order and the memory depth of the power amplifier memory polynomial model, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, K and Q are respectively the highest nonlinear order and the maximum memory depth of the power amplifier memory polynomial model, odd represents an odd set, and a signal z is a signal_i(n) output y affected by non-linear distortion and memory effect of power amplifier_i(n) according to the following formula:

further, in step eight, the step of obtaining the baseband signal s is performed according to the input baseband signal s_i(n) and a feedback signal g_i(n) calculating the independent predistorter parameter values in each branch, including:

one-frame input baseband signal s with continuous acquisition frame length L_i(n) forming an input baseband signal matrix S_i＝[s_i(1) s_i(2)...s_i(L)]^TCollecting the same frame s_i(n) corresponding feedback signal g_i(n); storing a memory polynomial matrix according to the following format

In (1),

is a matrix of LxK (Q +1), wherein the sub-matrices

Is a matrix of LxK with a matrix element of a_k(g_i(n))＝g_i(n)|g_i(n)|^k-1The structure is as follows:

according to S_iAnd

to calculate the independent predistorter coefficient matrix D in each branch_i＝{d_i，k，qK and Q are respectively the nonlinear order and the memory depth of the memory polynomial model of the predistorter, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, and K and Q are respectively the highest nonlinear order and the maximum memory depth of the memory polynomial model of the predistorter; odd represents an odd set, and the parameter value calculation process of the predistorter is carried out according to the following formula:

wherein inv (·) represents the inversion of the square matrix, and H represents the conjugate transpose of the matrix.

Further, in the ninth step, the input baseband signal s is predistorted by the predistorter_i(n) inverse processing of the non-linear characteristic and the memory effect is carried out to obtain the predistortion signal w_i(n) comprising:

modeling a predistorter using a memory polynomial with coefficient d_i，k，qK and Q are respectively the nonlinear order and the memory depth of the memory polynomial model of the predistorter, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, and K and Q are respectively the highest nonlinear order and the maximum memory depth of the memory polynomial model of the predistorter; odd denotes an odd set, input baseband signal s_i(n) obtaining a signal w after inverse processing of nonlinear characteristics and memory effect through a predistorter in a normal operation state_i(n) according to the following formula:

another objective of the present invention is to provide a full-loop distortion compensation digital predistortion method and system for MIMO transmitter, which specifically includes:

n independent predistorters, 1 joint pre-canceller, N independent predistorter parameter calculation modules and 1 joint pre-canceller parameter calculation module; the quadrature modulator error, the nonlinear crosstalk effect, the linear crosstalk effect and the quadrature demodulator error are respectively equivalent to a quadrature modulator error module, a nonlinear crosstalk module, a linear crosstalk module and a quadrature demodulator error module.

By combining all the technical schemes, the invention has the advantages and positive effects that: on one hand, the digital predistortion method and the digital predistortion system suitable for the MIMO transmitter full-loop distortion compensation can eliminate errors of an orthogonal modulator and an orthogonal demodulator and nonlinear and linear crosstalk effects in a combined manner in advance, so that mutually independent predistorter parameter value calculation and compensation can be carried out on distortion of a power amplifier in each MIMO branch, the modeling precision and the compensation performance of a compensation model are greatly improved, in-band distortion and out-band spectrum expansion are effectively inhibited, and the requirements of a broadband MIMO wireless communication system on the performance are met; on the other hand, the novel predistortion structure is adopted, so that the calculation complexity of the MIMO predistortion method is reduced to the maximum extent on the premise of ensuring the system performance, and the MIMO predistortion method is more beneficial to practical application.

The invention can be applied to the actual MIMO transmitter to jointly eliminate the comprehensive influence caused by the orthogonal modulator error, the orthogonal demodulator error, the nonlinear crosstalk, the linear crosstalk effect, the nonlinear distortion of the power amplifier and the memory effect due to the nonideal characteristic of the analog device of the MIMO transmitter. The MIMO predistortion method of the invention can effectively compensate the errors of the above five systems, so that the compensation system not only has higher performance, but also keeps lower complexity, thereby being beneficial to practical application. The invention can effectively improve the comprehensive performance of the MIMO transmitter system and simultaneously reduce the realization complexity of the MIMO predistortion method.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a flowchart of a full-loop distortion compensation digital predistortion method suitable for a MIMO transmitter according to an embodiment of the present invention.

