WO2020019240A1

WO2020019240A1 - Method, apparatus and computer readable media for data processing

Info

Publication number: WO2020019240A1
Application number: PCT/CN2018/097217
Authority: WO
Inventors: Guangling Zhao
Original assignee: Nokia Shanghai Bell Co Ltd; Nokia Solutions and Networks Oy
Current assignee: Nokia Shanghai Bell Co Ltd; Nokia Solutions and Networks Oy
Priority date: 2018-07-26
Filing date: 2018-07-26
Publication date: 2020-01-30
Anticipated expiration: 2021-01-26
Also published as: CN112262369B; CN112262369A

Abstract

Embodiments of the present disclosure relate to methods, apparatuses and computer program products for data processing. A method comprises obtaining input reference data and a first output reference data for training an artificial neural network (ANN); generating a second output reference data by suppressing noise in the first output reference data; and training the ANN based on the input reference data and the second output reference data. In some embodiments, the trained ANN may be used for determining configuration of digital pre-distortion (DPD) in a power amplifier (PA) system accurately.

Description

METHOD, APPARATUS AND COMPUTER READABLE MEDIA FOR DATA PROCESSING

FIELD

Non-limiting and example embodiments of the present disclosure generally relate to a technical field of data processing, and specifically to methods, apparatuses and computer program products for training an artificial neural network (ANN) .

BACKGROUND

This section introduces aspects that may facilitate better understanding of the disclosure. Accordingly, the statements of this section are to be read in this light and are not to be understood as admissions about what is in the prior art or what is not in the prior art.

Modern wireless services demand an efficient and linear transmission of a radio frequency (RF) carrier modulated in amplitude as well as in phase by an envelope signal. The conflicting demands of power efficiency and linearity impose very stringent requirements on a transmitter, in particular on its power amplifier (PA) .

Although class-APA are best in terms of linearity, their efficiency is rather poor as compared with other amplification classes such as “AB” , “C” and Doherty amplifiers. However, higher efficiency leads to higher nonlinearity and PA output will be distorted, often to extent that fails the system performance requirements. Therefore, class-AB power amplifiers or other variations are usually used together with some suitable form of linearization schemes.

Digital pre-distortion (DPD) has been considered as a popular method for compensating the nonlinearity of the PA. In a PA system with DPD, the transfer characteristics of the PA may be modeled by sampling the output of the PA and the inverse characteristics are calculated. Then the digital baseband signal is multiplied by the inverse of the nonlinear transfer characteristics of the PA, up-converted to RF frequencies and applied to the PA input. In this way, the DPD engines can correct output distortion of the PA and achieve higher efficiencies.

A challenge with DPD technology is that the distortion (i.e., non-linear) characteristics of the PA may vary with time, temperature, and biasing, and it is not easy to design a correct pre-distortion algorithm.

SUMMARY

Various embodiments of the present disclosure mainly aim at providing methods, apparatuses and computer storage media for data processing.

In a first aspect of the disclosure, there is provided a method of data processing. The method comprises: obtaining input reference data and first output reference data for training an ANN; generating second output reference data by suppressing noise in the first output reference data; and training the ANN based on the input reference data and the second output reference data.

In some embodiments, generating the second output reference data further may comprise: generating the second output reference data through polynomial fitting based on the input reference data and the first output reference data. In some embodiments, generating the second output reference data may further comprise generating the second output reference data based on a least square (LS) criterion.

In some embodiments, generating the second output reference data may comprise determining an amplitude and a phase of the second output reference data respectively, based on the input reference data and the first output reference data; and generating the second output reference data based on the determined amplitude and phase. In some further embodiments, determining the amplitude and the phase of the second output reference data may comprise: determining the amplitude through polynomial fitting based on an amplitude of the first output reference data relative to the input reference data; and determining the phase through polynomial fitting based on a phase of the first output reference data relative to the input reference data.

In some embodiments, generating the second output reference data may comprise: determining an inphase component and a quadrature component of the second output reference data respectively, based on the input reference data and the first output reference data; and generating the second output reference data based on the determined inphase and quadrature components.

In some embodiments, the method may further comprise determining a parameter for DPD to be applied to a PA based on the trained ANN. In some embodiments, obtaining the input reference data and the first output reference data may comprise: obtaining training data input to the PA as the input reference data; and obtaining feedback data output from the PA in response to the training data as the first output reference data.

