Self-adaptive power amplifier nonlinear correction system based on independent architecture

Abstract

The invention provides a self-adaptive power amplifier nonlinear correction system based on an independent architecture, the system is composed of a digital processing unit, an analog-to-digital converter and a first digital-to-analog converter, a gating dynamic neural network is configured in the digital processing unit, and the gating dynamic neural network is composed of a gating network and a backbone network; a baseband signal in the communication equipment forms an initial analog signal after passing through a second digital-to-analog converter of the communication equipment, and the initial analog signal is converted into a digital signal through an analog-to-digital converter and is transmitted to a digital processing unit. And data processing is carried out through a gating network and a backbone network in the digital processing unit to form a corrected signal, then the corrected signal is converted into a target analog signal through a first digital-to-analog converter, and the target analog signal is sent to a power amplifier of the communication equipment. In combination with the gating dynamic neural network, efficient, flexible and adaptive dynamic configuration can be realized, and the requirement of an intelligent communication system on the linearization performance of the power amplifier can be met.

Description

Technical Field

[0001] This application belongs to the field of communication technologies, and particularly relates to an adaptive power amplifier non-linear correction system based on an independent architecture. Background Art

[0002] Power amplifier linearization technology is a fundamental method for solving the non-linear problems of power amplifiers. Traditional power amplifier linearization methods, such as predistortion, power feeding, etc., are mainly designed and optimized for fixed-frequency communication systems in the civilian field. On the one hand, traditional technologies are difficult to implement in some application scenarios. For example, the algorithm needs to be integrated into the baseband unit of the original device, which limits the flexibility and optimization space of the system. On the other hand, for an intelligent communication system with dynamic transmission configurations (such as power, frequency, modulation mode), when facing rapid changes in parameters such as power, frequency, and modulation mode, traditional methods are difficult to track parameter changes in real time and adjust the linearization strategy in a timely manner, resulting in poor effects and being unable to meet the strict requirements of the intelligent communication system for power amplifier linearization performance. Summary of the Invention

[0003] In view of this, this application aims to propose an adaptive power amplifier non-linear correction system based on an independent architecture to solve at least one of the above problems.

[0004] To achieve the above object, the technical solution of this application is realized as follows: This application provides an adaptive power amplifier non-linear correction system based on an independent architecture, characterized in that: The system consists of a digital processing unit, an analog-to-digital converter, and a first digital-to-analog converter. Among them, a gated dynamic neural network is configured in the digital processing unit, and the gated dynamic neural network is composed of a gating network and a backbone network; The baseband signal in the communication device forms an initial analog signal after passing through the second digital-to-analog converter of the communication device itself. The initial analog signal is converted into a digital signal by the analog-to-digital converter and transmitted to the digital processing unit, and is processed by the gating network and the backbone network in the digital processing unit to form a corrected signal, and then is converted into a target analog signal by the first digital-to-analog converter, and the target analog signal is sent to the power amplifier of the communication device.

[0005] Furthermore, the backbone network is configured to characterize and correct the non-linear behavior of the power amplifier, and it is composed of an input data layer, a filtering layer, a vector decomposition layer, a non-linear layer, a phase recovery layer, and a synthesis layer.

[0006] Furthermore, the filtering layer consists of n parallel FIR networks, where the FIR network formula is as follows: ; ; In the formula, represents the weight parameter of each network, represents the bias parameter of each network, represents the Hadamard product, represents the gating signal of the filtering layer output by the gating network, represents the output data of the input data layer, 、 represents the output data of the filtering layer of component.

[0007] Furthermore, the vector decomposition layer decomposes the output of the filtering layer to obtain the amplitude and phase components. The specific formula is as follows: ; ; In the formula, 、 represent the of the output data of the filtering layer component, represents the phase component, represents the amplitude.

[0008] Furthermore, the non-linear layer is configured to correct the amplitude obtained after decomposition by the vector decomposition layer. The specific formula is as follows: ; In the formula, represents the weight parameter of the network, represents the bias parameter of the network, represents the activation function, represents the Hadamard product, represents the non-linear layer gating signal output by the gating network, represents the corrected amplitude.

