KR102550079B1 - An apparatur for compensation of nonlinearly distorted signals caused by power amplifiers based on deep learning and method thereof - Google Patents

Abstract

A compensation method for nonlinear distortion of a power amplifier is provided. The method includes receiving a signal amplified based on a power amplifier from a transmitting device and determining an input signal of the power amplifier corresponding to the received signal using a nonlinear distortion compensation model based on an artificial neural network. can include

Description

Compensation method and apparatus for nonlinear distortion of power amplifier based on deep learning

The present invention relates to a power amplifier, and more particularly to an apparatus and method for compensating for nonlinear distortion of a power amplifier.

Orthogonal Frequency Division Multiplexing (ODFM) is a type of multi-carrier modulation. After converting a sequence of symbols input serially into a parallel form in units of N blocks, each element symbol is It can be transmitted by modulating subcarriers having mutual orthogonality and adding them respectively. OFDM is gradually expanding its application area due to the advantages of being resistant to multi-path fading occurring in a wireless communication environment and enabling high-speed data transmission.

However, the OFDM signal has a relatively high peak-to-average power ratio (PAPR), which requires a large input power back-off for the power amplifier, which drastically reduces the efficiency of the power amplifier. . In particular, this is a very important technical issue for Internet-of-Things (IoT) devices requiring low power consumption and/or unmanned aerial vehicles (UAVs) of an aerial communication network.

Korean Patent Registration No. 10-1679230 ("Polynomial digital predistortion device and method for compensating nonlinear characteristics of power amplifier", Sogang University Industry-University Cooperation Foundation)

One object of the present invention to solve the above problems is to compensate for nonlinear distortion of a power amplifier based on an artificial neural network in a receiving end of a wireless signal, so that an additional calculation procedure is not required in a transmitting end and a conventional repetition-based compensation technique is performed in a receiving end. It is to provide a method capable of efficiently compensating for nonlinear distortion of a power amplifier even with a reduced computational load compared to the above.

Another object of the present invention to solve the above problems is to compensate for nonlinear distortion of a power amplifier based on an artificial neural network at a receiving end of a wireless signal, so that an additional calculation procedure is not required at the transmitting end and a conventional repetition-based compensation technique is performed at the receiving end. An object of the present invention is to provide a device capable of efficiently compensating for nonlinear distortion of a power amplifier even with a reduced computational load compared to the above.

However, the problem to be solved by the present invention is not limited thereto, and may be expanded in various ways without departing from the spirit and scope of the present invention.

A method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention for achieving the above object is performed by a processor of a computer, and the method is based on a power amplifier from a transmission device. Receiving the amplified signal with; and determining an input signal of the power amplifier corresponding to the received signal by using an artificial neural network-based nonlinear distortion compensation model.

According to one aspect, the signal from the transmitter may be an Orthogonal Frequency Division Multiplexing (ODFM) signal.

According to one aspect, the power amplifier may be configured to non-linearly amplify the signal for at least a portion of a power magnitude range.

According to one aspect, the nonlinear distortion compensation model may be generated by learning a deep neural network (DNN) based on a plurality of training data.

According to one aspect, each of the plurality of training data includes, as inputs, a conventional amplified received signal and a channel fading gain corresponding to the conventional amplified received signal, and An input signal of a corresponding conventional power amplifier may be included as an output.

According to one aspect, the amplified received signal may be a time-domain signal.

According to one aspect, the determining of the input signal of the power amplifier may include inputting an amplified signal received from the transmitter and a measurement value of a channel fading gain to the nonlinear distortion compensation model; and receiving an input signal of the power amplifier corresponding to the received signal from the nonlinear distortion compensation model.

According to one aspect, the receiving may be configured to receive the signal from at least one of an IoT device or an unmanned aerial vehicle (UAV) of an aerial communication network.

According to one aspect, at least one of the IoT device or the unmanned aerial vehicle may be configured to transmit/receive a signal based on a frequency-nonselective channel.

