Digital pre-distortion processing method, WIFI broadband system and computer storage medium
By performing energy normalization and polynomial compensation coefficient update methods on the output signal of the power amplifier, the unstable and cumbersome digital predistortion processing in the prior art is solved, and effective nonlinear correction of broadband OFDM signals is achieved, which improves the System stability and simplicity.
Patent Information
- Application Number
- CN202011378320.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-12-01
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2040-12-01
AI Technical Summary
The existing digital predistortion processing methods have problems with instability and cumbersome implementation processes when correcting the nonlinear influence of RF power amplifiers, especially in broadband OFDM signal processing, which is difficult to effectively eliminate in-band distortion and spectrum growth.
A digital predistortion processing method is proposed. By normalizing the output signal of the power amplifier and performing digital predistortion compensation based on the polynomial compensation coefficient, the polynomial compensation coefficient is updated using nonlinear processing signals, combined with normalization of the minimum mean square error The criteria perform coefficient updates, simplifying the computational process and improving stability.
It realizes a more stable digital predistortion processing for the signal, eliminates the nonlinear characteristics of the signal after passing through the power amplifier, improves the stability and simplicity of the system, and is suitable for memoryless and memory-free power amplifier models.
Smart Images

Figure CN114584081B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of signal processing, and in particular to a digital predistortion processing method, a WIFI broadband system and a non-temporary computer storage medium. Background Art
[0002] DPD (Digital Pre-Distortion) is a baseband signal processing method to correct the nonlinear effects of RF PA (Radio Frequency Power Amplifiers). Nonlinearity can cause out-of-band leakage, spectrum growth, and in-band distortion. The distortion of in-band signals can cause an increase in the bit error rate. For broadband OFDM (Orthogonal Frequency Division Multiplexing) signals, their high peak-to-average ratio makes them more susceptible to nonlinearity. In order to correct the nonlinear effects of RF power amplifiers, pre-distortion processing is usually performed at the digital end. There are two main methods for implementing DPD: a lookup table-based implementation method and a polynomial-based training method. For the lookup table-based implementation method, it is mainly used for power amplifier models without memory. For power amplifier models with memory, there is no effect after compensation. For the polynomial-based training method, it is applicable to both power amplifier models without memory and with memory, but in its implementation, it is affected by the system model, there is a certain instability and the process is cumbersome. Summary of the invention
[0003] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, one object of the present invention is to provide a digital predistortion processing method, which can obtain a more stable processing signal and has a relatively simple operation method.
[0004] The second objective of the present invention is to provide a WIFI broadband system.
[0005] A third object of the present invention is to provide a non-transitory computer storage medium.
[0006] In order to achieve the above-mentioned purpose, the digital predistortion processing method of the first aspect embodiment of the present invention is used in a WIFI broadband system, and the digital predistortion processing method includes: obtaining an output signal of a power amplifier and a transmitting end signal in the WIFI broadband system, performing energy normalization on the output signal of the power amplifier based on the transmitting end signal, and obtaining an energy-normalized signal; performing digital predistortion compensation on the energy-normalized signal according to a polynomial compensation coefficient, and obtaining a digital predistortion compensated signal, wherein the digital predistortion compensation includes nonlinear processing; performing a subtraction operation on the digital predistortion compensated signal and the transmitting end signal to obtain an error signal; obtaining a nonlinear processing signal in the digital predistortion compensation, and obtaining the polynomial compensation coefficient according to the error signal and the nonlinear processing signal; and feeding the polynomial compensation coefficient back to the digital predistortion compensation.
[0007] According to the digital predistortion processing method of the embodiment of the present invention, the signal is digitally predistorted, so that the nonlinear characteristics of the signal can be eliminated after the digital predistortion signal passes through the power amplifier. By normalizing the energy of the output signal of the power amplifier based on the transmitting end signal, the energy of the output signal of the power amplifier and the signal to be compensated at the transmitting end can be kept consistent, and the signal is more stable during the digital predistortion processing.
[0008] In some embodiments, energy normalization is performed on an output signal of a power amplifier based on a transmitting end signal to obtain an energy-normalized signal, including: obtaining an absolute value of the output signal of the power amplifier, adding the absolute value of the output signal of the power amplifier to a first time-delay signal to obtain a first operation value, wherein the first operation value is subjected to time-delay processing to obtain the first time-delay signal; obtaining an absolute value of the transmitting end signal, adding the absolute value of the transmitting end signal to a second time-delay signal to obtain a second operation value, wherein the second operation value is subjected to time-delay processing to obtain the second time-delay signal; dividing the first operation value and the second operation value to obtain a third operation value; and multiplying the third operation value by the output signal of the power amplifier to obtain an energy-normalized signal.
[0009] In some embodiments, digital pre-distortion compensation is performed on the energy normalized signal according to the polynomial compensation coefficient to obtain the digital pre-distortion compensated signal, including: delay processing the energy normalized signal to obtain a third time-delayed signal; nonlinear processing the third time-delayed signal to obtain a nonlinear processed signal; obtaining the polynomial compensation coefficient, multiplying the nonlinear processed signal with the polynomial compensation coefficient to obtain the multiplied signal; and accumulating the multiplied signal to obtain the digital pre-distortion compensated signal.
[0010] In some embodiments, delay processing is performed on the energy-normalized signal to obtain a third delayed signal, including: performing multiple levels of delay processing on the energy-normalized signal in sequence to obtain multiple levels of third delayed signals.
[0011] In some embodiments, the energy-normalized signal is sequentially subjected to multi-level delay processing to obtain a multi-level third delayed signal, including: using the energy-normalized signal as a 0th-level third delayed signal; performing a first-level delay processing on the 0th-level third delayed signal to obtain a first-level third delayed signal; performing a second-level delay processing on the first-level third delayed signal to obtain a second-level third delayed signal; performing a third-level delay processing on the second-level third delayed signal to obtain a third-level third delayed signal; performing a fourth-level delay processing on the third-level third delayed signal to obtain a fourth-level third delayed signal.
[0012] In some embodiments, the third time-delayed signal is nonlinearly processed to obtain a nonlinear processed signal, including: performing nonlinear processing on the 0th level third time-delayed signal and obtaining three first nonlinear processed signals; performing nonlinear processing on the first level third time-delayed signal and obtaining three second nonlinear processed signals; performing nonlinear processing on the second level third time-delayed signal and obtaining three third nonlinear processed signals; performing nonlinear processing on the third level third time-delayed signal and obtaining three fourth nonlinear processed signals; performing nonlinear processing on the fourth level third time-delayed signal and obtaining three fifth nonlinear processed signals.
[0013] In some embodiments, the polynomial compensation coefficient is obtained, and the nonlinear processing signal is multiplied by the polynomial compensation coefficient to obtain the signal after the multiplication operation, including: the polynomial compensation coefficient includes fifteen compensation coefficients; based on the fifteen compensation coefficients, three channels of the first nonlinear processing signals, three channels of the second nonlinear processing signals, three channels of the third nonlinear processing signals, three channels of the fourth nonlinear processing signals and three channels of the fifth nonlinear processing signals are multiplied respectively to obtain the signal after the multiplication operation.
[0014] In some embodiments, the signals after the multiplication operation are accumulated to obtain the signals after digital pre-distortion compensation, including: accumulating the signals after the three-way first nonlinear processing signal multiplication operation to obtain a first accumulated value; accumulating the signals after the three-way second nonlinear processing signal multiplication operation to obtain a second accumulated value; accumulating the signals after the three-way third nonlinear processing signal multiplication operation to obtain a third accumulated value; accumulating the signals after the three-way fourth nonlinear processing signal multiplication operation to obtain a fourth accumulated value; accumulating the signals after the three-way fifth nonlinear processing signal multiplication operation to obtain a fifth accumulated value; accumulating the first accumulated value, the second accumulated value, the third accumulated value, the fourth accumulated value and the fifth accumulated value to obtain a total accumulated value as the signal after digital pre-distortion compensation.