In a second aspect of the present disclosure, there is provided an apparatus for data processing. The apparatus comprises at least one processor; and at least one memory including computer program codes; the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus at least to: obtain input reference data and a first output reference data for training an ANN; generate a second output reference data by suppressing noise in the first output reference data; and train the ANN based on the input reference data and the second output reference data.

In a third aspect of the present disclosure, there is provided another apparatus for data processing. The apparatus comprises means for obtaining input reference data and a first output reference data for training an ANN; means for generate a second output reference data by suppressing noise in the first output reference data; and means for training the ANN based on the input reference data and the second output reference data.

In a fourth aspect of the disclosure, there is provided a computer program. The computer program comprises instructions which, when executed by an apparatus, causes the apparatus to carry out the method according to the first aspect of the present disclosure.

In a fifth aspect of the disclosure, there is provided a computer readable medium with a computer program stored thereon which, when executed by at least one processor of a device, causes the device to carry out the method of the first aspect of the present disclosure.

In a six aspect of the present disclosure, there is provided a device for communication. The device comprises: a PA; and a DPD couple to an input of the PA; where a parameter of the DPD is obtained based on an ANN which is trained with input reference data and an output reference data; and wherein the output reference data is generated by suppressing noise in feedback data output from the PA.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other aspects, features, and benefits of various embodiments of the present disclosure will become more fully apparent from the following detailed description with reference to the accompanying drawings, in which like reference signs are used to designate like or equivalent elements. The drawings are illustrated for facilitating better understanding of the embodiments of the disclosure and are not necessarily drawn to scale, in which:

FIG. 1 illustrates a wireless communication network in which embodiments of the present disclosure may be implemented;

FIG. 2 illustrates a flow chart of a method of data processing according to an example embodiment of the present disclosure;

FIG. 3 shows a diagram of an ANN schematically;

FIG. 4 shows an example of reconstructing clean training data via polynomial fitting according to an embodiment of the present disclosure;

FIGs. 5-6 show another example of reconstructing clean training data according to an embodiment of the present disclosure;

FIG. 7 shows a flow chart of a method of reconstructing clean training data according to an embodiment of the present disclosure;

FIG. 8 shows an example for configuring DPD in a PA system based on an ANN according to an embodiment of the present disclosure;

FIG. 9 shows a curve of amplitude to amplitude characteristic of DPD which is configured based on an ANN trained with clean data, according to an embodiment of the present disclosure;

FIG. 10 shows an original spectrum of a PA system without DPD;

FIG. 11 shows a spectrum of a PA system with conventional DPD;

FIG. 12 shows a spectrum of a PA system with DPD designed according to an embodiment of the present disclosure; and

FIG. 13 shows a simplified block diagram of an apparatus which may be utilized for data processing according to an example embodiment of the present disclosure.

DETAILED DESCRIPTION

Hereinafter, the principle and spirit of the present disclosure will be described with reference to illustrative embodiments. It should be understood that all these embodiments are given merely for one skilled in the art to better understand and further practice the present disclosure, but not for limiting the scope of the present disclosure. For example, features illustrated or described as part of one embodiment may be used with another embodiment to yield still a further embodiment. In the interest of clarity, not all features of an actual implementation are described in this specification.

References in the specification to “one embodiment, ” “an embodiment, ” “an example embodiment, ” and the like indicate that the embodiment described may include a particular feature, structure, or characteristic, but it is not necessary that every embodiment includes the particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to affect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.

It shall be understood that although the terms “first” and “second” etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and similarly, a second element could be termed a first element, without departing from the scope of example embodiments. As used herein, the term “and/or” includes any and all combinations of one or more of the listed terms.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be liming of example embodiments. As used herein, the singular forms “a” , “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises” , “comprising” , “has” , “having” , “includes” and/or “including” , when used herein, specify the presence of stated features, elements, and/or components etc., but do not preclude the presence or addition of one or more other features, elements, components and/or combinations thereof.

As used in this application, the term “circuitry” may refer to one or more or all of the following:

(a) hardware-only circuit implementations (such as implementations in only analog and/or digital circuitry) and

(b) combinations of hardware circuits and software, such as (as applicable) :

(i) a combination of analog and/or digital hardware circuit (s) with software/firmware and

(ii) any portions of hardware processor (s) with software (including digital signal processor (s) ) , software, and memory (ies) that work together to cause an apparatus, such as a mobile phone or server, to perform various functions) and

(c) hardware circuit (s) and or processor (s) , such as a microprocessor (s) or a portion of a microprocessor (s) , that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation.