[0009] Furthermore, the phase recovery layer is configured to recombine the corrected amplitude and the original phase into a complex signal. The specific formula is as follows: ; Among them, , .

[0010] Furthermore, the synthesis layer is configured to weightedly fuse the correction results of each parallel branch to output m signal data. The specific formula is as follows: ; In the formula, Represents the weight parameters of the network, Represents the bias parameters of the network, Represents the output data of the phase recovery layer.

[0011] Furthermore, the gating network is configured to learn the pattern of the power amplifier behavior changing with the transmission configuration, and dynamically adjust the weights of the backbone network during the power amplifier linearization correction process, wherein the gating network consists of an input data layer, a feature extraction layer, and a gating weight generation layer.

[0012] Furthermore, the feature extraction layer is configured to extract the data features of the input data layer, and the specific formula is as follows: ; Wherein, the input data , , ; In the formula, Represents the maximum value of power, Represents the power value of the current configuration, Represents the maximum value of frequency, Represents the frequency of the current configuration, , Represents the weight parameters of the network, , Represents the bias parameters of the network, Represents the activation function, Represents the data features extracted by the feature extraction layer.

[0013] Furthermore, the gating weight generation layer is configured to map the extracted data features into gating weights, and the specific formula is as follows: ; ; In the formula, , Represents the weight parameters of the network, , Represents the bias parameters of the network, Represents the Sigmod activation function, , Represents the gating weights, Represents the data features extracted by the feature extraction layer.

[0014] Compared with the prior art, the adaptive power amplifier nonlinear correction system based on an independent architecture described in this application has the following beneficial effects: An adaptive power amplifier nonlinear correction system based on an independent architecture described in this application avoids the dependence on the baseband unit of traditional baseband-side power amplifier linearization correction algorithms and is applicable to scenarios where it is difficult to perform customized development on the baseband. This system combines a gated dynamic neural network and can achieve high efficiency, flexibility, and adaptation to dynamic configuration, meeting the strict requirements of intelligent communication systems for power amplifier linearization performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which form a part of this application, are used to provide a further understanding of this application. The schematic embodiments and descriptions thereof of this application are used to explain this application and do not constitute an improper limitation to this application. In the drawings: Figure 1 FIG. is a schematic diagram of the radio frequency architecture of an adaptive power amplifier nonlinear correction system based on an independent architecture described in an embodiment of this application; Figure 2 FIG. is a schematic diagram of the network structure of an adaptive power amplifier nonlinear correction system based on an independent architecture described in an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the following further describes this application in detail with reference to specific embodiments and the accompanying drawings.

[0017] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the embodiments of this application should have the ordinary meaning understood by those of ordinary skill in the field to which this application belongs. The "first", "second", and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this term cover the elements or objects listed after this term and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left", "right", etc. are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0018] Please refer to Figure 1 As shown, this embodiment provides an adaptive power amplifier nonlinear correction system based on an independent architecture. The system architecture is independent of the original communication device. Without changing the original communication system, it is directly connected to the original communication system through a radio frequency cable, and by adopting analog signal input and analog signal output This method avoids the dependence on the baseband unit in the traditional baseband-side power amplifier linearization correction algorithm and is applicable to scenarios where it is difficult to perform customized development on the baseband.

[0019] This system consists of three parts: a digital processing unit, an analog-to-digital converter (ADC), and a first digital-to-analog converter (DAC).

[0020] The baseband signal in the communication device forms an initial analog signal after passing through the second digital-to-analog converter (DAC) of the communication device itself , and this initial analog signal is converted into a digital signal by the digital-to-analog converter of this patent system and is connected to the digital unit of this system; The digital signal forms a corrected signal after being processed by the gated dynamic neural network (i.e., the predistortion network, which consists of a gated network and a backbone network) in the digital processing unit , is converted into an analog signal by the first digital-to-analog converter , and the signal is sent to the power amplifier (PA) of the communication device.