[0015] In some embodiments, a nonlinear processing signal in digital predistortion compensation is obtained, and the polynomial compensation coefficient is obtained according to the error signal and the nonlinear processing signal, including: obtaining multiple nonlinear processing signals in digital predistortion compensation; obtaining the conjugate signal of each nonlinear processing signal; multiplying the conjugate signal of each nonlinear processing signal with the error signal and the coefficient update step to obtain a plurality of product values; obtaining the polynomial compensation coefficient at a previous moment, and respectively adding the plurality of product values and the polynomial compensation coefficient at the previous moment to obtain the polynomial compensation coefficient at a current moment.
[0016] In some embodiments, the digital pre-distortion processing method further includes: normalizing the coefficient update step size.
[0017] In some embodiments, the coefficient update step is normalized, including: acquiring multiple nonlinear processing signals; counting the energy of each nonlinear processing signal; calculating the energy sum of the multiple nonlinear processing signals; performing LUT processing on the energy sum to obtain the inverse of the energy sum; and multiplying the inverse of the energy sum with a fixed step value to obtain a normalized coefficient update step.
[0018] In some embodiments, the polynomial compensation coefficients satisfy the following normalized minimum mean square error convergence formula:
[0019] in,
[0020] Among them, c m,k (n) represents the polynomial compensation coefficient at the nth moment, c m,k (n-1) represents the polynomial compensation coefficient at the previous moment, μ represents the fixed step size, e(n) represents the error signal, z m,k(n) represents the signal after nonlinear processing, the symbol “*” represents conjugation, m represents the memory length index, k represents the nonlinear order index, and P(n) represents the normalized energy of the polynomial structure at the nth moment.
[0021] In some embodiments, M=5, K=3.
[0022] In order to achieve the above-mentioned object, the WIFI broadband system of the second aspect of the present invention includes: a power amplifier; and a digital predistortion processing device, wherein the digital predistortion processing device is connected to the power amplifier and is used to implement any one of the digital predistortion processing methods described above. The digital predistortion processing device can be implemented by hardware or by software to implement digital predistortion processing of the signal.
[0023] In the WIFI broadband system according to the embodiment of the present invention, the power amplifier is used to output the amplified baseband signal, but the broadband signal may be affected by nonlinearity. The digital pre-distortion processing device of the broadband system can correct the nonlinear influence of the power amplifier and perform pre-distortion processing at the digital end, thereby ensuring system stability while being simple to implement.
[0024] In order to achieve the above-mentioned purpose, the non-temporary computer storage medium of the third aspect embodiment of the present invention stores a computer program thereon, and is characterized in that when the computer program is executed, any one of the digital pre-distortion processing methods described is implemented.
[0025] According to the non-temporary computer storage medium of the embodiment of the present invention, the computer program stored thereon is run and can send executable instructions. When the instructions are executed, the digital pre-distortion processing method of the embodiment of the present invention can be implemented, so that the original signal is digitally pre-distorted before passing through the power amplifier to eliminate the influence of the signal's nonlinear characteristics caused by the signal passing through the power amplifier.
[0026] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:
[0028] Figure 1 is a flow chart of a digital predistortion processing method according to an embodiment of the present invention;
[0029] Figure 2 is a flow chart of obtaining an energy-normalized signal according to an embodiment of the present invention;
[0030] Figure 3 is a flow chart of updating polynomial compensation coefficients based on the normalized minimum mean square error criterion according to an embodiment of the present invention;
[0031] Figure 4 is a flow chart of digital pre-distortion compensation according to one embodiment of the present invention;
[0032] Figure 5 is a flow chart of a calibration process of a pre-distortion calibration device according to an embodiment of the present invention;
[0033] Figure 6 is a block diagram of a broadband system according to an embodiment of the present invention;
[0034] Figure 7 is a schematic diagram of a simulation model structure of a digital pre-distortion system according to an embodiment of the present invention;
[0035] Figure 8 is a schematic diagram describing the structure of a Saleh model simulating a memoryless power amplifier according to an embodiment of the present invention;
[0036] Figure 9a is a waveform diagram of the relationship between the error signal amplitude and the sampling time according to one embodiment of the present invention;
[0037] Figure 9b is a waveform diagram showing the relationship between the amplitude of the error signal after filtering and the sampling time according to one embodiment of the present invention;
[0038] Fig.9c is a waveform diagram showing the relationship between the amplitude (dB) value of the error signal after filtering and the sampling time according to one embodiment of the present invention;
[0039] Fig.10a is a waveform diagram of the time domain response of the original signal of one embodiment of the present invention;
[0040] Fig.10b It is a waveform diagram of the time domain response of the output signal after the original signal is processed by digital pre-distortion and amplified by the power amplifier according to one embodiment of the present invention;
[0041] Fig.11a is a waveform diagram of a frequency domain response of an original signal according to an embodiment of the present invention;
[0042] Fig.11b It is a waveform diagram of the frequency domain response of the original signal after being amplified by the power amplifier and the frequency domain response of the original signal after being processed by digital predistortion and amplified by the power amplifier according to an embodiment of the present invention;
[0043] Fig.12ais a waveform diagram of the amplitude of an original signal and the amplitude of a signal after the original signal is amplified by a power amplifier according to an embodiment of the present invention;
[0044] Figure 12b It is a waveform diagram of the amplitude of the original signal and the amplitude of the signal after the original signal is processed by digital pre-distortion and amplified by a power amplifier according to an embodiment of the present invention;
[0045] Fig.13a It is a waveform diagram of the phase of an original signal and the phase of a signal after the original signal is amplified by a power amplifier according to an embodiment of the present invention;
[0046] Fig.13b It is a waveform diagram of the phase of an original signal according to an embodiment of the present invention and the phase of a signal after the original signal is processed by digital predistortion and amplified by a power amplifier.
[0047] Reference numerals:
[0048] A broadband system 1000 , a digital predistortion processing device 100 of the broadband system, and a power amplifier 200 . DETAILED DESCRIPTION
[0049] Embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. Embodiments of the present invention are described in detail below.
[0050] The digital pre-distortion processing device and the WIFI broadband system according to the embodiments of the present invention will be described below with reference to the accompanying drawings.
[0051] Figure 1 is a flow chart of a digital pre-distortion processing method according to an embodiment of the present invention.
[0052] like Figure 1 As shown, the digital pre-distortion processing method of the embodiment of the present invention may at least include steps S101-S105.
[0053] S101, obtaining the output signal of the power amplifier and the transmitting end signal of the WIFI broadband system, and performing energy normalization on the output signal of the power amplifier based on the transmitting end signal to obtain the energy normalized signal. By performing normalization processing on the output signal of the power amplifier before performing compensation processing on the output signal of the power amplifier, the energy of the output signal of the power amplifier and the signal to be compensated at the transmitting end can be kept consistent, and a more stable processed signal can be obtained, so as to be more stable during the system simulation process.
[0054] S102, performing digital predistortion compensation on the energy normalized signal according to the polynomial compensation coefficient to obtain a digital predistortion compensated signal, wherein the digital predistortion compensation includes nonlinear processing. Digital predistortion compensation is performed on the energy normalized signal to eliminate the influence of the signal nonlinear characteristics caused by the signal passing through the power amplifier.
[0055] S103, performing a subtraction operation on the signal after digital predistortion compensation and the transmitting end signal to obtain an error signal. For example, the obtained error signal can be used in the calculation of updating the polynomial compensation coefficient.