This definition of circuitry applies to all uses of this term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers an implementation of merely a hardware circuit or processor (or multiple processors) or portion of a hardware circuit or processor and its (or their) accompanying software and/or firmware. The term circuitry also covers, for example and if applicable to the particular claim element, a baseband integrated circuit or processor integrated circuit for a computing device.

As used herein, the term “communication network” refers to a network following any suitable communication standards, such as 5G, New Radio (NR) , Long Term Evolution (LTE) , LTE-Advanced (LTE-A) , Wideband Code Division Multiple Access (WCDMA) , High-Speed Packet Access (HSPA) , and so on. The “communication network” may also be referred to as a “communication system. ” Furthermore, communications between network devices, between a network device and a terminal device, or between terminal devices in the communication network may be performed according to any suitable communication protocol, including, but not limited to, Global System for Mobile Communications (GSM) , Universal Mobile Telecommunications System (UMTS) , Long Term Evolution (LTE) , New Radio (NR) , 5G, wireless local area network (WLAN) standards, such as the IEEE 802.11 standards, and/or any other appropriate communication standard either currently known or to be developed in the future.

As used herein, the term “network device” refers to a node in a communication network via which a terminal device receives services. For example, the network device may include, but is not limited to, a base station (BS) , and Node B (NB) , an evolved NB (eNB) , a 5G NB (gNB) , or an access point (AP) , etc.

The term “terminal device” refers to any end device that may be capable of communication. By way of example rather than limitation, a terminal device may also be referred to as a communication device, UE, a Subscriber Station (SS) , a Portable Subscriber Station, a Mobile Station (MS) , or an Access Terminal (AT) . The terminal device may include, but not limited to, a mobile phone, a cellular phone, a smart phone, voice over IP (VoIP) phones, wireless local loop phones, a tablet, a wearable terminal device, a personal digital assistant (PDA) , portable computers, desktop computer, image capture terminal devices such as digital cameras, gaming terminal devices, music storage and playback appliances, vehicle-mounted wireless terminal devices, wireless endpoints, mobile stations, laptop-embedded equipment (LEE) , laptop-mounted equipment (LME) , USB dongles, smart devices, wireless customer-premises equipment (CPE) and the like. In the following description, the terms “terminal device” , “communication device” , “terminal” , “user equipment” and “UE” may be used interchangeably.

As yet another example, in an Intemet of Things (lOT) scenario, a terminal device may represent a machine or other device that performs monitoring and/or measurements, and transmits the results of such monitoring and/or measurements to another terminal device and/or network equipment. The terminal device may in this case be a machine-to-machine (M2M) device, which may in a 3GPP context be referred to as a machine-type communication (MTC) device. As one particular example, the terminal device may be a UE implementing the 3GPP narrow band internet of things (NB-IoT) standard. Examples of such machines or devices are sensors, metering devices such as power meters, industrial machinery, or home or personal appliances, for example refrigerators, televisions, personal wearables such as watches etc. In other scenarios, a terminal device may represent a vehicle or other equipment that is capable of monitoring and/or reporting on its operational status or other functions associated with its operation.

FIG. 1 illustrates an example wireless communication network 100 in which embodiments of the present disclosure may be implemented. As shown in FIG. 1, the wireless communication network 100 may include one or more network devices (also referred to as network nodes) , for example, a network device 101, which may be in a form of an eNB or gNB. It will be appreciated that the network device 101 can also be in a form of a NB, Base Transceiver Station (BTS) , and/or Base Station Subsystem (BSS) , AP and the like. The network device 101 provides radio connectivity to a set of terminal devices, e.g., terminal device 102. Both the network device 101 and the terminal device 102 are equipped with a transmitter and a receiver (or a transceiver) to enable communication between them.

Power amplifier (PA) is an important component in a transmitter (or a transceiver) , and it has to be well designed to enable efficient communication. Due to high efficiency, class-AB and class C PA have been widely used in transmitters/transceivers. However, high efficiency comes with high nonlinearity which may cause degraded system performance and is not desired.

By compensating the nonlinearity of the PA, system performance may be improved. DPD has been considered as a candidate for the compensating. In a PA system with DPD, an input signal may be pre-distorted before entering the PA, and in this way, the distortion at the output of the PA may be corrected.