An apparatus for compensating for nonlinear distortion of a power amplifier according to another embodiment of the present invention for solving the above-described problem includes a processor and a transceiver, and the processor uses the transceiver to transmit power to the power amplifier (Power Amplifier) to receive an amplified signal based on; And it may be configured to determine an input signal of the power amplifier corresponding to the received signal using an artificial neural network-based nonlinear distortion compensation model.

According to one aspect, the signal from the transmitter may be an Orthogonal Frequency Division Multiplexing (ODFM) signal.

According to one aspect, the power amplifier may be configured to non-linearly amplify the signal for at least a portion of a power magnitude range.

According to one aspect, the nonlinear distortion compensation model may be generated by learning a deep neural network (DNN) based on a plurality of training data.

According to one aspect, each of the plurality of training data includes, as inputs, a conventional amplified received signal and a channel fading gain corresponding to the conventional amplified received signal, and An input signal of a corresponding conventional power amplifier may be included as an output.

According to one aspect, the amplified received signal may be a time-domain signal.

According to one aspect, determining the input signal of the power amplifier may include inputting an amplified signal received from the transmitter and a measurement value of a channel fading gain to the nonlinear distortion compensation model; and receiving an input signal of the power amplifier corresponding to the received signal from the nonlinear distortion compensation model.

According to one aspect, the receiving may be configured to receive the signal from at least one of an IoT device or an unmanned aerial vehicle (UAV) of an aerial communication network.

According to one aspect, at least one of the IoT device or the unmanned aerial vehicle may be configured to transmit/receive a signal based on a frequency-nonselective channel.

A computer-readable storage medium according to another embodiment of the present invention for solving the above problems is a computer-readable storage medium storing instructions executable by a processor, wherein the instructions are stored in the processor when executed by the processor. Receive a signal amplified based on a power amplifier from a transmission device using a transceiver; And it may be configured to determine an input signal of the power amplifier corresponding to the received signal using an artificial neural network-based nonlinear distortion compensation model.

The disclosed technology may have the following effects. However, it does not mean that a specific embodiment must include all of the following effects or only the following effects, so it should not be understood that the scope of rights of the disclosed technology is limited thereby.

According to an apparatus and method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention, a receiving end of a communication system having better computing power and power resources than a transmitting end compensates for distortion of a power amplified signal, , For example, it is possible to compensate for distortion of a signal due to a power amplifier without increasing computational complexity at all in a type of transmitter that operates with low power, such as IoT and UAV. In particular, there is an advantage in that the signal distortion problem can be solved at the receiving end without modifying the operation of the transmitters in an environment where a large number of devices exist, such as in a massive IoT. In this regard, it is possible to compensate for signal distortion without increasing the computational complexity of the transmitting end even by an iterative decoding algorithm, but according to the apparatus and method according to an embodiment of the present invention, the performance for signal distortion or the performance of the receiving end It is superior to iterative-based decoding algorithms in both terms of computational complexity. In addition, unlike an iteration-based decoding algorithm, there is an advantage in that computational complexity does not increase even when the number of subcarriers of OFDM increases by deploying a deep neural network in the time domain.

1 shows an OFDM transmitter using a power amplifier.
2 shows a receiving end of an iteration-based decoding algorithm in the case of using hard decoding at every iteration (iteration) to estimate OFDM frequency domain symbols.
3 shows a receiving end of an iteration-based decoding algorithm in the case of performing channel decoding in every iteration for OFDM frequency domain symbol estimation.
4 is a schematic flowchart of a method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention.
FIG. 5 is a detailed flowchart of the input signal determining step in FIG. 4;
6 shows offline training for decoding a DNN-based nonlinear power amplified signal according to an embodiment of the present invention.
7 illustrates online deployment (test stage) of decoding of a DNN-based nonlinear power amplified signal according to an embodiment of the present invention.
8 shows a comparison of BLER performance between a DNN-based system and an iterative-based algorithm.
9 is a block diagram showing the configuration of a computing system in which a method according to an embodiment of the present invention may be performed.

Since the present invention can make various changes and have various embodiments, specific embodiments are illustrated in the drawings and described in detail.

However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. These terms are only used for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present invention. The terms and/or include any combination of a plurality of related recited items or any of a plurality of related recited items.