Fig. 2 is a schematic structural diagram of a MIMO predistortion method in a2x 2 MIMO transmitter according to an embodiment of the present invention.

Fig. 3 is a schematic structural diagram of a joint pre-canceller in a2x 2 MIMO transmitter according to a MIMO pre-distortion method provided in an embodiment of the present invention.

Fig. 4 is a schematic structural diagram of a predistorter of a MIMO predistortion method in a2x 2 MIMO transmitter according to an embodiment of the present invention.

Fig. 5 is a schematic diagram comparing power spectral density performance of the MIMO predistortion method provided by the embodiment of the present invention with that of the existing three methods.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

In view of the problems in the prior art, the present invention provides a digital predistortion method and system suitable for the full loop distortion compensation of a MIMO transmitter, and the present invention is described in detail below with reference to the accompanying drawings.

As shown in fig. 1, the digital predistortion method for full-loop distortion compensation of a MIMO transmitter according to the embodiment of the present invention includes the following steps:

s101, turning on a short-circuit channel to make N power amplifiers invalid, and inputting a baseband signal S at the nth moment_i(n) passing s through predistorter in through state and joint canceller in through state in sequence_i(n) as received output u_i(n); wherein, i ═ 1.·, N, denotes the ith MIMO branch; the through state represents that the output signal of the module is equal to the input signal;

s102, combining the output signal u of the pre-canceller_i(n) outputting x after being interfered by the quadrature modulator error module_i(n), quadrature modulatorError interference signal x_i(n) output z after being interfered by nonlinear crosstalk module_i(n)；

S103, nonlinear crosstalk interference signal z_i(n) passing through the short-circuit path and outputting as it is as y_i(n)，y_i(n) entering a feedback loop and obtaining a signal f after being interfered by the linear crosstalk module_i(n), interference signal f of linear crosstalk_i(n) obtaining a signal g after being disturbed by quadrature demodulator errors_i(n)；

S104, the quadrature demodulator error interference signal gi (n) and the input baseband signal S_i(n) sending the parameters into a parameter calculation module of the joint pre-eliminator, copying the calculated parameter values into the joint pre-eliminator, and enabling the joint pre-eliminator to enter a normal operation state;

s105, disconnecting the short-circuit channel to enable the N power amplifiers to work normally, and inputting a baseband signal S_i(n) passing s through a predistorter in the through state_i(n) as-received output of w_i(n)；

S106, output signal w of predistorter_i(n) pre-compensating for quadrature modulator errors, non-linear crosstalk, linear crosstalk and quadrature demodulator errors by a joint pre-canceller in a normal operating state to obtain a pre-processed signal u_i(n)；

S107, repeating the step S102, and outputting a signal z of the nonlinear crosstalk module_i(n) output y affected by non-linear distortion and memory effects in the power amplifier_i(n)；

S108, power amplifier output signal y_i(n) entering a feedback loop, and obtaining a signal g after being sequentially interfered by linear crosstalk and quadrature demodulator errors_i(n) converting the signal g_i(n) and an input baseband signal s_i(N) sending the parameters into a predistorter parameter calculation module, and updating the calculated parameter values into N predistorters to enable the predistorters to enter a normal operation state;

s109, inputting a baseband signal S_i(n) obtaining a signal w after inverse processing of the nonlinear distortion and the memory effect of the power amplifier through a predistorter in a normal operation state_i(n) repeating the steps S106,S107；

The technical solution of the present invention is further described with reference to the following examples.

The digital predistortion method and system suitable for full loop distortion compensation of a MIMO transmitter provided by the embodiment of the invention are carried out in an N x N MIMO transmitter system, and the MIMO predistortion method and system comprise the following steps: n independent predistorters, 1 joint pre-canceller, N independent predistorter parameter calculation modules and 1 joint pre-canceller parameter calculation module; the quadrature modulator error, the nonlinear crosstalk effect, the linear crosstalk interference effect and the quadrature demodulator error are respectively equivalent to a quadrature modulator error module, a nonlinear crosstalk module, a linear crosstalk module and a quadrature demodulator error module.