A challenge with DPD technology is that the distortion (i.e., non-linearity) characteristics of the PA may vary (e.g., with time, temperature, and biasing) , and therefore it may not be easy to determine a proper parameter/algorithm of the DPD operation.

In DPD realm, a conventional method for designing the DPD parameter/algorithm is to use Volterra series, some variant thereof, or a combination of the Volterra series and other techniques (e.g., orthogonal processing) . However, these methods are usually very complicated and have low capability for solving nonlinearity fitting problems.

Another way for determining the DPD parameter is to use a feedback mechanism, i.e., sample an output signal of the PA and use it to correct the parameter of the DPD. This mechanism makes use of input training reference data and output reference data. The output reference data may be gathered from feedback of the PA which is noisy and nonlinear. With such feedback mechanism, the noise in the feedback may cause inaccurate estimation of the transfer characteristic of the PA, and result in improper DPD design.

In some embodiments of the present disclosure, it is proposed to train the DPD based on an ANN. For low-pass equivalent behavioral modeling of wireless transmitters, both the ANN and the Volterra series have received particular interest from microwave community. Inventor of the present disclosure has observed that ANN has much stronger fitting capability than Volterra, but does not perform well in a noisy scenario. Though some techniques e.g., regulation used in ANN, may be used to suppress sensitivity of ANN to noise, performance of which cannot meet requirements of DPD applications.

To solve this problem and other similar problems, in some embodiments of the present disclosure, it is proposed to use clean training data for ANN. With the clean training data, the ANN based DPD is more suitable for wide bandwidth nonlinearity applications.

In some embodiments, original noisy training data (which may be obtained from the feedback of the PA) may be pretreated (e.g., via polynomial fitting) , to construct new clean training data. This scheme overcomes weakness of the ANN in noisy scenarios while maintaining its advantage of non-linearity fitting ability. In some further embodiments, the new training data may be reconstructed based on some optimization criteria e.g., a LS criterion.

As an example rather than limitation, in some embodiments, amplitude (AM) vs AM and AM vs PM curves of the output reference data relative to the input training reference data are calculated firstly. Then a regression method e.g., LS based polynomial fitting, may be used to fit these curves. The fitted polynomial may be used to reconstruct new output reference data which has no or suppressed noise and keeps the non-linearity characteristic. Note that, many other fitting functions (e.g., segment fitting) may be used for this purpose in some other embodiments.

The reconstructed new output reference data is used to train the ANN (e.g., a delay tapped back propagation (BP) based ANN) , so as to determine a proper parameter of the DPD. Since noise is suppressed in the reconstructed new output reference data, the number of neuron in the ANN can be chosen higher to reach better performance without causing over-fitting.

To facilitating understanding of the solutions proposed herein, some embodiments will be described below with reference to FIGs. 2-13.

FIG. 2 shows an example method 200 according to an embodiment of the present disclosure. The method may be implemented by a training apparatus which may be implemented, for example, in a transceiver of the network device 101 or the terminal device 102 in FIG. 1, or may provide an input to the transceiver. However, it should be appreciated that the method 200 may also be implemented by other device, apparatus, or a cloud for data processing. Just for illustration purpose, and without limitation, the method 300 will be described below with reference to a training apparatus.

As shown in FIG. 2, at block 210, the training apparatus obtains input reference data and first output reference data for training an ANN. Note that embodiments are not limited to any specific application of the ANN. Just for illustration rather than limitation, the ANN may be used for determining configuration/parameter for DPD in a PA. In such embodiments, at block 210, the training apparatus may obtain training data input to the PA as the input reference data; and obtain feedback data output from the PA in response to the training data as the first output reference data.

In addition, embodiments are not limited to any specific structure of the ANN. Just for illustration, FIG. 3 shows a diagram of a delay-tapped BP ANN schematically, however, it should be appreciated that embodiments of the present disclosure are not limited thereto. The example ANN shown in FIG. 3 comprises a plurality of neurons (denoted as small circles in FIG. 3) . In addition, tapped-delay lines (denoted by symbol v in FIG. 3) are employed in neurons of inputs to simulate a memory effect of a PA. In FIG. 3, I _in and Q _in are inputs, and I _out and Q _out are outputs of the ANN. Though only one hidden layer is shown in the example ANN in FIG. 3, it should be appreciated that the ANN may include a plurality of hidden layers in some embodiments. Further, symbol b in FIG. 3 represents a threshold value, f represents an activation function where sigmoid function may be used, and w represents coefficients of the ANN model to be learnt via training.