[0056] S104, obtaining a nonlinear processing signal in digital predistortion compensation, and obtaining a polynomial compensation coefficient according to the error signal and the nonlinear processing signal. For example, multiple nonlinear processing signals and error signals may be calculated to obtain updated values of multiple polynomial compensation coefficients.
[0057] S105, feeding back the polynomial compensation coefficients to the digital predistortion compensation. For example, after the digital predistortion estimation is completed, the generated polynomial compensation coefficients can be fed back to the digital predistortion compensation process, and the estimation process and the compensation process are separated, which is simple to implement.
[0058] According to the digital predistortion processing method of the embodiment of the present invention, the signal is digitally predistorted, which can eliminate the nonlinear characteristics of the signal caused by the power amplifier. By normalizing the energy of the output signal of the power amplifier based on the transmitting end signal, the output signal of the power amplifier and the energy of the signal to be compensated at the transmitting end can be kept consistent, and the signal can be more stable during the digital predistortion processing.
[0059] Figure 2 The flowchart of obtaining an energy-normalized signal according to an embodiment of the present invention.
[0060] In an embodiment of the present invention, Figure 2 As shown, the output signal of the power amplifier at time n can be recorded as y(n), and the output signal of the system transmitter at time n can be recorded as x(n). The output signal y(n) of the power amplifier and the output signal x(n) of the system transmitter can be normalized to obtain an energy-normalized signal, which can keep the energy of the signal x(n) that needs to be compensated at the transmitter consistent. In addition, the calculation method can be simplified to obtain a more stable processed signal.
[0061] In an embodiment, the output signal of the power amplifier can be energy normalized based on the transmitting end signal, as shown in formula (1-1), wherein the signal energy can be normalized approximately by taking the absolute value of the signal for normalization, and the specific implementation process of obtaining the energy normalized signal can be achieved by a specific algorithm. For example, in the present invention, N can be selected as 4096 points, and the output signal y(n) of the power amplifier is calculated by taking the absolute value in the abs subunit, and the processed signal can be recorded as |y(n)|, and the processed signal |y(n)| is added to the first delayed signal to obtain a first operation value, wherein the obtained first operation value can be delayed to obtain the first delayed signal, and this process is iterated and looped, and the first operation value can be obtained after cumulative calculation. The absolute value of the transmitting end signal x(n) can also be calculated in the abs subunit, and the processed signal is recorded as |x(n)|. The processed signal |x(n)| is added to the second delayed signal to obtain a second operation value, wherein the obtained second operation value can be delayed to obtain a second delayed signal, and this process is iterated and looped, and the second operation value can be obtained after cumulative calculation. The first operation value and the second operation value may be divided to obtain a third operation value. The third operation value may be recorded as 1 / G, and the third operation value 1 / G and the output signal y(n) of the power amplifier may be multiplied to obtain the energy-normalized signal y1(n), and y1(n)=y(n)*1 / G may be recorded.
[0062]
[0063] According to the method for performing energy normalization processing on the output signal of a power amplifier based on a transmitting end signal according to an embodiment of the present invention, before the output signal y(n) of the power amplifier is compensated, the output signal y(n) of the power amplifier is first normalized. The operation method is relatively simple and the energy of the output signal y(n) of the power amplifier and the signal x(n) to be compensated at the transmitting end can be kept consistent, so that the signal can be more stable during the system simulation process.
[0064] Figure 3 The present invention is a flowchart of updating polynomial compensation coefficients based on the normalized minimum mean square error criterion according to an embodiment of the present invention.
[0065] In an embodiment, according to the polynomial compensation coefficient, the digital predistortion compensation module performs digital predistortion compensation processing on the energy normalized signal to obtain a digital predistortion compensated signal. The specific implementation process can be implemented by the specific algorithm of formula (1-2). For example, in formula (1-2), the energy normalized output signal of the power amplifier at time n can be recorded as y1(n), and the signal y1(n) is subjected to hierarchical delay processing in the tap delay line to obtain a third delayed signal recorded as y1(nm). The nonlinear processing module can perform nonlinear processing on the third delayed signal and output the nonlinearly processed signal y1(nm)|y1(nm)| k The signal after nonlinear processing is multiplied by the polynomial compensation coefficient. For example, the polynomial compensation coefficient at the current time n is recorded as c m,k (n), after multiplication, the output signal c after multiplication m,k (n)y1(nm)|y1(nm)| k , and then perform addition operation on the multiplication signal to obtain the signal after digital predistortion compensation Among them, |y1(nm)| k It is the kth power of the absolute value of y1(nm), k is the nonlinear order index, and m is the memory depth index. The present invention combines the existing digital pre-distortion compensation algorithm model to select the memory length M as 5, the nonlinear order K as 3, and the sampling clock at a frequency of 160Mhz, which is easy to implement.
[0066]
[0067] In an embodiment, Figure 3 As shown, the energy-normalized signal is delayed in a tapped delay line to obtain a third delayed signal, wherein the tapped delay line can perform multiple levels of delay processing on the energy-normalized signal in sequence to obtain multiple levels of third delayed signals. For example, the energy-normalized signal can be subjected to four levels of delay processing in sequence to obtain a five-level third delayed signal.
[0068] Figure 4 is a flow chart of digital pre-distortion compensation according to an embodiment of the present invention.
[0069] In the embodiment of the present invention, for example, the multi-stage delay processing may include a four-stage delay processing, and the obtained multi-stage third delay signal may be a five-stage third delay signal. Figure 4As shown, the power amplifier outputs a signal y(n) at time n. After normalization, the signal obtained is y1(n), which is recorded as the third time-delayed signal of level 0. The third time-delayed signal y1(n) of level 0 is subjected to the first-level delay processing to obtain the third time-delayed signal of level 1, and the first-level third time-delayed signal can be recorded as y1(n-1). The first-level third time-delayed signal y1(n-1) is subjected to the second-level delay processing to obtain the second-level third time-delayed signal recorded as y1(n-2). The second-level third time-delayed signal y1(n-2) is subjected to the third-level delay processing to obtain the third-level third time-delayed signal recorded as y1(n-3). The third-level third time-delayed signal y1(n-3) is subjected to the fourth-level delay processing to obtain the fourth-level third time-delayed signal recorded as y1(n-4). Among them, Figure 4 In the figure, the nonlinear processing modules for the output signals of the first-stage delay, the second-stage delay and the third-stage delay are omitted here for the simplification of drawing.
[0070] In an embodiment, the energy normalized signal can be subjected to four-level delay processing in sequence. Each level of delay device can perform delay processing on the input signal to obtain a signal after delay processing. The output signals corresponding to each level are y1(n), y1(n-1), y1(n-2), y1(n-3), y1(n-4), and the five signals are simultaneously sent to the nonlinear processing module for nonlinear operation.
[0071] In an embodiment of the present invention, the third time-delayed signal may be subjected to nonlinear processing to obtain a nonlinearly processed signal, such as Figure 4 As shown, the five third time-delayed signals after multi-stage time-delay processing can be simultaneously processed nonlinearly. The specific calculation method can be implemented according to formula (1-3), where z m,k (n) represents the signal after nonlinear processing. For example, y1(nm) can be taken as the third delayed signal after delay processing, |y1(nm)| k It is the kth power of the absolute value of y1(nm), k is the nonlinear order index, and the nonlinear order can be selected as 3 for easy implementation.
