A digital predistortion processing system and method
By using the coordinated operation of predistorter, power amplifier, 1/G module and learner in the digital predistortion processing system, the problems of high computational complexity and slow convergence speed are solved, and the coefficient interval value is quickly calculated and the calculation is reduced. The effect of time.
Patent Information
- Application Number
- CN202111259753.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-28
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-10-28
AI Technical Summary
In the existing digital predistortion processing system, there are problems such as high computational complexity, long calculation time and slow convergence speed.
Through coordinated operation of predistorter, power amplifier, 1/G module, computing module and learner, the coefficient interval value is quickly calculated, and the direct and indirect learning structures are used for rough adjustment and fine adjustment to optimize the calculation process.
The convergence speed is improved, the calculation time is reduced, the coefficient interval value is quickly calculated, the regulation convenience is improved, and the digital predistortion effect is achieved.
Smart Images

Figure CN114050795B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power amplifiers, and in particular to a digital predistortion processing system and method. Background Art
[0002] Due to the nonlinear characteristics of the power amplifier, power amplifier distortion will occur. In order to offset the distortion generated by the power amplifier, a pre-distortion function is added to the digital signal. At present, the digital pre-distortion processing model in the market mainly adopts the indirect learning type. This learning structure adds a first learner (coefficient estimation) with the same structure as the pre-distorter in the feedback path. This learning structure is behind the power amplifier, and the coefficient estimation value is solved from the output of the power amplifier. After learning, the coefficient estimation is copied to the pre-distorter, but it has the characteristics of computational complexity, long calculation time, and slow convergence speed. Summary of the invention
[0003] The present invention aims to solve at least one of the above technical problems in the prior art to a certain extent. To this end, one object of the present invention is to provide a digital predistortion processing system and method that improves convergence speed and reduces calculation time.
[0004] The technical solution of the present invention to solve the above technical problem is as follows: A digital predistortion processing system, comprising:
[0005] A predistorter, the predistorter is used to collect a signal x(n), perform predistortion processing on the signal x(n), and generate a signal z(n); the input end of the predistorter is electrically connected to the output end of the first learner, and the predistorter is also used to perform predistortion processing on the signal x(n) according to a coefficient interval value;
[0006] A power amplifier, wherein a signal input terminal of the power amplifier is electrically connected to an output terminal of the predistorter, and the power amplifier is used to amplify the signal z(n) to generate a signal y(n);
[0007] A 1 / G module, wherein an input end of the 1 / G module is electrically connected to an output end of the power amplifier, and the 1 / G module is used to reduce the signal amplitude of the signal y(n) by G times to generate a signal y(n) / G;
[0008] An operation module, wherein an input end of the operation module is electrically connected to an output end of the 1 / G module, and an input end of the operation module is also electrically connected to an input end of the predistorter, and the operation module is used to operate a signal x(n) and a signal y(n) / G to obtain a signal e(n);
[0009] A first learner, wherein the input end of the first learner is electrically connected to the output end of the operation module, and the first learner is used to estimate the fitting coefficient according to the signal e(n), determine the coefficient estimation interval, and generate the coefficient interval value.
[0010] The beneficial effects of the present invention are: through the coordinated operation of the predistorter, the power amplifier, the 1 / G module, the operation module and the first learner, the coefficient interval value can be quickly calculated, the convergence speed can be improved, the calculation time can be reduced, and the expected effect can be achieved.
[0011] Based on the above technical solution, the present invention can also be improved as follows.
[0012] Furthermore, a switch C is provided on the connection line between the 1 / G module and the operation module; and a switch A is provided on the connection line between the operation module and the input end of the predistorter.
[0013] The beneficial effect of adopting the above further scheme is that when switch C and switch A are turned on, the predistorter, power amplifier (2), 1 / G module, operation module and first learner can form a direct learning structure, and the coefficient interval value can be roughly adjusted, thereby improving the convenience of regulation.