The full-loop distortion compensation digital predistortion method suitable for the MIMO transmitter provided by the embodiment of the invention comprises the following steps:

step one, switching on a short-circuit channel to enable N power amplifiers to be invalid, and inputting a baseband signal s at the nth moment_i(N) (i ═ 1.. times.n, denoting the i-th MIMO branch) the predistorted signal w is obtained by means of a predistorter in the pass-through state (the output signal of the module is equal to the input signal)_i(n), predistortion signal w_i(n) obtaining the pre-processed signal u by a joint pre-canceller in a through state_i(n) the following relationship holds:

u_i(n)＝w_i(n)＝s_i(n)；

step two, inputting a baseband signal u_i(n) after the signal is sent to an error module of the quadrature modulator, an error interference signal x of the quadrature modulator is obtained_i(n), specifically including: calculating by utilizing unbalanced coefficients alpha and beta and a direct current offset gamma, wherein the unbalanced coefficients alpha and beta are mainly determined by amplitude unbalanced xi and phase unbalanced theta, and the specific relationship is as follows: α ═ 1+ (1+ ε) e^jθ]/2、β＝[1-(1+ε)e^-jθ]The quadrature modulator error interference process is carried out according to the following formula:

x_i(n)＝αRe[u_i(n)]+βIm[u_i(n)]j+γ；

wherein Re [. cndot ] is an operation of obtaining a real part of the complex signal, Im [. cndot ] is an operation of obtaining an imaginary part of the complex signal, and j represents an imaginary unit;

step three, the error interference signal x of the quadrature modulator_i(n) obtaining a nonlinear crosstalk interference signal z after the interference of the nonlinear crosstalk effect_i(n) using the crosstalk coefficient α_i，j(representing the crosstalk value of the ith branch to the jth branch) is calculated, and the nonlinear crosstalk effect interference process is carried out according to the following formula:

z_i(n)＝α_1，1x₁(n)+α_2，1x₂(n)+...+α_N，1x_N(n)；

step four, the nonlinear crosstalk interference signal passes through the short-circuit channel and then is output as is y_i(n), signal y_i(n) entering a feedback loop and obtaining a signal f after being interfered by the linear crosstalk module_i(n) according to the following formula:

f_i(n)＝α_1，1y₁(n)+α_2，1y₂(n)+...+α_N，1y_N(n)；

step five, linear crosstalk interference signal f_i(n) obtaining a signal g after being disturbed by quadrature demodulator errors_i(n) according to the following formula:

g_i(n)＝αRe[f_i(n)]+βIm[f_i(n)]j+γ；

step six, interfering the error of the orthogonal demodulator with a signal g_i(n) and an input baseband signal s_i(n) feeding the parameters into a parameter calculation module of the joint pre-eliminator, copying the calculated parameter values into the joint pre-eliminator to make the joint pre-eliminator enter a normal operation state, wherein the parameter calculation process of the joint pre-eliminator specifically comprises:

one-frame input baseband signal s with continuous acquisition frame length L_i(n) forming an input baseband signal matrix S_i＝[s_i(1) s_i(2)...s_i(L)]^T(ii) a Collecting the same frame s_i(n) corresponding feedback signal g_i(n) forming a feedback signal matrix G_i＝[g_i(1) g_i(2)...g_i(L)]^T(ii) a Wherein, the g_i(n) is s_i(n) signals interfered by the quadrature modulator error, the nonlinear crosstalk effect, the linear crosstalk effect and the quadrature demodulator error in sequence;

by using S_iMatrix of real part S_irAnd an imaginary matrix S_imOf a feedback signal matrix G_iAnd an all-one matrix I ═ 1.. 1 of size 1xL]The coefficient matrix M ═ M of the joint pre-canceller is calculated according to the following formula_i，j}^TWherein i is more than or equal to 1 and less than or equal to N, j is more than or equal to 1 and less than or equal to 3N:

wherein inv (·) represents square matrix inversion, matrix conjugation, T represents matrix transposition, and j represents an imaginary unit;

step seven, disconnecting the short-circuit channel to enable the N power amplifiers to work normally, and inputting a baseband signal s_i(n) outputting w after passing through predistorter in through state_i(n) the following relation holds:

w_i(n)＝s_i(n)；

step eight, an output signal w of the predistorter_i(n) pre-eliminating the influence of the quadrature modulator error, the non-linear crosstalk effect, the linear crosstalk effect and the quadrature demodulator error in advance by the joint pre-eliminator in the normal operation state to obtain a pre-processing signal u_i(n), specifically including:

coefficient matrix M ═ M using joint pre-canceller_i，j}^TFor input baseband signal w_i(n) inverse processing of quadrature modulator error, quadrature demodulator error, nonlinear crosstalk and linear crosstalk effect is performed to obtain a preprocessed signal u_i(n) according to the following formula:

u_i(n)＝m_i，1 Re[w₁(n)]+m_i，2Im[w₁(n)]j+m_i，3+

m_i，4Re[w₂(n)]+m_i，5Im[w₂(n)]j+m_i，6+

…

m_i，3N-2Re[w_N(n)]+m_i，3N-1 Im[w_N(n)]j+m_i，3N，；

wherein Re [. cndot ] is an operation of obtaining a real part of the complex signal, Im [. cndot ] is an operation of obtaining an imaginary part of the complex signal, and j represents an imaginary unit;

step nine, repeating step two and step three, output signal z of nonlinear crosstalk module_i(n) output y affected by non-linear distortion and memory effects in the power amplifier_i(n), as follows:

in the invention, a power amplifier is modeled by using a memory polynomial with the coefficient of h_i，k，qK and Q are respectively the nonlinear order and the memory depth of the power amplifier memory polynomial model, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, K and Q are respectively the highest nonlinear order and the maximum memory depth of the power amplifier memory polynomial model, odd represents an odd set, and a signal z is a signal_i(n) output y affected by non-linear distortion and memory effects in the power amplifier_i(n) according to the following formula:

step ten, the output signal y of the power amplifier_i(n) sending the signal into a feedback loop, and obtaining a signal g after being sequentially interfered by linear crosstalk and quadrature demodulator errors_i(n) converting the signal g_i(n) and an input baseband signal s_i(N) sending the parameters into a predistorter parameter calculation module, and updating the calculated parameter values into N predistorters to enable the predistorters to enter a normal operation state, wherein the predistorter parameter calculation process comprises the following steps:

one-frame input baseband signal s with continuous acquisition frame length L_i(n) forming an input baseband signal matrix S_i＝[s_i(1) s_i(2)...s_i(L)]^TCollecting the same frame s_i(n) corresponding feedback signal g_i(n); storing a memory polynomial matrix according to the following format

In (1),

is a matrix of LxK (Q +1), wherein the sub-matrices

Is a matrix of LxK with a matrix element of a_k(g_i(n))＝g_i(n)|g_i(n)|^k-1The structure is as follows:

according to S_iAnd

to calculate the independent predistorter coefficient matrix D in each branch_i＝{d_i，k，qK and Q are respectively the nonlinear order and the memory depth of the memory polynomial model of the predistorter, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, and K and Q are respectively the highest nonlinear order and the maximum memory depth of the memory polynomial model of the predistorter; odd represents an odd set, and the parameter value calculation process of the predistorter is carried out according to the following formula:

wherein inv (·) represents the inversion of a square matrix, and H represents the conjugate transpose of the matrix;

step eleven, inputting a baseband signal s_i(n) obtaining a signal w after inverse processing of nonlinear characteristics and memory effect through a predistorter in a normal operation state_i(n), as follows:

in the present invention, a memory polynomial is used to model a predistorter, which is storedMemory polynomial coefficient of d_i，k，qK and Q are respectively the nonlinear order and the memory depth of the memory polynomial model of the predistorter, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, and K and Q are respectively the highest nonlinear order and the maximum memory depth of the memory polynomial model of the predistorter; odd represents the odd set, and the predistorter compensation process proceeds as follows:

and repeating the eight steps and the nine steps.

For a2x 2 MIMO transmitter, the system block diagram, the joint pre-canceller and the pre-distorter of the MIMO predistortion method according to the present invention are shown in fig. 2, fig. 3 and fig. 4, respectively.

