A digital predistorter for a power amplifier of a portable terminal

Abstract

The application provides a digital pre-distorter for a power amplifier of a portable terminal, which is suitable for the case of amplifier load mismatch, comprising: a coordinate rotation digital computer module for extracting the phase and amplitude of a plurality of input signals; a coefficient selection module for extracting the amplitude of the input signal and the timely signal and selecting the amplitude within a predetermined threshold interval as a coefficient; an operation module for generating a plurality of non-linear base functions by using all the amplitudes, phases and coefficients; and a multiplication addition module for obtaining an output signal by using the plurality of non-linear base functions and the phase of a cross signal. The application selects the coefficient by comparing the input signal with the threshold value, establishes a modified double-input pre-distorter, considers not only the original input signal but also the reflected signal, and processes the cross term of the input signal and the reflected signal, so that the complexity can be significantly reduced while the modeling accuracy is maintained.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of mobile communication, and particularly relates to a digital predistorter for a power amplifier of a portable terminal. BACKGROUND

[0002] A power amplifier (PA) is one of the core devices of a radio frequency device front end and is a device with the most power consumption in a transmitter. Therefore, the efficiency of the PA directly affects the performance of the transmitter. However, the PA usually suffers from severe nonlinear distortion when providing high efficiency, which requires additional linearization techniques. At present, in order to meet the requirements of linearity and efficiency, a digital predistortion (DPD) technique is widely applied in the transmitter to improve the linearity of the PA.

[0003] In a transmitter of a mobile phone, the load of the PA is sensitive to various factors, such as a handheld position, and the PA suffers from frequent load mismatch and produces complex nonlinear distortion. With the increase of the signal bandwidth of the 5G system, this problem becomes more serious. Traditional behavior models of the DPD technique, such as a memory polynomial (MP), a general memory polynomial (GMP), a decomposed vector rotation (DVR) model, and a model based on an amplitude selective affine (MSA) function, suffer from significant performance loss when compensating for the nonlinear distortion of the amplifier under load mismatch. Therefore, an expected DPD scheme is established under the condition of load mismatch, which can compensate for the nonlinear distortion of the amplifier while maintaining low complexity. In the current research, there are an extended amplitude selective affine (EMSA) function model and a time-domain poly-harmonic distortion (TD-PHD) device based on the theory of a poly-harmonic distortion (PHD) model. Although the modeling accuracy is significantly improved, the complexity is also increased due to the power and cost limitations of the mobile phone, which is also a key consideration. SUMMARY

[0004] In order to solve the above problems in the prior art, the application provides a digital predistorter for a power amplifier of a portable terminal. The technical problem to be solved by the application is solved by the following technical scheme:

[0005] The digital predistorter for the power amplifier of the portable terminal comprises:

[0006] A coordinate rotation digital computer module is configured to input a reflected signal, an input signal, and a cross signal, and decompose phases and amplitudes of the three signals to obtain corresponding amplitudes and phases.

[0007] A coefficient selection module is configured to extract amplitudes of the inputted timely signal and the input signal, and select coefficients corresponding to the amplitudes in a predetermined threshold interval.

[0008] An operation module is configured to generate a plurality of nonlinear base functions by using all the amplitudes, phases, and coefficients.

[0009] a multiplication-addition module, configured to obtain an output signal by using the plurality of non-linear base functions and the phase of the cross signal.

[0010] Optionally, the coordinate rotation digital computer module is specifically configured to:

[0011] input a reflected signal a2(n-m-l), an input signal a1(n-m), and a cross signal and respectively decompose the phase and amplitude of the three to obtain the amplitude |Φ1| = |a1(n-m)| and the phase of the input signal a1(n-m), and the amplitude |Φ2| = |a2(n-m-l)| and the phase of the reflected signal a2(n-m-l). wherein θ1 is a phase angle of the input signal, and θ2 is a phase angle of the cross signal.