[0014] Furthermore, it also includes:
[0015] A second learner, wherein the input end of the second learner is electrically connected to the output end of the 1 / G module, and the output end of the second learner is electrically connected to the input end of the operation module; the second learner is used to process the signal y(n) / G to generate a signal z`(n);
[0016] The input end of the operation module is also electrically connected to the output end of the predistorter. The operation module is also used to collect the signal z(n), operate the signal z(n) and the signal z`(n), and obtain the signal e(n).
[0017] The beneficial effect of adopting the above further solution is: fine adjustment of the coefficient interval value of the predistorter (1) and the rough adjustment of the coefficient interval value require at least two signal processings to achieve the digital predistortion effect, improve the convergence speed and reduce the calculation time.
[0018] Furthermore, a switch D is provided on the connection line between the 1 / G module and the second learner; and a switch B is provided on the connection line between the operation module and the output end of the pre-distorter.
[0019] The beneficial effect of adopting the above further scheme is that when switch B and switch D are turned on, the pre-distorter, power amplifier, 1 / G module, operation module, first learner and second learner can form an indirect learning structure, and the coefficient interval value can be fine-tuned to improve the convenience of regulation.
[0020] In order to solve the above technical problems, the present invention also provides a digital predistortion processing method, comprising the following coarse adjustment steps:
[0021] The predistorter collects a signal x(n), performs predistortion processing on the signal x(n), and generates a signal z(n); the power amplifier amplifies the signal z(n) and generates a signal y(n); the 1 / G module reduces the signal amplitude of the signal y(n) by G times and generates a signal y(n) / G; the operation module operates the signal x(n) and the signal y(n) / G to obtain a signal e(n); the first learner estimates the fitting coefficient according to the signal e(n), determines the coefficient estimation interval, and generates a coefficient interval value; the predistorter performs predistortion processing on the signal x(n) according to the coefficient interval value.
[0022] The beneficial effects of the present invention are: through the coordinated operation of the predistorter, the power amplifier, the 1 / G module, the operation module and the first learner, the coefficient interval value can be quickly calculated, the convergence speed can be improved, the calculation time can be reduced, and the expected effect can be achieved.
[0023] Based on the above technical solution, the present invention can also be improved as follows.
[0024] Furthermore, the following fine-tuning steps are included:
[0025] The second learner processes the signal y(n) / G to generate a signal z`(n); the operation module collects the signal z(n), operates the signal z(n) and the signal z`(n) to obtain the signal e(n).
[0026] The beneficial effect of adopting the above further solution is: fine adjustment of the coefficient interval value of the predistorter (1) and the rough adjustment of the coefficient interval value require at least two signal processings to achieve the digital predistortion effect, improve the convergence speed and reduce the calculation time.
[0027] Further, the method further comprises the following steps:
[0028] A threshold of the signal e(n) is set. When the signal e(n) obtained by calculating the signal z(n) and the signal z`(n) is greater than the set threshold, the fine-tuning step is repeated until the signal e(n) is less than or equal to the set threshold.
[0029] The beneficial effect of adopting the above further solution is that the digital pre-distortion effect can be achieved by obtaining the signal e(n) and performing multiple fine adjustments on the coefficient interval value.
[0030] Furthermore, the following steps are included: when the fine-tuning step is repeated more than a set number of times, switching to the coarse-tuning step, adjusting the coefficient estimation interval, and then performing the fine-tuning step processing until the signal e(n) is less than or equal to the set threshold.
[0031] The beneficial effect of adopting the above further solution is that the digital pre-distortion effect can be quickly achieved through the coordination of fine adjustment and coarse adjustment. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 A module block diagram of a digital predistortion processing system of the present invention;
[0033] Figure 2 A first flow chart of a digital predistortion processing method of the present invention;
[0034] Figure 3 This is a second flow chart of a digital predistortion processing method of the present invention.
[0035] In the accompanying drawings, the components represented by the reference numerals are listed as follows:
[0036] 1. Pre-distorter, 2. Power amplifier, 3. 1 / G module, 4. Operation module, 5. First learner, 6. Second learner. DETAILED DESCRIPTION
[0037] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.