It is understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. It should be. On the other hand, when an element is referred to as “directly connected” or “directly connected” to another element, it should be understood that no other element exists in the middle.

Terms used in this application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, the terms "include" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present application, they should not be interpreted in an ideal or excessively formal meaning. don't

Hereinafter, with reference to the accompanying drawings, preferred embodiments of the present invention will be described in more detail. In order to facilitate overall understanding in the description of the present invention, the same reference numerals are used for the same components in the drawings, and redundant descriptions of the same components are omitted.

As discussed above, OFDM is increasingly being applied to due to its strength against multi-path fading occurring in a wireless communication environment and high-speed data transmission.

Relatedly, FIG. 1 shows an OFDM transmitting end using a power amplifier. As shown in FIG. 1, the ODFM transmitter can encode bits (source bits) of information to be transmitted using a channel encoder and then map them to symbols. Afterwards, Inverse Fast Fourier Transformation (IFFT) is performed on the symbols, and a signal is amplified using a power amplifier to transmit time-domain signals.

However, the OFDM signal has a relatively high peak-to-average power ratio (PAPR), which requires a large input power back-off for the power amplifier, which drastically reduces the efficiency of the power amplifier. . In particular, this is a very important technical issue for Internet-of-Things (IoT) devices requiring low power consumption and/or unmanned aerial vehicles (UAVs) of an aerial communication network.

In order to solve the low efficiency of such a power amplifier, for example, the following techniques may be considered.

First, a method of reducing the PAPR of an OFDM signal at a transmitter may be considered. However, since this method requires a high level of computational complexity and coding overhead in the transmitter, it is not suitable when the transmitter's device is a low-power device.

Next, a method of linearizing the power amplifier using a digital predistorter can be considered. This not only increases the computational complexity of the transmitting end, but also has a disadvantage in that linearity of the channel must be maintained.

Meanwhile, a method of decoding a nonlinearly distorted signal at a receiving end using an iterative decoding algorithm may be considered. Relatedly, Figure 2 shows an iteration based decoding algorithm. More specifically, FIG. 2 shows a receiving end of an iteration-based decoding algorithm in the case of using hard decoding at every iteration to estimate an OFDM frequency domain symbol, and FIG. 3 shows a channel at every iteration for OFDM frequency domain symbol estimation. Indicates a receiving end of an iterative-based decoding algorithm when decoding is performed. As shown in FIGS. 2 and 3, after estimating a frequency domain symbol at the receiving end, a signal distortion component is estimated using a power amplifier model used at the transmitting end and subtracted from the received OFDM signal (time domain). Once the distortion component is reduced, more accurate frequency domain symbol estimation is possible in the next iteration, and the signal distortion component is also more accurately subtracted. The disadvantage of this algorithm is that FFT and IFFT, which are computationally complex, must be performed at each iteration. Moreover, in order to increase the accuracy of symbol estimation in an iterative process, channel decoding, which has a higher computational complexity than FFT, must be performed in every iteration (see FIG. 3).

According to the nonlinear distortion compensation method of a power amplifier according to an embodiment of the present invention, deep learning-based nonlinear distortion compensation can be performed in order to improve the performance of existing iteration-based algorithms and overcome excessive computational complexity.

For example, DNN based on a multilayer perceptron structure has the advantage of solving complex nonlinear problems through back propagation techniques, so deep learning techniques can be applied to areas such as channel coding, modulation recognition, channel estimation, and spectrum sensing. In this regard, according to a method and apparatus for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention, a signal amplified nonlinearly based on deep learning can be decoded at the receiving end.

4 is a schematic flowchart of a method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention, and FIG. 5 is a detailed flowchart of an input signal determining step in FIG. 4 . Hereinafter, a method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention will be described in more detail with reference to FIGS. 4 and 5 .

As shown in FIG. 4, a method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention can be performed by a processor of a computer, and first, from a transmission device, a power amplifier An amplified signal may be received based on (step 410).

According to one aspect, the signal from the transmitter may be an Orthogonal Frequency Division Multiplexing (ODFM) signal. As discussed above, since OFDM has a relatively high PAPR, it may have a disadvantage that the power amplifier may be out of the range of the power size of a linearly operated signal. In order not to deviate from the linear operating range, it is required to use a power amplifier having a high dynamic range, which may result in an increase in cost.