The above steps describe the preferred embodiment of the present invention, and it is obvious that those skilled in the art can make various modifications and substitutions to the present invention with reference to the preferred embodiment of the present invention and the accompanying drawings, and those modifications and substitutions should fall within the protection scope of the present invention.

The effect of the present invention will be further explained with the simulation experiment.

1) Simulation conditions are as follows: matlab simulation software is used in the simulation experiment, and the MIMO branch number N is 2; input baseband signal s_i(n) (i ═ 1, 2) is an OFDM signal, the number of subcarriers (frame length L) is 2048, the modulation scheme is 64QAM, the cyclic prefix is 128, and the upsampling is 8 times; the amplitude unbalance s of the error models of the quadrature modulator and the quadrature demodulator is 3%, the phase unbalance theta is 3 degrees, and the direct current offset gamma is 0.03+0.01 j; magnitude of crosstalk α_1，2＝α_2，1＝-20dB、α_1，1＝α_2，21 is ═ 1; the predistorter and the power amplifier both adopt a memory polynomial model, the highest nonlinear order K is 5, and the maximum memory depth Q is 3.

2) Simulation content and results:

in a2x 2 MIMO transmitter, the MIMO predistortion method and the existing three MIMO predistortion methods are respectively adopted to compensate signals subjected to combined interference of five errors, namely, an orthogonal modulator error, an orthogonal demodulator error, nonlinear crosstalk, linear crosstalk and power amplifier distortion, the power spectral density curve of related signals is shown in fig. 5, the adjacent channel power ratio and the error vector magnitude are shown in table 1, and the number of compensation model parameters is shown in table 2.

In the power spectral density curves plotted in fig. 5, curves a, b, c, e and f all represent power spectral density curves of output signals of the quadrature demodulator, where curve a represents compensation processing performed by using the ALCC-DPD method, curve b represents compensation processing performed by using the MIMO predistortion method according to the present invention, curve c represents compensation processing performed by using the CP-DPD-DCC method, curve e represents compensation processing performed by using the CP-DPD method, and f represents no compensation processing; curve d represents the power spectral density curve of the input baseband signal.

As can be seen from fig. 5 and table 1, the MIMO predistortion method of the present invention can efficiently compensate for quadrature modulator error, quadrature demodulator error, nonlinear crosstalk, linear crosstalk, and power amplifier distortion, and effectively suppress in-band distortion and out-of-band spectrum regeneration of signals. Compared with the three existing MIMO predistortion methods, the MIMO predistortion method has great improvement on three indexes of power spectral density, adjacent channel power ratio and error vector amplitude value.

As can be seen from table 2, compared to the three existing MIMO predistortion methods, the MIMO predistortion method of the present invention has advantages in the number of model parameters and has lower complexity, for example, the number of compensation model parameters in the MIMO predistortion method of the present invention is about one ninth of the ALCC-DPD method and one half of the CP-DPD-DCC method.

Table 1 is a performance comparison table of adjacent channel power ratio and error vector magnitude values of the present invention and the existing three methods, and table 2 is a comparison table of the number of compensation model parameters in a2x 2 MIMO system of the present invention and the existing three methods.

TABLE 1 adjacent channel power ratio and error vector magnitude performance comparison table for the present invention and the existing three methods

Table 2 comparison table of number of compensation model parameters in2x 2 MIMO system according to the present invention and the existing three methods

In conclusion, compared with the existing three MIMO predistortion methods, the MIMO predistortion method has better performance and lower complexity, and is more beneficial to practical application.

The embodiment of the present application further provides another digital predistortion processing method and system, which are applicable to a MIMO transmitter system having quadrature modulator errors, nonlinear crosstalk, and power amplifier errors at the same time, and have an ideal feedback loop, that is, there are no linear crosstalk and quadrature demodulator errors, and the specific implementation steps are as follows:

step one, turning on a short-circuit channel to make N power amplifiers invalid, and inputting a baseband signal s at the nth moment_i(n) passing s through predistorter in through state and joint canceller in through state in sequence_i(n) as received output u_i(n); wherein, i ═ 1.·, N, denotes the ith MIMO branch; the through state represents that the output signal of the module is equal to the input signal;

step two, combining the output signal u of the pre-canceller_i(n) outputting x after being interfered by the quadrature modulator error module_i(n), quadrature modulator error interference signal x_i(n) output z after being interfered by nonlinear crosstalk module_i(n)；