[0038] Embodiment 1:
[0039] like Figure 1 The figure is a block diagram of a digital predistortion processing system of the present invention. A digital predistortion processing system of the present invention comprises a predistorter 1, a power amplifier 2, a 1 / G module 3, a calculation module 4 and a first learner 5;
[0040] Specifically, the predistorter 1 is used to collect a signal x(n), perform predistortion processing on the signal x(n), and generate a signal z(n); the signal input end of the power amplifier 2 is electrically connected to the output end of the predistorter 1, and the power amplifier 2 is used to amplify the signal z(n) to generate a signal y(n); the input end of the 1 / G module 3 is electrically connected to the output end of the power amplifier 2, and the 1 / G module 3 is used to reduce the signal amplitude of the signal y(n) by G times to generate a signal y(n) / G; the input end of the operation module 4 is electrically connected to the output end of the 1 / G module 3 The input end of the operation module 4 is also electrically connected to the input end of the predistorter 1, and the operation module 4 is used to operate the signal x(n) and the signal y(n) / G to obtain the signal e(n); the input end of the first learner 5 is electrically connected to the output end of the operation module 4, and the first learner 5 is used to estimate the fitting coefficient according to the signal e(n), determine the coefficient estimation interval, and generate the coefficient interval value; the input end of the predistorter 1 is electrically connected to the output end of the first learner 5, and the predistorter 1 is also used to perform predistortion processing on the signal x(n) according to the coefficient interval value;.
[0041] In actual application, the signal x(n) is the signal before pre-distortion, G is the power amplifier gain, and the collected feedback signal y(n) is processed by the 1 / G module 3 to reduce the amplitude of the signal by G times; in addition, the first learner 5 adjusts the delay, gain and phase of the data according to the collected signal e(n), calculates the linear term coefficient and the power amplifier intermodulation term coefficient of the power amplifier 2; calculates the fitting error, that is, the difference between the fitted power amplifier 2 data and the actual data, and determines the coefficient or coefficient range.
[0042] In this embodiment, the predistorter 1, the power amplifier 2, the 1 / G module 3, the operation module 4 and the first learner 5 form a direct learning structure, the predistorter 1 is placed in front of the power amplifier 2, the predistorter 1 directly obtains the inverse of the power amplifier 2, so that the input of the predistorter and the output of the power amplifier are in a linear relationship; the operation module 4 obtains the signal e(n)=x(n)-y(n) / G, and continuously updates the predistorter 1. When e(n) approaches 0, the output of the power amplifier 2 is a linear amplification of the input of the predistorter 1, and the coefficients simulated by the ideal predistorter 1 can be obtained. Through the coordinated operation of the predistorter 1, the power amplifier 2, the 1 / G module 3, the operation module 4 and the first learner 5, the coefficient interval value can be quickly calculated, the convergence speed can be improved, the calculation time can be reduced, and the expected effect can be achieved.
[0043] In the above embodiment, a switch C is provided on the connection line between the 1 / G module 3 and the operation module 4 ; a switch A is provided on the connection line between the operation module 4 and the input end of the predistorter 1 .
[0044] In this embodiment, the conduction and disconnection of the line between the 1 / G module 3 and the operation module 4 are controlled by the switch C, and the conduction and disconnection of the line between the operation module 4 and the predistorter 1 are controlled by the switch A; Figure 1 As shown, by disconnecting switch B and switch D and connecting switch C and switch A, the predistorter 1, power amplifier 2, 1 / G module 3, operation module 4 and first learner 5 can form a direct learning structure, and the coefficient interval value can be roughly adjusted to improve the convenience of regulation.
[0045] To match the above digital predistortion processing system, the present invention also provides a digital predistortion processing method, such as Figure 2 FIG. 1 is a first flow chart of a digital predistortion processing method of the present invention. A digital predistortion processing method of the present invention includes the following coarse adjustment steps:
[0046] S1. The predistorter 1 collects the signal x(n), performs predistortion processing on the signal x(n), and generates a signal z(n); the power amplifier 2 amplifies the signal z(n) and generates a signal y(n); the 1 / G module 3 reduces the amplitude of the signal y(n) by G times and generates a signal y(n) / G;
[0047] S2. The operation module 4 operates the signal x(n) and the signal y(n) / G to obtain the signal e(n); the first learner 5 estimates the fitting coefficient according to the signal e(n), determines the coefficient estimation interval, and generates the coefficient interval value; the predistorter 1 performs predistortion processing on the signal x(n) according to the coefficient interval value.