[0072] z m,k (n)=y1(nm)|y1(nm)| k (1-3)
[0073] In an embodiment, Figure 4 As shown, for example, the third delayed signal y1(n) of level 0 can be recorded as z 0,0 (n), z 0,0 (n) = y1(n). The signal y1(n) is subjected to the second nonlinear processing, and the processing method is to multiply the absolute value of the third time-delayed signal y1(n) of the 0th level by the signal y1(n), and is recorded as z0,1 (n), z 0,1 (n) = y1(n)|y1(n)|. The signal y1(n) is subjected to the third nonlinear processing. The processing method is to square the absolute value of the third time-delayed signal y1(n) at level 0 and then multiply it by the signal y1(n), which is recorded as z 0,2 (n), z 0,2 (n)=y1(n)|y1(n)| 2 Finally, we can get the three-way signal z after nonlinear processing. 0,0 (n),z 0,1 (n) and z 0,2 (n).
[0074] In an embodiment, the first-stage third delayed signal y1(n-1) may be denoted as z 1,0 (n), z 1,0 (n) = y1(n-1). The signal y1(n-1) is processed by the second nonlinear process, and the absolute value of the first-stage third time-delay signal y1(n-1) is multiplied by the signal y1(n-1), which is recorded as z 1,1 (n), z 1,1 (n) = y1(n-1)|y1(n-1)|. After the signal y1(n-1) is processed by the third nonlinear process, the absolute value of the first-stage third time-delay signal y1(n-1) can be squared and then multiplied with the signal y1(n-1) to obtain z. 1,2 (n), z 1,2 (n) = y1(n-1)|y1(n-1)| 2 Finally, we can get the three-way signal z after nonlinear processing. 1,0 (n),z 1,1 (n) and z 1,2 (n) Figure 4 In the embodiment, the implementation process of performing nonlinear processing on the third time-delayed signal of the first stage and obtaining the signal after nonlinear processing is omitted here for the simplification of drawing.
[0075] In an embodiment, the second-stage third delayed signal y1(n-2) may be denoted as z 2,0 (n), z 2,0 (n) = y1(n-2). After the signal y1(n-2) is processed by the second nonlinear process, the absolute value of the second-stage third time-delay signal y1(n-2) can be multiplied by the signal y1(n-2), which is recorded as z 2,1 (n), z 2,1(n) = y1(n-2)|y1(n-2)|. After the signal y1(n-2) is processed by the third nonlinear process, the absolute value of the second-stage third time-delay signal y1(n-2) can be squared and then multiplied with the signal y1(n-2), which is recorded as z 2,2 (n), z 2,2 (n) = y1(n-2)|y1(n-2)| 2 Finally, we can get the three-way signal z after nonlinear processing. 2,0 (n),z 2,1 (n) and z 2,2 (n) Figure 4 In the embodiment, the implementation process of performing nonlinear processing on the second-stage third time-delayed signal and obtaining the signal after nonlinear processing is omitted here for the sake of simplifying the drawing.
[0076] In an embodiment, the third-stage third delayed signal y1(n-3) may be denoted as z 3,0 (n), z 3,0 (n) = y1(n-3). After the signal y1(n-3) is processed by the second nonlinear process, the absolute value of the third-stage third-delay signal y1(n-3) can be multiplied by the signal y1(n-3), which is recorded as z 3,1 (n), z 3,1 (n) = y1(n-3)|y1(n-3)|. After the signal y1(n-3) is processed by the third nonlinear path, the absolute value of the third-stage third-delay signal y1(n-3) can be squared and then multiplied with the signal y1(n-3), which is recorded as z 3,2 (n), z 3,2 (n) = y1(n-3)|y1(n-3)| 2 Finally, we can get the three-way signal z after nonlinear processing. 3,0 (n),z 3,1 (n) and z 3,2 (n) Figure 4 In the embodiment, the implementation process of performing nonlinear processing on the third-stage third-delay signal and obtaining the signal after nonlinear processing is omitted here for the sake of simplifying the drawing.
[0077] In an embodiment, the fourth-stage third delayed signal y1(n-4) may be recorded as z 4,0 (n), z 4,0 (n) = y1(n-4). After the signal y1(n-4) is processed by the second nonlinear process, the absolute value of the fourth-stage third time-delay signal y1(n-4) can be multiplied by the signal y1(n-4), which is recorded as z 4,1 (n), z 4,1(n) = y1(n-4)|y1(n-4)|. After the signal y1(n-4) is processed by the third nonlinear process, the absolute value of the fourth-stage third time-delay signal y1(n-4) can be squared and then multiplied with the signal y1(n-4) to obtain z. 4,2 (n), z 4,2 (n) = y1(n-4)|y1(n-4)| 2 Finally, we can get the three-way signal z after nonlinear processing. 4,0 (n),z 4,1 (n) and z 4,2 (n).
[0078] In the embodiment of the present invention, the five-level third time-delay signals y1(n), y1(n-1), y1(n-2), y1(n-3), and y1(n-4) obtained by the four-level time-delay processing can be subjected to nonlinear operations respectively. After the third time-delay signal of each level is subjected to nonlinear processing, three nonlinearly processed signals can be obtained. Therefore, in the calculation process, the nonlinear order can be selected as 3. Then, after the five-level third time-delay signals are subjected to nonlinear processing, a total of fifteen nonlinearly processed output signals can be obtained. The fifteen nonlinearly processed output signals can be respectively recorded as z 0,0 (n),z 0,1 (n),z 0,2 (n),z 1,0 (n),z 1,1 (n),z 1,2 (n),z 2,0 (n),z 2,1 (n),z 2,2 (n),z 3,0 (n),z 3,1 (n),z 3,2 (n),z 4,0 (n),z 4,1 (n), and z 4,2 (n).
[0079] In an embodiment of the present invention, a polynomial compensation coefficient may be obtained, and the nonlinear processing signal and the polynomial compensation coefficient may be multiplied to obtain a signal after the multiplication operation. Figure 4 As shown, the nonlinear processing signal is multiplied by the polynomial compensation coefficient. The specific calculation method can be implemented according to formula (1-2), where z m,k (n) represents the signal after nonlinear processing, y1(nm) is the signal after time delay processing, |y1(nm)| k is the kth power of the absolute value of y1(nm), k is the nonlinear order index, and the nonlinear order can be selected as 3 for easy implementation. Among them, the polynomial compensation coefficient is recorded as c m,k(n), for example, the compensation coefficients of the fifteen-way polynomial at time n can be respectively denoted as c 0,0 (n),c 0,1 (n),c 0,2 (n),c 1,0 (n),c 1,1 (n),c 1,2 (n),c 2,0 (n),c 2,1 (n),c 2,2 (n),c 3,0 (n),c 3,1 (n),c 3,2 (n),c 4,0 (n),c 4,1 (n) and c 4,2 (n). Each polynomial compensation coefficient is multiplied by the corresponding input signal after nonlinear processing to obtain a signal after multiplication. For example, the signal after multiplication can be recorded as c m,k (n)z m,k (n). Fifteen signals after nonlinear processing z m,k (n) and the 15-way polynomial compensation coefficient c m,k (n) correspond to multiplication and obtain fifteen signals, Figure 4 In order to simplify the drawing, the implementation process of multiplying the nine-way polynomial compensation coefficients with the corresponding input nonlinear processed signals is omitted.
[0080] In an embodiment of the present invention, the signal after the multiplication operation can be accumulated to obtain a signal after digital predistortion compensation. Figure 4 As shown, first, the signal c after the three-way multiplication operation is m,k (n)z m,k (n) is accumulated to obtain one-way accumulated value, then the signal after the fifteen-way multiplication operation can obtain five-way accumulated value. The five-way accumulated values are recorded as the first accumulated value, the second accumulated value, the third accumulated value, the fourth accumulated value and the fifth accumulated value. Figure 4 In the figure, the second accumulated value, the third accumulated value, and the fourth accumulated value are omitted here for the sake of simplifying the drawing. The five accumulated values, namely the first accumulated value, the second accumulated value, the third accumulated value, the fourth accumulated value, and the fifth accumulated value, are accumulated to obtain a total accumulated value. After the total accumulated value is output, a signal after digital pre-distortion compensation can be obtained. Among them, the specific calculation method of the signal after digital pre-distortion compensation can be implemented according to formula (1-4).