Step three, nonlinear crosstalk interference signal z_i(n) outputting y as it is after passing through the short-circuited path_i(n) converting the signal y_i(n) and an input baseband signal s_i(n) sending the parameters into a parameter calculation module of the joint pre-eliminator, copying the calculated parameter values into the joint pre-eliminator, and enabling the joint pre-eliminator to enter a normal operation state;

step four, the short-circuit channel is disconnected to ensure thatN power amplifiers work normally, and a baseband signal s is input_i(n) passing s through a predistorter in the through state_i(n) as-received output of w_i(n)；

Step five, an output signal w of the predistorter_i(n) precompensating quadrature modulator errors and nonlinear crosstalk with a joint pre-canceller in normal operation to obtain a pre-processed signal u_i(n)；

Step six, repeating step two, output signal z of nonlinear crosstalk module_i(n) output y affected by non-linear distortion and memory effects in the power amplifier_i(n)；

Step seven, the power amplifier outputs a signal y_i(n) entering a feedback loop to provide a feedback signal y_i(n) and an input baseband signal s_i(N) sending the parameters into a predistorter parameter calculation module, and updating the calculated parameter values into N predistorters to enable the predistorters to enter a normal operation state;

step eight, inputting a baseband signal s_i(n) obtaining a signal w after inverse processing of the nonlinear distortion and the memory effect of the power amplifier through a predistorter in a normal operation state_i(n), repeating the sixth step and the seventh step;

in the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When used in whole or in part, can be implemented in a computer program product that includes one or more computer instructions. When loaded or executed on a computer, cause the flow or functions according to embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.)). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, which is intended to cover all modifications, equivalents and improvements that are within the spirit and scope of the invention as defined by the appended claims.

Claims (10)

1. A digital predistortion method for full loop distortion compensation of a MIMO transmitter, the digital predistortion method for full loop distortion compensation of a MIMO transmitter comprising the steps of:

first, turning on the short-circuit channel to make N power amplifiers ineffective, and inputting baseband signal s at nth time_i(n) passing s through predistorter in through state and joint canceller in through state in sequence_i(n) as received output u_i(n); wherein, i ═ 1.·, N, denotes the ith MIMO branch; the through state represents that the output signal of the module is equal to the input signal;

second step of combining the output signal u of the pre-canceller_i(n) outputting x after being interfered by the quadrature modulator error module_i(n), quadrature modulator error interference signal x_i(n) output z after being interfered by nonlinear crosstalk module_i(n)；

Third, the non-linear crosstalk interference signal z_i(n) passing through the short-circuit path and outputting as it is as y_i(n)，y_i(n) entering a feedback loop and obtaining a signal f after being interfered by the linear crosstalk module_i(n), interference signal f of linear crosstalk_i(n) obtaining a signal g after being disturbed by quadrature demodulator errors_i(n)；

Fourthly, interfering the error of the quadrature demodulator with the signal g_i(n) and an input baseband signal s_i(n) sending the parameters into a parameter calculation module of the joint pre-eliminator, copying the calculated parameter values into the joint pre-eliminator, and enabling the joint pre-eliminator to enter a normal operation state;

fifthly, disconnecting the short-circuit channel to make the N power amplifiers work normally, and inputting the baseband signal s_i(n) passing s through a predistorter in the through state_i(n) as-received output of w_i(n)；

Sixthly, the output signal w of the predistorter_i(n) pre-compensating for quadrature modulator errors, non-linear crosstalk, linear crosstalk and quadrature demodulator errors by a joint pre-canceller in a normal operating state to obtain a pre-processed signal u_i(n)；

Seventh step, repeat second step, output signal z of nonlinear crosstalk module_i(n) output y affected by non-linear distortion and memory effects in the power amplifier_i(n)；

Eighth step, the power amplifier outputs signal y_i(n) entering a feedback loop, and obtaining a signal g after being sequentially interfered by linear crosstalk and quadrature demodulator errors_i(n) converting the signal g_i(n) and an input baseband signal s_i(N) sending the parameters into a predistorter parameter calculation module, and updating the calculated parameter values into N predistorters to enable the predistorters to enter a normal operation state;

ninth, inputting the baseband signal s_i(n) obtaining a signal w after inverse processing of the nonlinear distortion and the memory effect of the power amplifier through a predistorter in a normal operation state_iAnd (n), repeating the sixth step and the seventh step.