[0048] This embodiment corresponds to an embodiment of a digital pre-distortion processing system, which will not be described in detail.
[0049] Embodiment 2:
[0050] like Figure 1 , which is a block diagram of a digital predistortion processing system of the present invention. The difference between this embodiment and embodiment 1 is that: it also includes: a second learner 6, the input end of the second learner 6 is electrically connected to the output end of the 1 / G module 3, and the output end of the second learner 6 is electrically connected to the input end of the operation module 4; the second learner 6 is used to process the signal y(n) / G to generate a signal z`(n);
[0051] The input end of the operation module 4 is also electrically connected to the output end of the predistorter 1. The operation module 4 is also used to collect the signal z(n), operate the signal z(n) and the signal z`(n), and obtain the signal e(n).
[0052] In this embodiment, the predistorter 1, the power amplifier 2, the 1 / G module 3, the operation module 4, the first learner 5 and the second learner 6 constitute an indirect learning structure. A second learner 6 having the same structure as the predistorter is added to the feedback path. The second learner 6 is behind the power amplifier 2 and solves the coefficient interval value from the output of the power amplifier 2. After learning, the coefficient interval value is transmitted to the predistorter 1, so that the input of the predistorter 1 and the output of the power amplifier 2 are in a linear relationship.
[0053] The feedback channel takes y(n) / G as its input and z'(n) as its output. The signal e(n) = z(n) – z'(n) is continuously updated to make the signal e(n) approach 0. Ideally, we want y(n) = G*x(n), so that the output of predistorter 1 is z(n) = z'(n), and the signal e(n) = 0.
[0054] After obtaining y(n) and x(n), the signal z'(n) is directly calculated through this indirect learning structure, that is, the coefficient interval value of the pre-distorter 1 is directly obtained, which improves the convergence speed, reduces the calculation time, and achieves the expected effect.
[0055] In the above embodiment, a switch D is provided on the connection line between the 1 / G module 3 and the second learner 6 ; a switch B is provided on the connection line between the operation module 4 and the output end of the predistorter 1 .
[0056] In this embodiment, the conduction and disconnection of the line between the 1 / G module 3 and the second learner 6 are controlled by the switch D, and the conduction and disconnection of the line between the operation module 4 and the predistorter 1 are controlled by the switch B; Figure 1 As shown, by disconnecting switch C and switch A and connecting switch B and switch D, the predistorter 1, power amplifier 2, 1 / G module 3, operation module 4, first learner 5 and second learner 6 can form an indirect learning structure, and the coefficient interval value can be fine-tuned to improve the convenience of regulation.
[0057] To match the above digital predistortion processing system, the present invention also provides a digital predistortion processing system, such as Figure 3 FIG. 1 is a second flow chart of the digital predistortion processing method of the present invention. A digital predistortion processing method of the present invention further includes the following fine adjustment steps:
[0058] S1. Predistorter 1 collects signal x(n), performs predistortion processing on signal x(n) to generate signal z(n); power amplifier 2 amplifies signal z(n) to generate signal y(n); 1 / G module 3 reduces the amplitude of signal y(n) by G times to generate signal y(n) / G;
[0059] S2. The second learner 6 processes the signal y(n) / G to generate a signal z`(n); the operation module 4 collects the signal z(n), operates the signal z(n) and the signal z`(n), and obtains the signal e(n).
[0060] This embodiment corresponds to an embodiment of a digital pre-distortion processing system, which will not be described in detail.
[0061] The above embodiment also includes the following steps: setting a threshold of the signal e(n); when the signal z(n) and the signal z`(n) are operated to obtain a signal e(n) greater than the set threshold, repeating the fine-tuning step until the signal e(n) is less than or equal to the set threshold.