[0012] Optionally, the coefficient selection module is specifically configured to:

[0013] extract the amplitude |a1(n)| of the input prompt signal a1(n) and the amplitude |a1(n-m)| of the input signal;

[0014] compare |a1(n)| with a threshold sequence β1, β2,..., β K to determine a first target amplitude value of a1(n) in a predetermined threshold interval [β q-1 , β q ).

[0015] compare |a1(n-m)| with a threshold sequence β1, β2,..., β K to determine a second target amplitude value in a predetermined threshold interval [β q-1 , β q ).

[0016] find a gain and an offset corresponding to the first target amplitude value and a gain and an offset corresponding to the second target amplitude value in the pre-stored coefficients.

[0017] wherein M1, M2, and M3 are all memory depths, and L2 and L3 are cross term delay lengths.

[0018] all the gains and offsets are taken as coefficients of the model.

[0019] Optionally, the pre-stored coefficients include coefficients corresponding to a plurality of regions, and each region corresponding coefficient is obtained through training.

[0020] Optionally, the process of training the coefficients of each region comprises:

[0021] Obtain a training set with 5000 samples;

[0022] According to the amplitude range of the input signals a'(n) and a'(n-m) contained in the training set, divide (K-1) regions, and the threshold of each region is β1, β2,..., β K ; wherein the relationship between the input signal a'(n-m) and the output signal u(n-m) of each region is u(n-m) = (A|a'(n-m)|+B)e jθ(n-m) ;

[0023] Use the least square method to estimate the coefficients A and B corresponding to each region, and store the coefficients A and B corresponding to each region.

[0024] Optionally, the operation module is specifically configured to:

[0025] Generate three nonlinear basis functions using all amplitudes, phases, and coefficients, denoted as:

[0026]

[0027] Optionally, the multiplication and addition module is specifically configured to:

[0028] Add the nonlinear basis functions f1 and f2, and then multiply the result by the phase of the input signal to obtain a first multiplication result;

[0029] Multiply the nonlinear basis function f3 by the phase of the cross signal to obtain a second multiplication result;

[0030] Obtain the output signal using the first multiplication result and the second multiplication result.

[0031] Optionally, the output signal is denoted as:

[0032]

[0033] Beneficial effects:

[0034] The application provides a digital predistorter for a portable terminal power amplifier, which is suitable for the case of amplifier load mismatch, and comprises a coordinate rotation digital computer module, which is used for inputting a reflected signal, an input signal and a cross signal, and decomposing the phase and amplitude of the three signals respectively to obtain corresponding amplitudes and phases; a coefficient selection module, which is used for extracting the amplitudes of the input timely signal and the input signal, and selecting the coefficients corresponding to the amplitudes in a predetermined threshold interval; an operation module, which is used for generating a plurality of nonlinear base functions by using all the amplitudes, phases and coefficients; and a multiplication addition module, which is used for obtaining an output signal by using the plurality of nonlinear base functions and the phase of the cross signal. The digital predistorter selects the coefficients of the model by comparing the input signal with the threshold, and establishes a modified double-input predistorter. The input of the predistorter of the application not only considers the original input signal, but also considers the reflected signal, and by processing the cross term of the input signal and the reflected signal, the complexity can be significantly reduced while maintaining high modeling accuracy.

[0035] The application will be further described in detail below in combination with the drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a schematic diagram of a digital predistorter for a portable terminal power amplifier provided by the application;

[0037] Figure 2 is a schematic diagram of a multiplication addition module calculation process provided by the application. DETAILED DESCRIPTION

[0038] The application will be further described in detail below in combination with the drawings and embodiments.

[0039] REFERENCE Figure 1 The application provides a digital predistorter for a portable terminal power amplifier, which comprises:

[0040] The coordinate rotation digital computer module is used for inputting a reflected signal, an input signal and a cross signal, and decomposing the phase and amplitude of the three signals respectively to obtain corresponding amplitudes and phases.