Here, the power amplifier may be configured to non-linearly amplify the signal for at least a part of the power size range. For example, a power amplifier has a dynamic range that can be operated linearly, and when the input power is out of this dynamic range, the linearity is lost and a non-linearly amplified signal is output, so that the signal before amplification is lost at the receiving end. It can also make it difficult to estimate.

Referring back to FIG. 4 , a nonlinear distortion compensation method according to an embodiment of the present invention determines an input signal of a power amplifier corresponding to a signal received from a transmitter using an artificial neural network-based nonlinear distortion compensation model. Yes (step 420). That is, despite the nonlinear distortion of the power amplifier, it is possible to decode the signal prior to being amplified by the power amplifier or the bits of information to be transmitted.

According to one aspect, the nonlinear distortion compensation model may be generated by learning a deep neural network (DNN) based on a plurality of training data. In this regard, according to one aspect of the present invention, each of the plurality of training data includes, as an input, a conventional amplified received signal and a channel fading gain corresponding to the conventional amplified received signal, and An input signal of a conventional power amplifier corresponding to the amplified received signal may be included as an output. Also, for example, the amplified received signal may be a time-domain signal.

More specifically, FIG. 6 shows offline training for decoding a DNN-based nonlinear power amplified signal according to an embodiment of the present invention. As shown in FIG. 6, according to the nonlinear distortion compensation method according to an embodiment of the present invention, a fully-connected layer (FCL) neural network is placed in the time domain (Time) of an OFDM receiver to input a received signal and a channel estimation value. After receiving it as , it is possible to output a signal compensated for nonlinear distortion. In the off-line training step as shown in FIG. 6, the DNN can be trained using the supervised learning method by using the input signal of the transmitting end power amplifier as a label message. At this time, the loss function 650 can be set to mean-squared error (MSE) calculated from the DNN output signal and the label message.

As shown in FIG. 6, an artificial neural network that can be used for the nonlinear distortion compensation method according to an embodiment of the present invention may be, for example, a deep neural network (DNN), and includes an input layer 620, a hidden layer ( 631, 633, 635) and an output layer 640. As exemplarily shown in FIG. 6, the hidden layer may include, for example, three hidden layers such as a second layer 631, a third layer 633, and a fourth layer 635, but this is an example. , but not limited to this.

A training data set for training a deep neural network may include a conventional received time domain signal 611 and a corresponding channel fading gain 613 as inputs, and a corresponding input of a power amplifier signal 615 as an output. That is, the label message may be a received time domain signal obtained from conventional data, a channel fading gain, and a corresponding input signal of a power amplifier.

By learning the deep neural network based on conventional data, a nonlinear distortion compensation model capable of outputting an input signal of a power amplifier according to the received amplification signal and measured values of the channel fading gain can be obtained.

In this regard, referring again to FIGS. 4 and 5 , in the method for compensating for nonlinear distortion according to an embodiment of the present invention, determining an input signal of a power amplifier using a nonlinear distortion compensation model (step 420). Step 421 of inputting the amplified signal received from the transmitter and the measured value of the channel fading gain to the nonlinear distortion compensation model and the power amplifier corresponding to the received signal from the nonlinear distortion compensation model It may include a step of receiving an input signal of (step 423).

7 illustrates online deployment (test stage) of decoding of a DNN-based nonlinear power amplified signal according to an embodiment of the present invention. As shown in FIG. 7, if the received time-domain signals and the measured value of the channel fading gain are input to the learned time-domain deep neural network (nonlinear distortion compensation model) (Time-domain deep neural network), power A signal before being amplified by an amplifier can be obtained. Subsequently, FFT is performed on these signals, demapping is performed on symbols, and then decoded through a channel decoder to obtain decoded source bits of information that the transmitter side wants to transmit.