[0081]
[0082] In an embodiment of the present invention, Figure 3As shown, the polynomial compensation coefficient can be obtained according to the signal after digital predistortion compensation and the signal after nonlinear processing. The coefficient adaptive updating method based on the polynomial structure in the prior art has certain instability factors. The coefficient update of the embodiment of the present invention can provide a specific implementation structure for updating the coefficient based on the normalized minimum mean square error criterion during the algorithm implementation process. The process of implementing the coefficient update operation is relatively simple, and the result of the coefficient update is also more stable.
[0083] In an embodiment of the present invention, in the process of performing digital predistortion compensation processing on the energy normalized signal, a nonlinear processed signal is obtained by performing nonlinear processing on the signal, and a polynomial compensation coefficient is obtained according to the error signal and the nonlinear processed signal. The polynomial compensation coefficient can be obtained by a specific algorithm.
[0084] In an embodiment of the present invention, Figure 3 As shown, the signal after digital pre-distortion compensation Subtract the signal x(n) from the transmitting end to obtain the error signal e(n). The transmitting end signal is recorded as x(n), and the specific calculation method of the error signal can be implemented according to formula (1-5). e(n) represents the error signal calculated at the nth moment. Subtract the signal x(n) from the signal Subtraction is performed to obtain an error signal e(n).
[0085]
[0086] In an embodiment, the implementation process of the polynomial compensation coefficient update algorithm can be implemented based on an existing algorithm model. For example, equation (1-6) defines a polynomial compensation coefficient update method, where c can be written as m,k (n) is the polynomial compensation coefficient at the nth moment, μ is a fixed step size, and the symbol “*” represents conjugation. However, the updating method of the polynomial compensation coefficient defined in equation (1-6) has certain instability in the actual system model simulation. Through simulation, it is found that z defined in equation (1-3) m,k (n) has a large value variation range and it is difficult to select the update step size. Therefore, the embodiment of the present invention changes formula (1-6) into a polynomial compensation coefficient update method as defined in formula (1-7). This method is based on the normalized minimum mean square error criterion and can prevent the polynomial compensation coefficient from being too large and causing divergence, wherein m can be represented as a memory length index, k can be represented as a nonlinear order index, P(n) is the normalized energy of the polynomial structure at the nth moment, and the calculation method of P(n) is defined in formula (1-8). In the present invention, the memory length M=5 and the nonlinear order K=3 are selected.
[0087] In an embodiment, the implementation process of the polynomial compensation coefficient update algorithm can be implemented based on the algorithm model defined in formula (1-7). For example, the fixed step value can be recorded as μ, and the update step value of the polynomial compensation coefficient can be recorded as μ n , the polynomial compensation coefficient of the previous moment is recorded as c m,k (n-1), the updated value of the polynomial compensation coefficient at the current time n is recorded as c m,k (n). First, the nonlinear processing signal z m,k (n) Perform conjugate operation and convert the signal z obtained by conjugate operation into m,k * (n) and the error signal e(n) and the update step value μ of the polynomial compensation coefficient n Perform multiplication operation. And the output signal obtained after the multiplication operation of the three is combined with the polynomial compensation coefficient C at the previous moment m,k (n-1) is added to obtain the updated value c of the polynomial compensation coefficient at the current time n m,k (n). In the process of algorithm implementation, the polynomial compensation coefficient c at the current moment can be m,k (n) performing time delay processing, and performing addition calculation in the next step of updating the calculation process of the polynomial compensation coefficient. As mentioned above, there are at least fifteen nonlinear processing signals z m,k (n), so at least fifteen updated values of the polynomial compensation coefficients can be obtained.
[0088]
[0089]
[0090]
[0091] In the embodiment of the present invention, during the digital pre-distortion compensation process, the nonlinear processing signal z obtained by performing nonlinear operation on the output signal of the power amplifier is m,k (n) has a large value variation range at different times. If a constant step size is used to update the polynomial compensation coefficients during the implementation process, it is easy to cause system divergence. Figure 3 As shown, the update step length of the polynomial compensation coefficient can be normalized in the normalization processing module to obtain the update step length of the polynomial compensation coefficient. represents a fixed step size, and ε represents a constant. m,k (n) Energy statistics are performed to obtain the energy of fifteen nonlinear processed signals. The energy of the fifteen nonlinear processed signals is accumulated to obtain the energy sum which can be recorded as P(n). After passing the energy sum P(n) through the LUT, we can get the reciprocal of the sum of the energy sum and ε, which is recorded as ε represents a constant, which is used to prevent P(n) from being too small and its derivative value from being too large, thus causing the update step size μ n becomes very large, which brings about the instability of coefficient convergence. In this paper, ε=0.1 is selected. To save the division operation, the normalized energy lookup table LUT shown in Table 1-1 can be used to implement it. Multiplying with the fixed step value μ to obtain the normalized update step size μ of the polynomial compensation coefficient n , The fixed step value μ can be configured by the register. In this paper, μ=0.01 is selected. n During the updating process of the input polynomial compensation coefficient, the updating operation of the polynomial compensation coefficient is implemented. The implementation process of the embodiment of the present invention is simpler and the calculation result is more stable.
[0092] Table 1-1 is a normalized energy lookup table LUT according to an embodiment of the present invention.
[0093] Table 1-1
[0094]
[0095] Figure 5 4 is a flow chart of a calibration process of a pre-distortion calibration device according to an embodiment of the present invention.
[0096] The digital pre-distortion processing method of the embodiment of the present invention is as follows: Figure 5 As shown, the overall implementation process from the power amplifier 200 outputting the signal to the digital pre-distortion processing of the signal may include the following: for example, the power amplifier outputs a signal y(n) at time n, and after normalization processing, a signal y1(n)=y(n)*1 / G is obtained, and then the signal y1(n) is subjected to digital pre-distortion compensation processing, for example, after tapped delay line processing, nonlinear processing, multiplication processing and accumulation processing, a signal Then the digital pre-distortion compensation signal Subtract the signal x(n) from the transmitting end to obtain the error signal e(n). m,k(n) and the error signal e(n) are used to perform update calculation of the polynomial compensation coefficient, and the iterative update step size is also normalized, thereby ensuring the stability of the polynomial compensation coefficient during the digital pre-distortion estimation process. The specific calculation process can refer to the above-mentioned update calculation content for the polynomial compensation coefficient. The digital pre-distortion processing method of the embodiment of the present invention, in combination with the existing digital pre-distortion processing method, provides a calculation method based on a normalized minimum mean square error algorithm, selects a memory length of 5, a nonlinear order of 3, and can set the sampling clock to work at a frequency of 160Mhz, which can ensure signal stability while being simple to implement.
[0097] Figure 6 is a structural block diagram of a broadband system according to an embodiment of the present invention.
[0098] In an embodiment of the present invention, Figure 6 As shown, the broadband system 1000 may include at least a power amplifier 200 and a digital predistortion processing device 100. The digital predistortion processing device 100 may be implemented by hardware. For example, in the process of hardware implementation, a certain number of adders, multipliers, accumulators, and nonlinear processing units may be combined into a complete digital circuit to realize signal calculation and transmission. The digital predistortion processing device 100 may also be implemented by software. For example, after a series of program settings, the calculation and processing of the signal may be realized in the process of running the program.