At block 220, the training apparatus generates second output reference data by suppressing noise in the first output reference data. In some embodiments, the first output reference data may be gathered from feedback of a PA and may include noise. In such a case, a relation between the input reference data and first output reference data cannot reflect transfer characteristic of the PA accurately. By suppressing noise in the first output reference data, the second output reference data generated at block 220 is cleaner and more suitable for training the ANN.

Embodiments are not limited to any specific way for suppressing the noise in the first output reference data in order to obtain the clean second output reference data at block 220. In other words, any proper pretreatment or preprocessing already known or to be developed in the future may be used for this purpose. Just for illustration, without limitation, in some embodiments, the training apparatus may generate the second output reference data through polynomial fitting based on the input reference data and the first output reference data. For instance, at block 220, the training apparatus may generate the second output reference data through polynomial fitting based on a LS criterion.

FIG. 4 shows an example of reconstructing the second output reference data via polynomial fitting. Specifically, an AM-AM curve 410 of the first output reference data (which is obtained at block 210, and may be referred to as the original output) relative to the input reference data (which is also obtained at block 210 and may be referred to as the original input) , and an AM-AM curve 420 reconstructed via polynomial fitting of the curve 410 are shown. In FIG. 4, the horizontal axis stands for an amplitude of the input reference data (which may be denoted as A_I herein) , and the vertical axis stands for an amplitude of the output reference data (which may be denoted as A_O herein) . As shown in FIG. 4, due to noises in the first output reference data, the black dots forming the AM-AM curve 410 are dispersed. In contrast, the AM-AM curve 420, which is reconstructed by polynomial fitting of the curve 410, is thinner, which implies that noises are suppressed. The second output reference data can be derived from the AM-AM curve 420 directly.

Alternatively or in addition, in some embodiments, the training apparatus may determine an amplitude and a phase of the second output reference data respectively, based on the input reference data and the first output reference data, and generate the second output reference data based on the determined amplitude and phase.

As an example rather than limitation, at block 220, the training apparatus may determine the amplitude of the second output reference data through a polynomial fitting based on an amplitude of the first output reference data relative to the input reference data, e.g., based on an AM-AM gain curve of the first output reference data. Likewise, the training apparatus may determine the phase of the second output reference data through a polynomial fitting based on a phase of the first output reference data relative to the input reference data, e.g., based on an AM-PM gain curve of the first output reference data.

FIGs. 5-6 show an example of reconstructing the second output reference data via polynomial fitting based on an AM-AM gain curve and an AM-PM gain curve of the first output reference data.

In particular, FIG. 5 shows a curve 510 of AM gain of the first output reference data relative to the input reference data, and a curve 520 of AM gain of the second output reference data (which is obtained at block 220 and may be referred to as reconstructed training data) relative to the input reference data. In FIG. 5, the horizontal axis stands for the amplitude of the input reference data (which may be denoted as A_I herein) , and the vertical axis stands for a gain in amplitude which may be represented as G_A=|A_O/A_I|. By polynomial fitting of the curve 510, the curve 520, and correspondingly the amplitude for the second output reference data, is obtained. Obviously, the gain in amplitude shown by curve 520 is more definite than that shown by curve 510, which implies suppressed noise in the reconstructed second output reference data.

Likewise, FIG. 6 shows a curve 610 of PM gain of the first output reference data relative to the input reference data, and a curve of 620 PM gain of the reconstructed training data relative to the input reference data. In FIG. 6, the horizontal axis stands for the amplitude of the input reference data (i.e., A_I) , and the vertical axis stands for a gain in phase which may be represented as G_P=phase (A_O/A_I) . By polynomial fitting of the curve 610, the curve 620, and correspondingly phase for the second output reference data, is obtained. Obviously, the gain in phase shown by curve 620 is more definite than that shown by curve 610, which also shows suppressed noise in the reconstructed second output reference data. Then based on the amplitude for the second output reference data in FIG. 5 and the phase for the second output reference data in FIG. 6, the second output reference data, i.e., the clean training data, is determined.