[0041] The coordinate rotation digital computer module inputs a reflected signal a2(n-m-l), an input signal a1(n-m) and a cross signal and decomposes the phase and amplitude of the three signals respectively to obtain the amplitude |Φ1|=|a1(n-m)| and the phase of the input signal a1(n-m), the amplitude |Φ2|=|a2(n-m-l)| and the phase of the reflected signal a2(n-m-l). Wherein, θ1 is the phase angle of the input signal, and θ2 is the phase angle of the cross signal.

[0042] a coefficient selection module, configured to extract the input prompt signal, the amplitude of the input signal, and select a coefficient corresponding to the amplitude in a predetermined threshold interval;

[0043] The coefficient selection module extracts the amplitude |a1(n)| of the input prompt signal a1(n) and the amplitude |a1(n-m)| of the input signal; compares |a1(n)| with the threshold sequence β1, β2,..., β K , and determines the first target amplitude value of a1(n) in the predetermined threshold interval [β q-1 , β q ); compares |a1(n-m)| with the threshold sequence β1, β2,..., β K , and determines the second target amplitude value in the predetermined threshold interval [β q-1 , β q ); and finds the gain and offset corresponding to the first target amplitude value and the gain and offset corresponding to the second target amplitude value in the pre-stored coefficients.

[0044] It is worth noting that the predetermined threshold interval can be determined according to the distribution of the amplitude and the modeling accuracy. For example, the distribution range of the amplitude of the whole signal is analyzed, such as 0-2.5, and β q is 2.5. If the modeling accuracy is a little higher, 0-2.5 can be divided into 10 segments, or if the modeling accuracy is a little higher, it can be divided into 20 segments. If the modeling accuracy is a little lower, it can be divided into 5 segments. If the number of segments is larger, the corresponding complexity is higher. Therefore, the predetermined threshold interval can be selected according to the actual modeling accuracy and complexity.

[0045] The pre-stored coefficients include coefficients corresponding to a plurality of regions, and the coefficients corresponding to each region are obtained by training. The process of training the coefficients of each region includes:

[0046] obtain a training set with 5000 samples; divide (K-1) regions according to the amplitude range of the input signals a'(n) and a'(n-m) contained in the training set, and the threshold values corresponding to each region are β1, β2,..., β K ; wherein the relationship between the input signal a'(n-m) and the output signal u(n-m) of each region training set is u(n-m) = (A|a'(n-m)|+B)e jθ(n-m); using the least squares (Least Squares, LS) to estimate the coefficients A and B corresponding to each region, and store the coefficients A and B corresponding to each region.

[0047] The operation module is configured to generate a plurality of nonlinear base functions by using all of the amplitudes, phases and coefficients.

[0048] The operation module generates three nonlinear base functions by A|Φ|+B using all of the amplitudes, phases and coefficients, and the three nonlinear base functions are represented as:

[0049]

[0050] The multiplication-addition module is configured to obtain an output signal by using the plurality of nonlinear base functions and the phase of the cross signal.

[0051] The multiplication-addition module adds the nonlinear base functions f1 and f2, and then multiplies the phase of the input signal to obtain a first multiplication result; multiplies the nonlinear base function f3 and the phase of the cross signal to obtain a second multiplication result; and obtains the output signal by using the first multiplication result and the second multiplication result, and the output signal is represented as:

[0052]

[0053] The digital predistorter for the power amplifier of the portable terminal provided in the application uses a 5G NR signal with a 100MHz bandwidth and a 491.52Msps sampling rate to verify the Doherty PA in the 3.3-3.8GHz band. The synchronous vector signal generator R&S SMW200A is used to generate radio frequency signal streams, which are fed to the amplifier and then captured by the spectrum analyzer R&S FSW43. The adjustable load is connected to the output of the PA through a coupler to control the load impedance. The performance of the CR-TD-PHD is verified in the forward modeling process, and compared with the TD-PHD, the application realizes comparable linearization performance under load mismatch conditions, while reducing the flop number by 85%. Through experimental verification, it can be concluded that the application can significantly reduce the model complexity while maintaining good linearization performance.