2. The digital predistortion method for full loop distortion compensation of a MIMO transmitter as set forth in claim 1, wherein the error interference procedure of said quadrature modulator in the second step and the error interference procedure of said quadrature demodulator in the third step comprise:

calculating by using unbalanced coefficients alpha and beta and a direct current offset gamma, wherein the unbalanced coefficients alpha and beta are mainly determined by amplitude unbalanced epsilon and phase unbalanced theta, and the relation is as follows: α ═ 1+ (1+ ε) e^jθ]/2、β＝[1-(1+ε)e^-jθ]/2；

In a second step, an input baseband signal u is applied_i(n) after the signal is sent to an error module of the quadrature modulator, an error interference signal x of the quadrature modulator is obtained_i(n) according to the following formula:

x_i(n)＝αRe[u_i(n)]+βIm[u_i(n)]j+γ；

wherein Re [. cndot ] is an operation of obtaining a real part of the complex signal, Im [. cndot ] is an operation of obtaining an imaginary part of the complex signal, and j represents an imaginary unit;

in a third step, the linear crosstalk interference signal f_i(n) obtaining a signal g after being disturbed by quadrature demodulator errors_i(n) according to the following formula:

g_i(n)＝αRe[f_i(n)]+βIm[f_i(n)]j+γ。

3. the digital predistortion method for full loop distortion compensation of a MIMO transmitter of claim 1, wherein the interference procedure of the nonlinear crosstalk module in the second step and the interference procedure of the linear crosstalk module in the third step comprise:

using the crosstalk coefficient alpha_i,jPerforming a calculation of where_i,jRepresenting the crosstalk value of the ith branch to the jth branch; in a second step, the quadrature modulator error interference signal x_i(n) obtaining a signal z after being interfered by the nonlinear crosstalk module_i(n) according to the following formula:

z_i(n)＝α_1,1x₁(n)+α_2,1x₂(n)+...+α_N,1x_N(n)；

in a third step, the signal y_i(n) entering a feedback loop and obtaining a signal f after being interfered by the linear crosstalk module_i(n) according to the following formula:

f_i(n)＝α_1,1y₁(n)+α_2,1y₂(n)+...+α_N,1y_N(n)。

4. the digital predistortion method for full loop distortion compensation of MIMO transmitter as claimed in claim 1, wherein in the fourth step, the calculation process of said joint canceller parameter calculation module comprises:

one-frame input baseband signal s with continuous acquisition frame length L_i(n) forming an input baseband signal matrix S_i＝[s_i(1) s_i(2) ... s_i(L)]^T(ii) a Collecting the same frame s_i(n) corresponding feedback signal g_i(n) forming a feedback signal matrix G_i＝[g_i(1) g_i(2) ... g_i(L)]^T(ii) a Wherein, the g_i(n) is s_i(n) signals subjected to quadrature modulator error, nonlinear crosstalk effect, linear crosstalk effect interference and quadrature demodulator interference in sequence;

by using S_iMatrix of real part S_irAnd an imaginary matrix S_imOf a feedback signal matrix G_iAnd an all-one matrix I ═ 1.. 1 of size 1xL]The coefficient matrix M ═ M of the joint pre-canceller is calculated according to the following formula_i,j}^TWherein i is more than or equal to 1 and less than or equal to N, j is more than or equal to 1 and less than or equal to 3N:

wherein inv (·) represents the inversion of the square matrix, x represents the conjugation of the matrix, T represents the transposition of the matrix, and j represents the unit of imaginary number.