As another alternative, at block 220, the training apparatus may generate the second output reference data via operations 700 shown in FIG. 7. Specifically, in the example shown in FIG. 7, the training apparatus may determine an inphase (I) component of the second output reference data based on the input reference data and the first output reference data at block 710, determine a quadrature (Q) component of the second output reference data based on the input reference data and the first output reference data at block 720; and generate the second output reference data based on the determined I and Q components at block 730. Note that in some embodiments, each of the I and Q components may be generated in a way similar to that described with reference to FIGs. 4-6.

Now refer back to FIG. 2. At block 230, the training apparatus trains the ANN based on the input reference data and the second output reference data which is generated at block 220 and cleaner than the original first output reference data. For illustration rather than limitation, a criterion for training the ANN may include minimizing a sum of squared error of target data and the outputs of ANN.

In some embodiments, the trained ANN may be used to determining a configuration/parameter of DPD which may be applied to a PA. That is, in some example embodiments, the method 200 may further comprise a block 240 where the training apparatus determines the configuration/parameter for DPD based on the trained ANN.

FIG. 8 shows an example for configuring the DPD in a PA system based on an ANN according to an embodiment of the present disclosure. The ANN for configuring the DPD may be trained, for example, using method 200. As shown in FIG. 8, data gathered from the feedback chain (which may include an attenuator 802, an IQ modulator 803, and ADCs 804 and 805) of the PA 801 is input to a preprocessing module 806 to generate clean training data before entering the ANN 807. The feedback data input to the pre-processing module 806 may be represented by an I component I_out and a Q component Q_out. For instance, the pre-processing module 806 may generate the clean training data with I component I_out_cln and Q component Q_out_cln using operations described with reference to block 220 of method 200, using the feedback data I_out and Q_out as the first output reference data. As shown in FIG. 8, the clean training data output from the pre-processing module 806 is input to the ANN 807, together with input reference data with I component I_in and Q component Q_in, for training the ANN 807. Any proper criterion known or to be developed in the future may be used for the training, and embodiments are not limited to any specific training algorithm. In some embodiments, operations similar to that described with reference to block 230 of method 200 may be used for the training.

Then the trained ANN 807 may be used to determine a parameter/coefficient for the DPD 808 based on the input reference data (I_in and Q_in) which may be obtained from the input side of the PA 801, for example before the IQ modulator 809. As shown in FIG. 8, a copy of the determined coefficient (Coeff) by the ANN 807 is applied to the DPD 808.

FIG. 9 shows AM-AM characteristic of the DPD which is configured based on an ANN trained with clean data, according to an embodiment of the present disclosure. Compared with conventional DPD, the AM-AM characteristic of the DPD shown in FIG. 9 is more accurate.

The accurate transfer characteristic of the DPD designed according to embodiments of the present disclosure results in better performance of the PA system, as shown in FIGs. 10-12. For comparison, FIG. 10 shows original spectrum of a PA system without DPD. It can be observed from FIG. 10 that out-band attenuation is about -70dBm, which is only about 25dBm lower than the in-band response, which means strong out-band interference.

FIG. 11 shows spectrum of a PA system with conventional DPD. It can be observed that the out-band attenuation is about -90dBm, which means reduced out-band interference compared with FIG. 10. FIG. 12 shows spectrum of a PA system with DPD designed according to an embodiment of the present disclosure. In this case, the out-band attenuation is reduced to -100dBm, which means even lower out-band interference than the PA system with conventional DPD shown in FIG. 11.

Though some embodiments are described with reference to DPD and a PA system, it should be appreciated that embodiments proposed herein are not limited to such specific application scenarios. Instead, proposed solutions for obtaining clean training data for an ANN via pre-processing may be applied to any application where similar problem exists, and/or clean training data is desired.

Note that in some embodiments, the training apparatus implementing the method 200 may be a part of an ANN. In another embodiment, the training apparatus may be a separate apparatus which may be connected to the ANN when needed.

Alternatively, or in addition, the ANN and/or the training apparatus may be a part of a DPD module. In another embodiment, the ANN and/or the training apparatus may be connected to the DPD module only when needed.

In some embodiments, the ANN, the training apparatus, and/or the DPD module may be a part of a PA system. In another embodiment, the ANN, the training apparatus, and/or the DPD module may be connected to the PA system only when needed.