[0099] The power amplifier 200 can be connected to the digital predistortion processing device 100, and can perform digital predistortion processing on the original signal in the digital predistortion processing device 100 before passing through the power amplifier 200. For example, before the original signal passes through the power amplifier 200, the digital signal predistortion processing is first performed in the digital predistortion processing device 100, and the signal after the digital signal predistortion processing can be passed through a DAC (Digital to Analog Converter) to convert a discrete signal in digital form into a continuous signal in analog form, and then the continuous signal in analog form is passed through the power amplifier 200 to obtain an amplified signal, and then the amplified signal is passed through an ADC (Analog to digital converter) to convert the continuous signal in analog form into a discrete signal in digital form. The present invention combines the existing digital predistortion processing method to provide a calculation method based on a normalized minimum mean square error algorithm, which is simple to implement while ensuring signal stability.
[0100] According to the non-temporary computer storage medium of the embodiment of the present invention, when the instructions in the storage medium are executed, the system can adopt the digital pre-distortion processing method of the embodiment of the present invention in the process of processing the signal, wherein the main program is stored internally in the non-temporary computer storage medium, and executable instructions can be issued through the running of the program. It has memory for information such as related processing signals, correlation coefficients, and operation methods, and responds quickly, which can effectively shorten the response time. When the computer program is executed, the digital pre-distortion processing method of the embodiment of the present invention is implemented, so that the original signal is digitally pre-distorted before passing through the power amplifier to eliminate the influence of the signal nonlinear characteristics caused by the signal passing through the power amplifier.
[0101] Figure 7 Schematic diagram of a simulation model structure of a digital pre-distortion system according to an embodiment of the present invention.
[0102] In the embodiment, in order to verify the processing effect of the digital pre-distortion processing device of the embodiment of the present invention on the output signal of the power amplifier, a simulation experiment is set. Figure 7 As shown, the simulation process can be implemented by the simulation model of the digital pre-distortion system. The simulation model can be implemented by hardware settings or by software to process the signal. The specific simulation process may include the following contents:
[0103] The simulation model can be divided into two stages: digital pre-distortion estimation and digital pre-distortion compensation. For example, in the digital pre-distortion estimation stage, the broadband signal generator can generate the original signal according to the corresponding broadband. For example, in the embodiment of the present invention, the original signal is taken as an OFDM digital broadband training signal, and the parameters are set to vht, mcs0, and the data packet is 8192 bytes. The OFDM digital broadband training signal is filtered to obtain a filtered signal and enters the multiplier. The filter is taken as an example of an 80->160 filter, indicating that the signal input frequency of the filter is 80MHz and the output frequency is 160MHz. The filtered signal can be multiplied with the digital gain in the multiplier. This step is mainly used to adjust the energy of the digital signal obtained after filtering by the 80->160 filter. In the digital pre-distortion estimation stage, the data selector can bypass the digital pre-distortion compensation module and directly send the digital signal adjusted by the multiplication operation into the filter. Taking the 160->320 filter as an example, it means that the signal input frequency of the filter is 160MHz and the output frequency is 320MHz. The filtered signal may be passed through a communication channel simulation model with memory, wherein the communication channel simulation model with memory is mainly composed of a FIR (Finite Impulse Response, finite unit impulse response) filter and a Saleh model, and is mainly used to implement the behavior of a simulated power amplifier. Then it passes through the transmitter I / Q (In-phase / Quadrature) unbalanced generation module. The transmitter I / Q unbalanced generation module is equivalently generated in the digital domain and is mainly used to simulate the IQ imbalance generated by the RF circuit. The digital signal passing through the transmitter I / Q unbalanced generation module can enter the multiplier and complete the multiplication calculation process with the RF gain in the multiplier. This process can adjust the output energy of the digital signal to a suitable range through the equivalent RF gain. For example, for broadband OFDM signals, considering factors such as the peak-to-average ratio, the output signal energy can be adjusted to about -12dB. Then the output signal of the multiplier is input into the adder, and a DC offset is added to simulate the DC signal generated by the RF circuit. In the present invention, the maximum offset value of the analog ADC can be 20mv, where the full amplitude range of the analog ADC can be + / -600mv. Then the output signal of the adder is input into the Gaussian white noise module, and then passes through the filter and the receiver I / Q imbalance correction module, where the filter here can be a 320->160 filter, indicating that the signal input frequency of the filter is 320MHz and the output frequency is 160MHz. The receiver I / Q imbalance correction module is mainly used to compensate for the IQ imbalance generated by the RF circuit. Finally, the processed digital signal is sent to the digital pre-distortion estimation module, and the other input of the digital pre-distortion estimation module comes from the output of the data selector through the output signal of the FIFO delay module.After the digital predistortion estimation module completes the calculation of the coefficients, in the digital predistortion compensation stage, the coefficients calculated in the digital predistortion estimation module can be copied to the digital predistortion compensation module, and the digital predistortion estimation module is closed at this time.
[0104] In the digital predistortion compensation stage, the OFDM digital broadband training signal is predistorted according to the estimated coefficients, and then sent to the communication channel simulation model with memory, and finally the OFDM digital broadband training signal after digital predistortion processing and power amplifier amplification is output. We can measure whether the estimated coefficients are correct by calculating the signal-to-noise ratio module, and we can also measure the effectiveness of the digital predistortion system by comparing the original OFDM digital broadband training signal with the OFDM digital broadband training signal after digital predistortion processing and power amplifier amplification, and the original OFDM digital broadband training signal with the OFDM digital broadband training signal after power amplifier amplification. The more the signal-to-noise ratio is improved, the better the signal processing effect is.
[0105] Table 1-2 is a comparison table of the signal-to-noise ratio of the original signal and the signal amplified by the power amplifier at different powers according to an embodiment of the present invention, as well as the signal-to-noise ratio of the original signal and the signal amplified by the power amplifier after the digital predistortion processing. As shown in Table 1-2, the signal-to-noise ratios of the signals obtained by turning on the digital predistortion compensation module and turning off the digital predistortion compensation module under 20M, 40M and 80M broadband conditions, and passing the original signal through a communication channel simulation model with memory. In the simulation process of the present invention, it can be seen from Table 1-2 that under the conditions of 20M, 40M and 80M broadbands and digital gains of -8dB, -6dB, -4dB, -2dB, and 0dB, respectively, the signal-to-noise ratio of the digital signal obtained by turning on the digital predistortion compensation module is greater than the signal-to-noise ratio of the digital signal obtained by turning off the digital predistortion compensation module. For example, the original OFDM digital broadband training signal can be set to 80M, vht, mcs0, the data packet can be 8192 bytes, the digital gain can be -8db, the power amplifier input power can be -16.6db, the power amplifier output power can be -11.2db, the signal-to-noise ratio of the original signal and the signal after the original signal is amplified by the power amplifier can be 13.2dB, and the signal-to-noise ratio of the original signal and the signal after the original signal is processed by digital predistortion and amplified by the power amplifier can be 44.1dB. It can be seen that when the digital gain, input power and output power are the same, the signal-to-noise ratio of the original signal and the signal after the original signal is processed by digital predistortion and amplified by the power amplifier is compared with the signal-to-noise ratio of the original signal and the signal after the original signal is amplified by the power amplifier. The former is significantly improved, and it can be seen from Table 1-2 that the smaller the energy of the original signal is, the more the signal-to-noise ratio is improved after the digital predistortion processing. As the energy of the original signal increases, the improvement of the signal-to-noise ratio after the digital predistortion processing becomes smaller, mainly because the signal has entered the saturation area of the power amplifier. The simulation results in Table 1-2 prove that the digital pre-distortion system of the embodiment of the present invention is effective.