Meanwhile, according to one aspect of the present invention, a scenario (frequency-nonselective channel) in which an IoT device and a UAV of an air communication network use a narrow frequency band may be considered. Most UAV flight control communication or low-power IoT uses narrowband transmission because of the advantages of wide coverage, high spectral efficiency, and long battery life due to reduced noise bandwidth. In relation to this, in the nonlinear distortion compensation method according to an embodiment of the present invention, the step of receiving the amplified signal (step 410) is, for example, an IoT device or an unmanned aerial vehicle of an aerial communication network. , UAV) may be configured to receive a signal from at least one of. Also, according to one aspect, at least one of the IoT device or the unmanned aerial vehicle may be configured to transmit/receive a signal based on a frequency-nonselective channel.

8 shows a comparison of BLER performance between a DNN-based system and an iterative-based algorithm. That is, block error rate (BLER) performance according to an exemplary implementation of the present invention is shown in FIG. 8 . The iteration-based decoding algorithm shown in FIG. 8 is the result of performing iterations 8 times, and hard decoding is performed instead of complex channel decoding in each iteration for symbol estimation (see FIG. 3). As shown in FIG. 8, the nonlinear distortion compensation method according to an embodiment of the present invention shows much better performance than the conventional iteration-based algorithm.

The system parameters used in the simulation are as follows.

,

,

,

)- Power amplifier model - TWT (Saleh model) (

,

,

,

)

- Number of OFDM symbols per frame: 12, Number of OFDM subcarriers: 128, Modulation: 64QAM

- Convolutional code: constratin length=11 & code rate=1/3.

- Fast fading assumption: coherence time = one OFDM symbol

- DNN structure: 4 layers (2 hidden layers). Number of nodes per layer: 2, 11, 11, 11, 1

As shown in FIG. 8, it can be seen that the DNN structure of the nonlinear distortion compensation method according to an embodiment of the present invention can sufficiently analyze and learn the nonlinearity of the power amplifier combined with channel fading using only three hidden layers. .

According to an apparatus and method for compensating for nonlinear distortion of a power amplifier according to an embodiment of the present invention, a receiving end of a communication system having better computing power and power resources than a transmitting end compensates for distortion of a power amplified signal, , For example, it is possible to compensate for distortion of a signal due to a power amplifier without increasing computational complexity at all in a type of transmitter that operates with low power, such as IoT and UAV. In particular, there is an advantage in that the signal distortion problem can be solved at the receiving end without modifying the operation of the transmitters in an environment where a large number of devices exist, such as in a massive IoT. In this regard, it is possible to compensate for signal distortion without increasing the computational complexity of the transmitting end even by an iterative decoding algorithm, but according to the apparatus and method according to an embodiment of the present invention, the performance for signal distortion or the performance of the receiving end It is superior to iterative-based decoding algorithms in both terms of computational complexity. In addition, unlike an iteration-based decoding algorithm, there is an advantage in that computational complexity does not increase even when the number of subcarriers of OFDM increases by deploying a deep neural network in the time domain.

On the other hand, an apparatus for compensating for nonlinear distortion of a power amplifier according to another embodiment of the present invention includes a processor and a transceiver, and the processor uses the transceiver to amplify from a transmission apparatus based on a power amplifier. It may be configured to receive a signal, and determine an input signal of the power amplifier corresponding to the received signal using an artificial neural network-based nonlinear distortion compensation model. A specific operation of the nonlinear distortion compensating device according to an embodiment of the present invention may be performed according to the above-described nonlinear distortion compensation method according to an embodiment of the present invention.

9 is a block diagram showing the configuration of a computing system in which a method according to an embodiment of the present invention can be performed. Referring to FIG. 9 , a computing system 800 may include a flash storage 810 , a processor 820 , a RAM 830 , an input/output device 840 and a power supply 850 . In addition, the flash storage 810 may include a memory device 811 and a memory controller 812 . Meanwhile, although not shown in FIG. 8 , the computing system 800 may further include ports capable of communicating with video cards, sound cards, memory cards, USB devices, etc., or with other electronic devices. .

The computing system 800 may be implemented as a personal computer or as a portable electronic device such as a notebook computer, a mobile phone, a personal digital assistant (PDA), and a camera.