Some embodiments of the present disclosure further propose a device for communication, which may include a network device (e.g., network device 101 in FIG. 1) or a terminal device (e.g., the terminal device 102 in FIG. 1) . The device for communication comprises a PA, and a DPD couple to an input of the PA. In addition, a parameter of the DPD is obtained based on an ANN which is trained with input reference data and an output reference data, and the output reference data is generated by suppressing noise in feedback data output from the PA, for example according to method 200.

FIG. 13 illustrates a simplified block diagram of an apparatus 1300 that may be embodied in/as a communication device which may include, but is not limited to, a network device or a terminal device. In some embodiments, the apparatus 1300 may be separate from the communication device and may be connected to the communication device when needed.

As shown by the example of FIG. 13, apparatus 1300 comprises a processor 1310 which controls operations and functions of apparatus 1300. For example, in some embodiments, the processor 1310 may implement various operations by means of instructions 1330 stored in a memory 1320 coupled thereto. The memory 1320 may be any suitable type adapted to local technical environment and may be implemented using any suitable data storage technology, such as semiconductor based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory, as non-limiting examples. In some example embodiments, the memory 1320 can be a non-transitory computer readable medium. Though only one memory unit is shown in FIG. 13, a plurality of physically different memory units may exist in apparatus 1300 in some embodiments.

The processor 1310 may be any proper type adapted to local technical environment, and may include one or more of general purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs) , central processing units (CPUs) , field-programmable gate arrays (FPGA) , application specific circuits (ASIC) , GPUs (Graphics Processing Unit) , NPUs (Neural Network Processing Unit) , AI (Artificial Intelligence) accelerators and processors based on multicore processor architecture, as non-limiting examples. The apparatus 1300 may also comprise a plurality of processors 1310 in any combination thereof.

The processors 1310 may also be coupled with one or more transceiver 1340 which enables communication with other apparatus, modules, or devices. In some embodiments, the processor 1310 and the memory 1320 may operate in cooperation to implement method 200 described with reference to FIGs. 2-7. It shall be appreciated that all the features described above with reference to FIGs. 2-12 may also apply to apparatus 1300, and therefore will not be detailed here.

Various embodiments of the present disclosure may be implemented by a computer program or a computer program product executable by one or more of the processors (for example processor 1310 in FIG. 13) , software, firmware, hardware or in a combination thereof.

Although some embodiments are described in the context of DPD and PA, it should not be construed as limiting the spirit and scope of the present disclosure. The principle and concept of the present disclosure may be more generally applicable to other application scenarios.

In addition, the present disclosure may also provide a carrier containing the computer program as mentioned above (e.g., computer instructions/grogram 1330 in FIG. 13) . The carrier includes a computer readable storage medium. The computer readable storage medium may include, for example, an optical compact disk or an electronic memory device like a RAM (random access memory) , a ROM (read only memory) , Flash memory, magnetic tape, CD-ROM, DVD, Blue-ray disc and the like.

The techniques described herein may be implemented by various means so that an apparatus implementing one or more functions of a corresponding apparatus described with an embodiment comprises not only prior art means, but also means for implementing the one or more functions of the corresponding apparatus and it may comprise separate means for each separate function, or means that may be configured to perform two or more functions. For example, these techniques may be implemented in hardware (e.g., circuit or a processor) , firmware, software, or combinations thereof. For a firmware or software, implementation may be made through modules (e.g., procedures, functions, and so on) that perform the functions described herein.

Some example embodiments herein have been described above with reference to block diagrams and flowchart illustrations of methods and apparatuses. It will be appreciated that each block of the block diagrams and flowchart illustrations, and combinations of blocks in the block diagrams and flowchart illustrations, respectively, may be implemented by various means including computer program instructions. These computer program instructions may be loaded onto a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions which execute on the computer or other programmable data processing apparatus create means for implementing the functions specified in the flowchart block or blocks.

While this specification contains many specific implementation details, these should not be construed as limitations on the scope of any implementation or of what may be claimed, but rather as descriptions of features that may be specific to particular embodiments of particular implementations. Certain features that are described in this specification in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable sub-combination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a sub-combination or variation of a sub-combination.

It will be obvious to a person skilled in the art that, as the technology advances, the inventive concept may be implemented in various ways. The above described embodiments are given for describing rather than limiting the disclosure, and it is to be understood that modifications and variations may be resorted to without departing from the spirit and scope of the disclosure as those skilled in the art readily understand. Such modifications and variations are considered to be within the scope of the disclosure and the appended claims. The protection scope of the disclosure is defined by the accompanying claims.