[0021] The gated dynamic neural network consists of two parts: a gated network and a backbone network. The gated dynamic neural network runs in the digital processing unit of this system and is used to correct the nonlinearity of the communication device PA; the backbone network can be any neural network-based power amplifier nonlinear correction model designed for a fixed configuration, and the backbone network is mainly used to characterize and correct the nonlinear behavior of the power amplifier.

[0022] The backbone network adopted in this system consists of six parts: an input data layer, a filtering layer, a vector decomposition layer, a nonlinear layer, a phase recovery layer, and a synthesis layer. The specific description is as follows: 1) Input data layer: The signal input into this system needs to be organized into two-dimensional data, which consists of the I / Q components of the input data, the cubic terms of the I / Q data, the quintic terms of the I / Q data, and the septic terms of the I / Q data. Specifically as follows:

[0023] In the formula, , represent the I / Q components of the data, , , are the cubic terms of the I / Q data, the quintic terms of the I / Q data, and the septic terms of the I / Q data respectively. The specific calculation formulas are as follows:

[0024] Wherein, n is the memory depth, indicating how many previous input data the power amplifier output data is related to. Usually, it needs to be selected according to the actual power amplifier situation, and m = 3, 5, 7.

[0025] 2) Filtering layer: The filtering layer consists of n parallel FIR networks. The number n of FIR networks is consistent with the memory depth of the input data and is used to model the memory effect of the power amplifier. The output after passing through the filtering layer is .

[0026]

[0027] Wherein, and represent the component of the output data of the filtering layer, represents the phase component, represents the amplitude.

[0028] The calculation formula of the FIR network is as follows:

[0029] Wherein, represents the weight parameter of each network, represents the bias parameter of each network, represents the Hadamard product, represents the gating signal of the filtering layer output by the gating network.

[0030] 3) Vector decomposition layer: The output of the FIR network is vector decomposed to obtain the amplitude and phase components. The specific calculation formula is as follows:

[0031]

[0032] 4) Nonlinear layer: The decomposed amplitude is sent to the nonlinear layer to model and compensate for the nonlinear behavior of the power amplifier. The output result of the nonlinear layer is . The calculation formula of the nonlinear layer is as follows:

[0033] Wherein, represents the weight parameter of the network, represents the bias parameter of the network, represents the activation function, represents the Hadamard product, represents the gating signal of the nonlinear layer output by the gating network.

[0034] 5) Phase recovery layer: recombine the corrected amplitude with the original phase to form a complex signal .

[0035]

[0036] In the formula, , The specific calculation formulas are as follows:

[0037]

[0038] 6) Synthesis layer: The synthesis layer weights and fuses the correction results of each parallel branch to enhance the non-linear compensation ability and adapt to the complex PA characteristics. The synthesis layer outputs m I / Q signal data. The specific calculation formula is as follows:

[0039] In the formula, represents the weight parameter of the network, represents the bias parameter of the network, represents the output of m I / Q signal data.

[0040] The gating network adopted in this system is used to learn the pattern of PA behavior changing with transmission configurations (such as power, frequency point, etc.), and dynamically adjust the weights of the backbone network during the power amplifier linearization correction process, enabling the system to quickly adapt to different transmission configurations.

[0041] The gating network consists of three parts: an input data layer, a feature extraction layer, and a gating weight generation layer. The specific description is as follows: 1) Input data layer: The input data includes information related to transmission configurations such as normalized power ( ), frequency point ( ), etc.

[0042]

[0043] Among them, , ; represents the maximum value of power, represents the power value of the current configuration; represents the maximum value of frequency, represents the frequency of the current configuration.

[0044] 2) Feature extraction layer: The feature extraction layer is used to extract the data features of the input layer. The specific calculation formula is as follows:

[0045] Wherein: 、 represent the weight parameters of the network, 、 represent the bias parameters of the network, represents the activation function.