Processor 820 can perform certain calculations or tasks. Depending on the embodiment, the processor 820 may be a micro-processor or a central processing unit (CPU). The processor 820 communicates with the RAM 830, the input/output device 840, and the flash storage 810 through a bus 860 such as an address bus, a control bus, and a data bus. communication can be performed. Flash storage 810 can be implemented using the flash storage of the embodiments shown in FIGS. 5-7 .

According to one embodiment, the processor 820 can also be coupled to an expansion bus, such as a Peripheral Component Interconnect (PCI) bus.

The RAM 830 may store data necessary for the operation of the computing system 800 . Any type of random access memory including, for example, DRAM, mobile DRAM, SRAM, PRAM, FRAM, MRAM, RRAM, RAM (830).

The input/output device 840 may include input means such as a keyboard, keypad, and mouse, and output means such as a printer and a display. The power supply 850 can supply an operating voltage necessary for the operation of the computing system 800 .

The method according to the present invention described above may be implemented as computer readable codes on a computer readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

Although the above has been described with reference to the drawings and examples, it does not mean that the scope of protection of the present invention is limited by the drawings or examples, and those skilled in the art will understand the scope of the present invention described in the claims below. It will be understood that various modifications and changes may be made to the present invention without departing from the spirit and scope.

Specifically, the described features may be implemented within digital electronic circuitry, or within computer hardware, firmware, or combinations thereof. Features may be implemented in a computer program product embodied within storage, eg, in a machine-readable storage device, for execution by a programmable processor. And features can be performed by a programmable processor executing a program of instructions to perform the functions of the described embodiments by operating on input data and generating output. The described features include at least one programmable processor, at least one input device, and at least one output device coupled to receive data and instructions from and to transmit data and instructions to the data storage system. It can be executed within one or more computer programs that can be executed on a programmable system including. A computer program includes a set of instructions that can be used directly or indirectly within a computer to perform a particular action for a given result. A computer program is written in any programming language, including compiled or interpreted languages, and contained as modules, components, subroutines, or other units suitable for use in other computer environments, or as stand-alone programs. can be used in any form.

Suitable processors for execution of a program of instructions include, for example, both general and special purpose microprocessors, and either a single processor or multiple processors in a computer of another kind. Also, storage devices suitable for embodying computer program instructions and data embodying the described features include, for example, semiconductor memory devices such as EPROM, EEPROM, and flash memory devices, magnetic memory devices such as internal hard disks and removable disks. devices, magneto-optical disks and all forms of non-volatile memory including CD-ROM and DVD-ROM disks. The processor and memory may be integrated within or added by ASICs (application-specific integrated circuits).

Although the present invention described above has been described based on a series of functional blocks, it is not limited by the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes are made within the scope of the technical spirit of the present invention. That this is possible will be apparent to those skilled in the art.

Combinations of the above-described embodiments are not limited to the above-described embodiments, and various types of combinations may be provided as well as the above-described embodiments according to implementation and/or needs.

In the foregoing embodiments, the methods are described on the basis of a flow chart as a series of steps or blocks, but the present invention is not limited to the order of steps, and some steps may occur in a different order or concurrently with other steps as described above. there is. In addition, those skilled in the art will understand that the steps shown in the flow chart are not exclusive, that other steps may be included, or that one or more steps of the flow chart may be deleted without affecting the scope of the present invention. You will understand.

The foregoing embodiment includes examples of various aspects. It is not possible to describe all possible combinations to represent the various aspects, but those skilled in the art will recognize that other combinations are possible. Accordingly, it is intended that the present invention cover all other substitutions, modifications and variations falling within the scope of the following claims.

Claims (19)