Claims

An apparatus for data processing, comprising:

at least one processor; and

at least one memory including computer program codes;

the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus at least to:

obtain input reference data and first output reference data for training an artificial neural network, ANN;

generate second output reference data by suppressing noise in the first output reference data; and

train the ANN based on the input reference data and the second output reference data.
The apparatus of Claim 1, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, further cause the apparatus to:

determine, based on the trained ANN, a parameter for digital pre-distortion, DPD, to be applied to a power amplifier, PA.
The apparatus of Claim 2, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, further cause the apparatus to:

obtain the input reference data and the first output reference data further comprising:

obtain training data input to the PA as the input reference data; and

obtain feedback data output from the PA in response to the training data as the first output reference data.
The apparatus of Claim 1, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus to

generate the second output reference data further comprising:

generate the second output reference data through polynomial fitting based on the input reference data and the first output reference data.
The apparatus of Claim 4, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus to:

generate the second output reference data further comprising:

generate the second output reference data based on a least square, LS, criterion.
The apparatus of any of Claims 1 to 5, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus to:

generate the second output reference data further comprising:

determine an amplitude and a phase of the second output reference data respectively, based on the input reference data and the first output reference data; and

generate the second output reference data based on the determined amplitude and phase.
The apparatus of Claim 6, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus to:

determine the amplitude and the phase of the second output reference data further comprising:

determine the amplitude through polynomial fitting based on an amplitude of the first output reference data relative to the input reference data; and

determine the phase through polynomial fitting based on a phase of the first output reference data relative to the input reference data.
The apparatus of any of Claims 1 to 5, wherein the at least one memory and the computer program codes are configured to, with the at least one processor, cause the apparatus to:

generate the second output reference data further comprising:

determine an inphase component and a quadrature component of the second output reference data respectively, based on the input reference data and the first output reference data; and

generate the second output reference data based on the determined inphase and quadrature components.
A method of data processing, comprising:

obtaining input reference data and first output reference data for training an artificial neural network, ANN;

generating second output reference data by suppressing noise in the first output reference data; and

training the ANN based on the input reference data and the second output reference data.
The method of Claim 9, further comprising:

determining, based on the trained ANN, a parameter for digital pre-distortion, DPD, to be applied to a power amplifier, PA.
The method of Claim 10, wherein obtaining the input reference data and the first output reference data comprises:

obtaining training data input to the PA as the input reference data; and

obtaining feedback data output from the PA in response to the training data as the first output reference data.
The method of Claim 9, wherein generating the second output reference data further comprises:

generating the second output reference data through polynomial fitting based on the input reference data and the first output reference data.
The method of Claim 12, wherein generating the second output reference data further comprises:

generating the second output reference data through polynomial fitting based on a least square, LS, criterion.
The method of any of Claims 9 to 13, wherein generating the second output reference data comprises:

determining an amplitude and a phase of the second output reference data respectively, based on the input reference data and the first output reference data; and

generating the second output reference data based on the determined amplitude and phase.
The method of claim 14, wherein determining the amplitude and the phase of the second output reference data comprises:

determining the amplitude through polynomial fitting based on an amplitude of the first output reference data relative to the input reference data; and

determining the phase through polynomial fitting based on a phase of the first output reference data relative to the input reference data.
The method of any of Claims 9 to 13, wherein generating the second output reference data comprises:

determining an inphase component and a quadrature component of the second output reference data respectively, based on the input reference data and the first output reference data; and

generating the second output reference data based on the determined inphase and quadrature components.
An apparatus for data processing, comprising:

means for obtaining input reference data and first output reference data for training an Artificial neural network, ANN;

means for generating second output reference data by suppressing noise in the first output reference data; and

means for training the ANN based on the input reference data and the second output reference data.
The apparatus of claim 17, wherein the means comprises

at least one processor; and

at least one memory including computer program code, the at least one memory and computer program code configured to, with the at least one processor, cause the performance of the apparatus.
A computer readable medium having a computer program stored thereon which, when executed by at least one processor of a device, causes the device to carry out the method of any of claims 9-16.
A device for communication, comprising:

a power amplifier, PA; and

a digital pre-distortion, DPD, coupled to an input of the PA;

where a parameter of the DPD is obtained based on an artificial neural network, ANN, the ANN being trained with input reference data and an output reference data; and

wherein the output reference data is generated by suppressing noise in feedback data output from the PA.