[0046] 3) Gating weight generation layer: The gating weight generation layer maps the extracted features to gating weights and . The specific calculation formula is as follows:

[0047]

[0048] In the formula, 、 represent the weight parameters of the network, 、 represent the bias parameters of the network, represents the Sigmod activation function.

[0049] Finally, it should be noted that: The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: They can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered by the scope of the claims and the description of the present invention.

[0050] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the embodiments of the present application shall be included in the protection scope of the present application.

Claims

1. An adaptive power amplifier nonlinear correction system based on an independent architecture, characterized in that: The system consists of a digital processing unit, an analog-to-digital converter, and a first digital-to-analog converter. Among them, a gated dynamic neural network is configured in the digital processing unit, and the gated dynamic neural network is composed of a gating network and a backbone network; The baseband signal in the communication device forms an initial analog signal after passing through the second digital-to-analog converter of the communication device itself. The initial analog signal is converted into a digital signal by the analog-to-digital converter and transmitted to the digital processing unit, and is processed by the gating network and the backbone network in the digital processing unit to form a corrected signal, and then is converted into a target analog signal by the first digital-to-analog converter, and the target analog signal is sent to the power amplifier of the communication device.

2. The system according to claim 1, characterized in that: The backbone network is configured to characterize and correct the nonlinear behavior of the power amplifier, and it is composed of an input data layer, a filtering layer, a vector decomposition layer, a nonlinear layer, a phase recovery layer, and a synthesis layer.

3. The system according to claim 2, wherein The filtering layer consists of n parallel FIR networks. Among them, the FIR network formula is as follows: ； ； Wherein, represents the weight parameter of each network, represents the bias parameter of each network, represents the Hadamard product, represents the filtering layer gating signal output by the gating network, represents the output data of the input data layer, and represent the output data of the filtering layer of components.

4. The system according to claim 2, characterized in that The vector decomposition layer performs vector decomposition on the output of the filtering layer to obtain amplitude and phase components. The specific formula is as follows: ； ； In the formula, , represent the output data of the filtering layer 's component, represents the phase component, represents the amplitude.

5. The system according to claim 4, characterized in that, The nonlinear layer is configured to correct the amplitude obtained after decomposition by the vector decomposition layer. The specific formula is as follows: ； In the formula, represents the weight parameter of the network, represents the bias parameter of the network, represents the activation function, represents the Hadamard product, represents the non-linear layer gating signal output by the gating network, represents the corrected amplitude.

6. The system according to claim 5, wherein The phase recovery layer is configured to recombine the corrected amplitude and the original phase into a complex signal. The specific formula is as follows: ； Among them, , .

7. The system according to claim 2, wherein The synthesis layer is configured to weighted-fuse the calibration results of each parallel branch to output m signal data, and the specific formula is as follows: ； wherein, represents the weight parameter of the network, represents the bias parameter of the network, represents the output data of the phase recovery layer.

8. The system according to claim 1, characterized in that: The gating network is configured to learn the mode of the power amplifier behavior changing with the transmission configuration and dynamically adjust the weights of the backbone network during the power amplifier linearization correction process. Among them, the gating network is composed of an input data layer, a feature extraction layer, and a gating weight generation layer.

9. The system according to claim 8, wherein The feature extraction layer is configured to extract the data features of the input data layer. The specific formula is as follows: ； Among them, the input data , , ; In the formula, represents the maximum value of power, represents the power value of the current configuration, represents the maximum value of frequency, represents the frequency of the current configuration, and represent the weight parameters of the network, and represent the bias parameters of the network, represents the activation function, represents the data features extracted by the feature extraction layer.

10. The system according to claim 8, characterized in that, The gating weight generation layer is configured to map the extracted data features into gating weights. The specific formula is as follows: ； ； wherein, , represent the weight parameters of the network, , represent the bias parameters of the network, represents the Sigmod activation function, , represent the gating weights, represents the data features extracted by the feature extraction layer.

Images