A method for compensating for nonlinear distortion of a power amplifier, performed by a processor of a computer, the method comprising:
Receiving a signal amplified based on a power amplifier from a transmitting device; and
Determining an input signal of the power amplifier corresponding to the received signal using an artificial neural network-based nonlinear distortion compensation model,
The distortion compensation model is generated by learning a deep neural network (DNN) based on a plurality of training data,
Each of the plurality of training data,
A conventionally amplified received signal and a channel fading gain corresponding to the conventionally amplified received signal as inputs;
An input signal of a conventional power amplifier corresponding to the conventional amplified received signal as an output,
The distortion compensation model is learned by using the conventionally amplified received signal and a channel fading gain corresponding to the conventionally amplified received signal as a label message.
According to claim 1,
The signal from the transmitting device is
A compensation method for nonlinear distortion of a power amplifier, which is an Orthogonal Frequency Division Multiplexing (ODFM) signal.
According to claim 1,
The power amplifier,
A method of compensating for nonlinear distortion in a power amplifier, configured to nonlinearly amplify the signal over at least a portion of the power magnitude range.
delete delete According to claim 1,
The amplified received signal,
A method for compensating for nonlinear distortion of a time-domain signal, a power amplifier.
According to claim 6,
Determining the input signal of the power amplifier,
inputting the amplified signal received from the transmitting device and the measured value of the channel fading gain into the nonlinear distortion compensation model; and
and receiving an input signal of the power amplifier corresponding to the received signal from the nonlinear distortion compensation model.
According to claim 7,
In the receiving step,
A method for compensating for nonlinear distortion of a power amplifier, configured to receive the signal from at least one of an IoT device or an Unmanned Aerial Vehicle (UAV) of an Aerial Communication Network.
According to claim 8,
At least one of the IoT device or the unmanned aerial vehicle,
A method for compensating for nonlinear distortion of a power amplifier, configured to transmit and receive a signal based on a frequency-nonselective channel.
An apparatus for compensating for nonlinear distortion of a power amplifier, wherein the apparatus for compensating for nonlinear distortion includes a processor and a transceiver,
the processor,
Receiving a signal amplified based on a power amplifier from a transmission device using the transceiver; and
And configured to determine an input signal of the power amplifier corresponding to the received signal using an artificial neural network-based nonlinear distortion compensation model,
The distortion compensation model is generated by learning a deep neural network (DNN) based on a plurality of training data,
Each of the plurality of training data,
A conventionally amplified received signal and a channel fading gain corresponding to the conventionally amplified received signal as inputs;
An input signal of a conventional power amplifier corresponding to the conventional amplified received signal as an output,
The distortion compensation model is learned by using the conventionally amplified received signal and a channel fading gain corresponding to the conventionally amplified received signal as a label message.
According to claim 10,
The signal from the transmitting device is
An orthogonal frequency division multiplexing (ODFM) signal, a compensation device for nonlinear distortion of a power amplifier.
According to claim 10,
The power amplifier,
An apparatus for compensating for nonlinear distortion of a power amplifier, configured to nonlinearly amplify the signal over at least a portion of a power magnitude range.
delete delete According to claim 10,
The amplified received signal,
Compensation device for non-linear distortion of a time-domain signal, a power amplifier.
According to claim 15,
Determining the input signal of the power amplifier,
inputting the amplified signal received from the transmitting device and a measurement value of a channel fading gain into the nonlinear distortion compensation model; and
and receiving an input signal of the power amplifier corresponding to the received signal from the nonlinear distortion compensation model.
17. The method of claim 16,
Receiving the
An apparatus for compensating for nonlinear distortion of a power amplifier, configured to receive the signal from at least one of an IoT device or an unmanned aerial vehicle (UAV) of an aerial communication network.
18. The method of claim 17,
At least one of the IoT device or the unmanned aerial vehicle,
An apparatus for compensating for nonlinear distortion of a power amplifier, configured to transmit and receive a signal based on a frequency-nonselective channel.
A computer-readable storage medium storing instructions executable by a processor, comprising:
The instructions, when executed by the processor, cause the processor to:
Receiving a signal amplified based on a power amplifier from a transmission device using a transceiver; and
It is configured to determine an input signal of the power amplifier corresponding to the received signal using an artificial neural network-based nonlinear distortion compensation model,
The distortion compensation model is generated by learning a deep neural network (DNN) based on a plurality of training data,
Each of the plurality of training data,
A conventionally amplified received signal and a channel fading gain corresponding to the conventionally amplified received signal as inputs;
An input signal of a conventional power amplifier corresponding to the conventional amplified received signal as an output,
The distortion compensation model is learned by using the conventionally amplified received signal and a channel fading gain corresponding to the conventionally amplified received signal as a label message.

Images