[0062] In this embodiment, the digital pre-distortion effect can be achieved by obtaining the signal e(n) and performing multiple fine adjustments on the coefficient interval value.
[0063] The above embodiment also includes the following steps: when the fine-tuning step is repeated more than a set number of times, switching to the coarse-tuning step, adjusting the coefficient estimation interval, and then performing the fine-tuning step processing until the signal e(n) is less than or equal to the set threshold.
[0064] This embodiment can achieve a digital predistortion effect through the coordination of fine adjustment and coarse adjustment.
[0065] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A digital predistortion processing system, It is characterized in that include: A predistorter, the predistorter is used to collect a signal x(n), perform predistortion processing on the signal x(n), and generate a signal z(n); the input end of the predistorter is electrically connected to the output end of the first learner, and the predistorter is also used to perform predistortion processing on the signal x(n) according to the coefficient interval value; A power amplifier, wherein a signal input terminal of the power amplifier is electrically connected to an output terminal of the predistorter, and the power amplifier is used to amplify the signal z(n) to generate a signal y(n); A 1 / G module, wherein an input end of the 1 / G module is electrically connected to an output end of the power amplifier, and the 1 / G module is used to reduce the signal amplitude of the signal y(n) by G times to generate a signal y(n) / G; An operation module, wherein an input end of the operation module is electrically connected to an output end of the 1 / G module, and an input end of the operation module is also electrically connected to an input end of the predistorter, and the operation module is used to operate a signal x(n) and a signal y(n) / G to obtain a signal e(n); A first learner, wherein an input end of the first learner is electrically connected to an output end of the operation module, and the first learner is used to estimate a fitting coefficient according to the signal e(n), determine a coefficient estimation interval, and generate a coefficient interval value; A switch C is provided on the connection line between the I / G module and the operation module; A switch A is provided on the connection line between the operation module and the input end of the predistorter; Also includes: A second learner, wherein an input end of the second learner is electrically connected to an output end of the 1 / G module, and an output end of the second learner is electrically connected to an input end of the operation module; The second learner is used to process the signal y(n) / G to generate a signal z`(n); The input end of the operation module is also electrically connected to the output end of the predistorter, and the operation module is also used to collect the signal z(n), perform operation on the signal z(n) and the signal z`(n), and obtain the signal e(n); A switch D is provided on the connection line between the 1 / G module and the second learner; a switch B is provided on the connection line between the operation module and the output end of the pre-distorter.
2. A digital predistortion processing method according to claim 1, It is characterized in that The following rough adjustment steps are included: The predistorter collects the signal x(n), performs predistortion processing on the signal x(n), and generates the signal z(n); the power amplifier amplifies the signal z(n) and generates the signal y(n); the 1 / G module reduces the signal amplitude of the signal y(n) by G times and generates the signal y(n) / G; the operation module operates the signal x(n) and the signal y(n) / G to obtain the signal e(n); the first learner estimates the fitting coefficient according to the signal e(n), determines the coefficient estimation interval, and generates the coefficient interval value; The predistorter performs predistortion processing on the signal x(n) according to the coefficient interval value.
3. The digital predistortion processing method according to claim 2, It is characterized in that The following fine-tuning steps are also included: The second learner processes the signal y(n) / G to generate a signal z`(n); the operation module collects the signal z(n), operates the signal z(n) and the signal z`(n) to obtain the signal e(n).
4. The digital predistortion processing method according to claim 3, It is characterized in that The following steps are also included: A threshold of the signal e(n) is set. When the signal e(n) obtained by calculating the signal z(n) and the signal z`(n) is greater than the set threshold, the fine-tuning step is repeated until the signal e(n) is less than or equal to the set threshold.
5. The digital predistortion processing method according to claim 4, It is characterized in that The following steps are also included: When the fine tuning step is repeated more than the set number of times, it switches to the coarse tuning step, adjusts the coefficient estimation interval, and then performs the fine tuning step processing until the signal e(n) is less than or equal to the set threshold.