[0106] Table 1-2
[0107]
[0108] Figure 8 It is a schematic diagram of the structure of a Saleh model simulating a memoryless power amplifier according to an embodiment of the present invention.
[0109] Table 1-3 is a table of coefficients and Saleh model parameters of a third-order FIR filter according to an embodiment of the present invention.
[0110] In the present invention, the simulation experiment of the digital signal is based on the communication channel simulation model with memory, which is mainly composed of FIR filter and Saleh model, and is mainly used to realize the behavior of the analog power amplifier. The FIR filter can be set as a third-order filter, such as Figure 8 As shown, when the input signal μ(n) passes through the Saleh module, the behavior of the analog power amplifier output signal y(n) can be realized through the algorithm. In the specific implementation process, the specific algorithm of formula (3-1) can be realized. For example, as shown in formula (3-1), it is used to simulate the input of the signal μ(n) into the Saleh module. The amplitude and phase calculation are first performed to obtain the amplitude and phase angle of the μ(n) signal, and then the amplitude of the μ(n) signal is obtained through AM / PM. The input is added to the adder and added to the phase angle of the μ(n) signal to obtain the phase angle of the output signal y(n). Then the modulus of the signal μ(n) is passed through AM / AM to obtain the modulus of the output signal y(n). Finally, the phase angle and modulus of the output signal y(n) are calculated using the Euler formula to obtain the output signal y(n).
[0111] Among them, the coefficients of the FIR filter and the parameters in formula (3-1) can be defined according to the table of coefficients of the FIR filter and the parameters of the Saleh model in Table 1-3. As can be seen from Table 1-3, the three coefficients of the FIR filter can be set to Coef0 = 0.7692, Coef1 = 0.1538, and Coef2 = 0.0769. The parameters in formula (3-1) include α a ,β a ,α p ,β p , you can set α a =2.1587, β a =1.1517,α p =4.003, β p =9.104.
[0112]
[0113] Table 1-3
[0114]
[0115] in, Figure 9a-9c : is a waveform diagram of the error signal of the digital pre-distortion estimation of the embodiment of the present invention, wherein the relationship between the error signal amplitude and the sampling time is as follows: Figure 9a As shown in the figure, the relationship between the amplitude of the error signal after filtering and the sampling time is as follows Figure 9b As shown, the relationship between the error signal amplitude dB value after filtering and the sampling time is as follows Fig.9cAs shown, the sampling time unit is μs.
[0116] In an embodiment, the simulated digital signal image and the original signal image may be compared, so as to evaluate the effect of digital pre-distortion of the digital signal.
[0117] like Fig.10a is a waveform diagram of the time domain response of the original signal of an embodiment of the present invention, Fig.10b : is a waveform diagram of the time domain response of the output signal after the original signal is processed by digital pre-distortion and amplified by the power amplifier according to an embodiment of the present invention. Fig.10a As shown in , the time domain response amplitude waveform of the original signal contains real and imaginary components. Fig.10b As shown, the time domain response amplitude waveform of the output signal obtained after the original signal is processed by digital predistortion and amplified by the power amplifier includes a real component and an imaginary component.
[0118] Depend on Fig.10a and Fig.10b By comparison, it can be found that the time domain response of the output signal after the original signal is processed by digital predistortion and amplified by the power amplifier basically coincides with the time domain response curve of the original signal, which can illustrate the correctness of the digital predistortion processing of the embodiment of the present invention.
[0119] like Fig.11a is a waveform diagram of the frequency domain response of the original signal of an embodiment of the present invention, Fig.11b The waveform diagram of the frequency domain response of the original signal after being amplified by the power amplifier and the waveform diagram of the frequency domain response of the original signal after being processed by digital predistortion and amplified by the power amplifier in one embodiment of the present invention. Fig.11b As shown in the figure, curve 1 is the waveform of the frequency domain response of the original signal after being amplified by the power amplifier, and curve 2 is the waveform of the frequency domain response of the original signal after being processed by digital predistortion and amplified by the power amplifier. It can be seen from curve 1 that the frequency domain response of the original signal after being amplified by the power amplifier is relatively Fig.11a Compared with the waveform of the frequency domain response of the original signal in the , the difference is quite large, which is caused by the nonlinearity of the power amplifier. From curve 2, it can be seen that the waveform of the frequency domain response of the original signal after digital pre-distortion processing and power amplifier amplification is different from that of the original signal after digital pre-distortion processing and power amplifier amplification. Fig.11a The waveform curves are basically the same as the waveform of the frequency domain response of the original signal in the embodiment of the present invention. According to the digital predistortion processing method of the embodiment of the present invention, the signal is digitally predistorted, so that the nonlinear characteristics of the signal after the digital predistortion processing can be eliminated after passing through the power amplifier, which illustrates the correctness of the digital predistortion processing.
[0120] like Fig.12ais a waveform diagram of the amplitude of an original signal and the amplitude of a signal after the original signal is amplified by a power amplifier according to an embodiment of the present invention, Figure 12b The waveform diagram is a diagram showing the amplitude of the original signal and the amplitude of the signal after the original signal is processed by digital pre-distortion and amplified by a power amplifier according to an embodiment of the present invention, wherein the sampling clock operates at a frequency of 160Mhz. Fig.12a The curves in include the amplitude curve of the original signal and the amplitude curve of the signal after the original signal is amplified by the power amplifier. As can be seen from the figure, the amplitude of the original signal and the amplitude curve of the signal after the original signal is amplified by the power amplifier do not overlap much, and there is a big difference. Figure 12b The curves in the figure include the amplitude curve of the original signal and the amplitude curve of the original signal after being processed by digital predistortion and amplified by the power amplifier. As can be seen from the figure, the amplitude of the original signal is basically coincident with the amplitude curve of the original signal after being processed by digital predistortion and amplified by the power amplifier. According to the digital predistortion processing method of the embodiment of the present invention, the signal is digitally predistorted, so that the nonlinear characteristics of the signal after the digital predistortion processing can be eliminated after passing through the power amplifier, which illustrates the correctness of the digital predistortion processing.
[0121] like Fig.13a is a waveform diagram of the phase of an original signal and the phase of a signal after the original signal is amplified by a power amplifier according to an embodiment of the present invention, Fig.13b It is a waveform diagram of the phase of the original signal and the phase of the signal after the original signal is processed by digital pre-distortion and amplified by the power amplifier according to an embodiment of the present invention, wherein the sampling clock operates at a frequency of 160Mhz. Fig.13a The curves in include the phase curve of the original signal and the phase curve of the signal after the original signal is amplified by the power amplifier. As can be seen from the figure, the phase of the original signal and the phase curve of the signal after the original signal is amplified by the power amplifier do not overlap much, and the difference is relatively large. Fig.13b The curves in the figure include the phase curve of the original signal and the phase curve of the original signal after being processed by digital predistortion and amplified by the power amplifier. As can be seen from the figure, the phase of the original signal is basically coincident with the phase curve of the original signal after being processed by digital predistortion and amplified by the power amplifier. According to the digital predistortion processing method of the embodiment of the present invention, the signal is digitally predistorted, so that the nonlinear characteristics of the signal after the digital predistortion processing can be eliminated after passing through the power amplifier, which illustrates the correctness of the digital predistortion processing.
[0122] In general, the digital predistortion processing device and broadband system of the embodiment of the present invention are applicable to both memoryless and memory-based power amplifier models. In combination with the existing digital predistortion processing method, a normalized minimum mean square error update calculation method is provided to update the polynomial compensation coefficient. The memory length is selected as 5, the nonlinear order is 3, and the sampling clock operates at a frequency of 160Mhz, which is simple to implement. In addition, the output signal of the power amplifier is energy normalized, and the iterative update step size is also normalized, thereby ensuring the stability of the polynomial compensation coefficient during digital predistortion processing.