[0054] It should be noted that the terms "first", "second" in the application are only used for description purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" can explicitly or implicitly include one or more of the features. In the description of the application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0055] Although the application has been described in connection with specific embodiments thereof, it will be understood that it is capable of further modifications and this application is intended to cover any variations using the principles of the application. For example, "comprising" shall not exclude other elements or steps. The inclusion of one particular limitation does not exclude others where these are not mutually exclusive. The skilled person will understand that many of the examples given above will cover one or a combination of alternative features.

[0056] The above description is further to specific preferred embodiments of the present application and is not to be construed as limiting the present application to only these descriptions. Those ordinarily skilled in the art to which the present application pertains will be able to understand and implement other variations of the disclosed embodiments without departing from the spirit of the present application. Such variations are intended to be within the scope of the present application.

Claims

1. A digital predistorter for a power amplifier of a portable terminal, characterized by, The application relates to a method for generating an output signal, comprising: a coordinate rotation digital computer module for inputting a reflection signal, an input signal and a cross signal, and respectively decomposing phases and amplitudes of the three to obtain corresponding amplitudes and phases; a coefficient selection module for extracting amplitudes of the input signal and the input signal, and selecting coefficients corresponding to the amplitudes in a predetermined threshold interval; an operation module for generating a plurality of nonlinear base functions by using all amplitudes, phases and coefficients; a multiplication and addition module for obtaining an output signal by using the plurality of nonlinear base functions and the phase of the cross signal; the coordinate rotation digital computer module is specifically used for: input reflection signal a2(n-m-l), input signal a1(n-m) and cross signal and the phase and amplitude of each of them are decomposed to get the amplitude |Φ1| = |a1(n-m)| and phase of a1(n-m) and the amplitude |Φ2| = |a2(n-m-l)| and phase of a2(n-m-l) where θ1 is the phase angle of the input signal and θ2 is the phase angle of the cross signal. the coefficient selection module is specifically used for: extracting an amplitude |a1(n)| of the input signal a1(n) and an amplitude |a1(n-m)| of the input signal; |a1(n)| is compared with a threshold sequence β1, β2,..., β K to determine a first target amplitude of a1(n) within a predetermined threshold interval [β q-1 , β q ). |a1(n-m)| is compared with a threshold sequence β1, β2,..., β K The predetermined threshold interval [β q-1 , β q ) is determined by comparing the second target amplitude. Looking up a gain corresponding to the first target amplitude value among the pre-stored coefficients and an offset corresponding to the second target amplitude value and an offset wherein M1, M2 and M3 are memory depths, and L2 and L3 are cross term delay lengths; all gains and offsets are used as coefficients of the model; the operation module is specifically used for: generating three nonlinear base functions by using all amplitudes, phases and coefficients, and the three nonlinear base functions are represented as: the multiplication and addition module is specifically used for: The nonlinear base functions f1 and f2 are added, and then multiplied with the phase of the input signal to obtain a first multiplication result; multiplying the non-linear base function f3 with the phase of the cross signal to obtain a second multiplication result; obtaining an output signal by using the first multiplication result and the second multiplication result; the output signal is represented as:

2. The digital predistorter for a power amplifier of a portable terminal according to claim 1, wherein the pre-stored coefficients include coefficients corresponding to a plurality of regions, and the coefficients corresponding to each region are obtained by training.

3. The digital predistorter for a power amplifier of a portable terminal according to claim 2, wherein The process of training the coefficients of each region comprises: obtaining a training set with 5000 samples; According to the amplitude range of the input signals a'(n) and a'(n-m) contained in the training set, K regions are divided, and the corresponding threshold values of each region are β1, β2,..., βK, respectively. K ; wherein the relationship between the input signal a'(n-m) and the output signal u(n-m) of each region training set is u(n-m) = (A|a'(n-m)|+B)e jθ(n-m) ; using a least square method to estimate coefficients A and B corresponding to each region, and storing the coefficients A and B corresponding to each region.

Images