5. The digital predistortion method for full loop distortion compensation of a MIMO transmitter as claimed in claim 1, wherein in the sixth step, said pre-canceling quadrature modulator error, nonlinear crosstalk effect, linear crosstalk effect and quadrature demodulator error using a joint pre-canceller comprises:

using joint predictionCoefficient matrix M ═ M for canceller_i,j}^TFor input baseband signal w_i(n) inverse processing of quadrature modulator error, nonlinear crosstalk effect, linear crosstalk effect and quadrature demodulator error is performed to obtain a preprocessed signal u_i(n) according to the following formula:

u_i(n)＝m_i,1Re[w₁(n)]+m_i,2Im[w₁(n)]j+m_i,3+

m_i,4Re[w₂(n)]+m_i,5Im[w₂(n)]j+m_i,6+

...

m_i,3N-2Re[w_N(n)]+m_i,3N-1Im[w_N(n)]j+m_i,3N,；

where Re [. cndot. ] is the operation of calculating the real part of the complex signal, Im [. cndot. ] is the operation of calculating the imaginary part of the complex signal, and j represents the imaginary unit.

6. The digital predistortion method for full loop distortion compensation of MIMO transmitter as claimed in claim 1, wherein in the seventh step, the nonlinear crosstalk interference signal z_i(n) are affected by power amplifier nonlinear distortion and memory effects, including:

modeling a power amplifier using a memory polynomial having a coefficient h_i,k,qK and Q are respectively the nonlinear order and the memory depth of the power amplifier memory polynomial model, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, K and Q are respectively the highest nonlinear order and the maximum memory depth of the power amplifier memory polynomial model, odd represents an odd set, and a signal z is a signal_i(n) output y affected by non-linear distortion and memory effect of power amplifier_i(n) according to the following formula:

7. the method of claim 1 for a MIMO transmitter full loopDigital predistortion method of path distortion compensation, characterized in that in the eighth step, said method is based on an input baseband signal s_i(n) and a feedback signal g_i(n) calculating the independent predistorter parameter values in each branch, including:

one-frame input baseband signal s with continuous acquisition frame length L_i(n) forming an input baseband signal matrix S_i＝[s_i(1) s_i(2) ... s_i(L)]^TCollecting the same frame s_i(n) corresponding feedback signal g_i(n); storing a memory polynomial matrix according to the following format

In (1),

is a matrix of LxK (Q +1), wherein the sub-matrices

Is a matrix of LxK with a matrix element of a_k(g_i(n))＝g_i(n)|g_i(n)|^k-1The structure is as follows:

according to S_iAnd

to calculate the independent predistorter coefficient matrix D in each branch_i＝{d_i,k,qK and Q are respectively the nonlinear order and the memory depth of the memory polynomial model of the predistorter, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, and K and Q are respectively the highest nonlinear order and the maximum memory depth of the memory polynomial model of the predistorter; odd represents an odd set, and the parameter value calculation process of the predistorter is carried out according to the following formula:

wherein inv (·) represents the inversion of the square matrix, and H represents the conjugate transpose of the matrix.

8. The digital predistortion method for full loop distortion compensation of MIMO transmitter as claimed in claim 1, wherein in the ninth step, said predistorter is used to apply the input baseband signal s_i(n) inverse processing of the non-linear characteristic and the memory effect is carried out to obtain the predistortion signal w_i(n) comprising:

modeling a predistorter using a memory polynomial with coefficient d_i,k,qWherein K and Q are respectively the nonlinear order and the memory depth of the memory polynomial model of the predistorter, K is more than or equal to 0 and less than or equal to K, K belongs to odd, Q is more than or equal to 1 and less than or equal to Q, K and Q are respectively the highest nonlinear order and the maximum memory depth of the memory polynomial model of the predistorter, odd is represented by odd set, and input baseband signal s_i(n) obtaining a signal w after inverse processing of nonlinear characteristics and memory effect through a predistorter in a normal operation state_i(n) according to the following formula:

9. the method and system of claim 1 wherein the method is performed in an N x N MIMO transmitter system.

10. A digital predistortion method for full loop distortion compensation of a MIMO transmitter as claimed in any of claims 1 to 9, wherein the digital predistortion system for full loop distortion compensation of a MIMO transmitter comprises:

n independent predistorters, 1 joint pre-canceller, N independent predistorter parameter calculation modules and 1 joint pre-canceller parameter calculation module; the quadrature modulator error, the nonlinear crosstalk effect, the linear crosstalk effect and the quadrature demodulator error are respectively equivalent to a quadrature modulator error module, a nonlinear crosstalk module, a linear crosstalk module and a quadrature demodulator error module.

Images