[0123] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. Or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0124] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example.
[0125] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.
Claims
1. A digital pre-distortion processing method, It is characterized in that For a WIFI broadband system, the digital predistortion processing method includes: Obtain an output signal of a power amplifier and a transmitting end signal in a WIFI broadband system, perform energy normalization on the output signal of the power amplifier based on the transmitting end signal, and obtain an energy normalized signal; Performing digital predistortion compensation on the energy-normalized signal according to the polynomial compensation coefficient to obtain a digital predistortion compensated signal, wherein the digital predistortion compensation includes nonlinear processing; Subtracting the digital predistortion compensated signal from the transmitting end signal to obtain an error signal; Acquire a nonlinear processing signal in digital predistortion compensation, and obtain the polynomial compensation coefficient according to the error signal and the nonlinear processing signal; Feeding back the polynomial compensation coefficients to the digital pre-distortion compensation; The method of performing digital pre-distortion compensation on the energy normalized signal according to the polynomial compensation coefficient to obtain the digital pre-distortion compensated signal includes: Performing time delay processing on the energy-normalized signal to obtain a third time-delayed signal; Performing nonlinear processing on the third time-delayed signal to obtain a nonlinear processed signal; Acquiring the polynomial compensation coefficient, performing a multiplication operation on the nonlinear processing signal and the polynomial compensation coefficient to obtain a signal after the multiplication operation; The signals after the multiplication operation are accumulated to obtain the signals after digital predistortion compensation.
2. The digital predistortion processing method according to claim 1, It is characterized in that The output signal of the power amplifier is energy normalized based on the transmitting end signal to obtain an energy normalized signal, including: Obtaining an absolute value of an output signal of the power amplifier, adding the absolute value of the output signal of the power amplifier to a first time-delayed signal to obtain a first operation value, wherein the first operation value is subjected to a time-delay process to obtain the first time-delayed signal; Obtaining an absolute value of the transmitting end signal, adding the absolute value of the transmitting end signal to a second time-delayed signal to obtain a second operation value, wherein the second operation value is subjected to a time-delay process to obtain the second time-delayed signal; Performing a division operation on the first operation value and the second operation value to obtain a third operation value; The third operation value is multiplied by the output signal of the power amplifier to obtain an energy-normalized signal.
3. The digital predistortion processing method according to claim 1, It is characterized in that Performing time delay processing on the energy normalized signal to obtain a third time delay signal includes: The energy-normalized signal is subjected to multi-level time delay processing in sequence to obtain a multi-level third time delay signal.
4. The digital predistortion processing method according to claim 3, It is characterized in that The energy-normalized signal is subjected to multi-level delay processing in sequence to obtain a multi-level third delay signal, including: The energy-normalized signal is used as the third time-delay signal of level 0; Performing a first-level delay processing on the level 0 third time-delayed signal to obtain a first-level third time-delayed signal; Performing a second-level time delay processing on the first-level third time delay signal to obtain a second-level third time delay signal; Performing a third-level time delay processing on the second-level third time delayed signal to obtain a third-level third time delayed signal; Perform a fourth-level delay processing on the third-level third time-delayed signal to obtain a fourth-level third time-delayed signal.
5. The digital predistortion processing method according to claim 4, It is characterized in that Performing nonlinear processing on the third time-delayed signal to obtain a nonlinear processed signal includes: Performing nonlinear processing on the third time-delayed signal of level 0, and obtaining three first nonlinear processing signals; Performing nonlinear processing on the first-stage third time-delayed signal, and obtaining three second nonlinear processed signals; Performing nonlinear processing on the second-stage third time-delayed signal, and obtaining three third nonlinear processed signals; Performing nonlinear processing on the third-stage third time-delayed signal to obtain three fourth nonlinear processed signals; The fourth-stage third time-delayed signal is subjected to nonlinear processing to obtain three fifth nonlinearly processed signals.
6. The digital predistortion processing method according to claim 5, It is characterized in that Acquiring the polynomial compensation coefficient, multiplying the nonlinear processing signal by the polynomial compensation coefficient to obtain a signal after the multiplication operation, including: The polynomial compensation coefficients include fifteen compensation coefficients; Based on the fifteen compensation coefficients, three of the first nonlinear processed signals, three of the second nonlinear processed signals, three of the third nonlinear processed signals, three of the fourth nonlinear processed signals and three of the fifth nonlinear processed signals are multiplied respectively to obtain signals after the multiplication operation.
7. The digital predistortion processing method according to claim 6, It is characterized in that The multiplication signal is accumulated to obtain a digital predistortion compensated signal, including: Performing an accumulation operation on the signals after the multiplication operation of the three first nonlinear processed signals to obtain a first accumulated value; Performing an accumulation operation on the signals after the multiplication operation of the three second nonlinear processed signals to obtain a second accumulated value; Performing an accumulation operation on the signals after the multiplication operation of the three third nonlinear processed signals to obtain a third accumulated value; Performing an accumulation operation on the signals after the multiplication operation of the three fourth nonlinear processed signals to obtain a fourth accumulated value; Performing an accumulation operation on the signals after the multiplication operation of the three fifth nonlinear processed signals to obtain a fifth accumulated value; The first accumulated value, the second accumulated value, the third accumulated value, the fourth accumulated value and the fifth accumulated value are accumulated to obtain a total accumulated value as the signal after digital pre-distortion compensation.
8. The digital predistortion processing method according to claim 1, It is characterized in that Acquiring a nonlinear processing signal in digital predistortion compensation, and obtaining the polynomial compensation coefficient according to the error signal and the nonlinear processing signal, comprising: Acquire multi-channel nonlinear processing signals in digital pre-distortion compensation; Obtain the conjugate signal of each nonlinearly processed signal; Perform multiplication operation on the conjugate signal of each nonlinear processing signal, the error signal and the coefficient update step length to obtain a plurality of product values; The polynomial compensation coefficient at a previous moment is obtained, and a plurality of product values and the polynomial compensation coefficient at the previous moment are respectively added to obtain the polynomial compensation coefficient at a current moment.
9. The digital predistortion processing method according to claim 8, It is characterized in that The digital pre-distortion processing method further includes: normalizing the coefficient update step size.
10. The digital predistortion processing method according to claim 9, It is characterized in that The coefficient update step size is normalized, including: Acquire multi-channel nonlinear processing signals; Count the energy of each nonlinear processing signal; Calculate the energy sum of multi-channel nonlinear processing signals; Performing LUT processing on the energy sum to obtain the reciprocal of the energy sum; The reciprocal of the energy sum is multiplied by the fixed step value to obtain a normalized coefficient update step.
11. The digital predistortion processing method according to any one of claims 1 to 10, It is characterized in that The polynomial compensation coefficients satisfy the following normalized minimum mean square error convergence formula: ,in, , in, represents the polynomial compensation coefficient at the nth moment, represents the polynomial compensation coefficient of the previous moment, μ represents the fixed step size, e(n) represents the error signal, It represents the signal after nonlinear processing, the symbol "*" represents conjugation, m represents the memory length index, and k represents the nonlinear order index. Represents the normalized energy of the polynomial structure at the nth time.
12. The digital predistortion processing method according to claim 11, It is characterized in that M=5,K=3.
13. A WIFI broadband system, It is characterized in that include: Power amplifier; and A digital predistortion processing device, wherein the digital predistortion processing device is connected to the power amplifier and is used to implement the digital predistortion processing method according to any one of claims 1 to 12.
14. A non-transitory computer storage medium having stored thereon a computer program, It is characterized in that When the computer program is executed, the digital pre-distortion processing method according to any one of claims 1 to 12 is implemented.
