Description:TECHNICAL FIELD
The present disclosure relates to wireless networks, and more specifically relates to methods and wireless networks (e.g., Open Radio Access Network (O-RAN)) for handling digital compensation for radio frequency (RF) power amplifier nonlinearities.

BACKGROUND
In an Open Radio unit (O-RU), power consumption can be very high due to radio-frequency power amplifiers (PAs). Further, a high peak-to-average power signal imposes a strict linearity requirement on the radio-frequency power amplifiers. Therefore, it is required to operate the radio-frequency power amplifiers in a saturation region for the highest efficiency, but it produces distortions. Hence, it is essential to compensate for such distortions using signal processing techniques. An additional challenge arises in practice where there is often a large amount of missing data, especially for multi-pollutant monitoring data. Thus, a separate imputation step is required prior to dimension reduction. Some of the prior art references are given below for handling digital compensation for the RF power amplifier nonlinearities:
US10523159B2 discloses a digital compensator for a non-linear system to linearize a non-linear power amplifier. The non-linear system describes the effective parametrization of a digital pre-distorter that is used for digital compensation.
WO2021148061A2 discloses a Digital Predistortion (DPD) device for a fifth generation (5G) broadband multiple input multiple output (MIMO) system. WO2021148061A2 describes the implementation of a new DPD model algorithm based on a dynamic deviation dimensionality reduction method. The dynamic deviation dimensionality reduction method proposes an effective order reduction method, which removes the high-order dynamic memory effect. The new DPD model algorithm reduces model complexity while maintaining high accuracy, saves hardware logic resources, and accelerates system response speed.
US20020080891A1 discloses a base station transmitter having a digital pre-distortion unit to perform a predistortion method. US20020080891A1 describes that if the digital signal pre-distorted by the pre-distortion unit is delayed for a predetermined time, an HPA modeling unit extracts an unknown coefficient of the non-linear characteristic model of the power amplifier using the digital signals outputted from the analog/digital converter and the delayed digital input signals.
A non-patent literature entitled “Digital Predistortion for Power Amplifier Based on Sparse Bayesian Learning” discloses a sparse-Bayesian-learning algorithm that is applied to estimate the coefficients of the power amplifier (PA) behavioral models and inverse models from the view of probability.
Another non-patent literature entitled “The Contribution to Dimensionality Reduction of Digital Predistorter Behavioral Models for RF Power Amplifier Linearization” discloses a DPD model to mitigate non-linearities in RF power amplifiers. Further, it discloses dimensionality reduction techniques to reduce the order of the DPD model.
Yet another non-patent literature entitled “The Independent Digital Predistortion Parameters Estimation Using Adaptive Principal Component Analysis” discloses an estimation/adaptation method based on the adaptive principal component analysis (APCA) technique to guarantee the identification of the minimum necessary parameters of a digital predistorter.
Yet another non-patent literature entitled “The Application of principal component analysis based effective digital predistortion technique for low-cost FPGA implementation” describes the impact of principle component analysis (PCA) on the bit-resolution of DPD solutions within the context of established DPD models.
Further, FIG. 1 illustrates a general block diagram of a DPD block (100), according to the prior art. The digital predistortion block (100) includes a complex gain adjustment unit (102), a digital to analog converter (DAC) (104), a modulation and up-conversion unit (106), a power amplifier (PA) (108), a demodulation and down conversion unit (110), an analog to digital converter (ADC) (112), an adaptation logic (114), a look-up table (116) and a delay logic (118).
The complex gain adjustment unit (102) processes a predistorted baseband input signal (vi) and then feeds the processed pre-distorted baseband input signal (vi) to the DAC (104). The DAC (104) performs digital-to-analog conversion on the processed pre-distorted baseband input signal that is further transmitted to the modulation and up-conversion unit (106). The modulation and up-conversion unit (106) modulates analog signals received from the DAC (104) and outputs the modulated analog signals as RF signals. The PA (108) amplifies the power of the RF signals received from the modulation and up-conversion unit (106). The PA (108) transmits and outputs the RF signal through the antenna.
The demodulation and down conversion unit (110) demodulates signals sampled from the RF signals amplified by the power amplifier (108). The ADC (112) converts analog signals outputted from the demodulation and down conversion unit (110). The adaptation logic (114) extracts an error function using the demodulated signal, reference signals, and feedback signal (vfb) and produces a coefficient from the predistorted baseband input signal (vi). Also, the adaptation logic (114) controls characteristics of the predistortion adaptively using an error function. The delay logic (118) delays the pre-distorted phase digital input signals, sampled from outputs of the adaptation logic (114), for a predetermined time. The error function is stored in the look-up table (116).
FIG. 2 illustrates an example graph (200) of an instantaneous power level of a signal with high peak power (Ppeak) compared to its average power (Pavg), according to prior art. This graph shows the peak power is higher than the average power and this peak saturates the PA and causes nonlinearities.
FIG. 3 illustrates an example plot (300) of a PA output that when an input signal is applied to it, then we get a distorted output along with harmonics of the actual signal according to the prior art.
FIG. 4 illustrates an example pictorial representation (400) of an output of an RF-power amplifier, according to the prior art. The output of the RF-power amplifier has a distorted fundamental as well as harmonics and performance mitigation in case of missing data during modeling.
While the prior arts cover various approaches for handling digital compensation for the RF power amplifier nonlinearities, none of the prior arts discloses techniques to enhance dimension reduction procedure under the presence of missing data to further improve the performance of DPD. In light of the above-stated discussion, there is a need to overcome the above stated disadvantages.

OBJECT OF THE DISCLOSURE
A principal object of the present disclosure is to provide a method and an O-RU for handling digital compensation for radio frequency (RF) power amplifier nonlinearities in an Open Radio Access Network (O-RAN).
Another object of the present disclosure is to enhance a dimension reduction procedure under the presence of missing data using an inverse modelling technique with Probabilistic Principal Component Analysis (PPCA) to compensate for the impact of missing data (if any) to further improve the performance of DPD (Digital Pre-Distortion).

SUMMARY
Accordingly, the present disclosure provides methods and an O-RU for handling digital compensation for radio frequency (RF) power amplifier nonlinearities in an Open Radio Access Network (O-RAN). The method includes receiving a baseband output signal z(n) by a power amplifier (PA). The baseband output signal z(n) is defined by a first predefined dimension. Further, the method includes comparing the baseband output signal z(n) with a baseband input signal x(n) to obtain one or more missing values m(n). The baseband input signal x(n) is received at the PA. Further, the method includes reconstructing the baseband output signal z(n) by adding one or more missing values m(n) to form a reconstructed signal g(n). The reconstructed signal g(n) is defined by the first predefined dimension, wherein the reconstructed signal g(n) is a linear distortion-free signal. Further, the method includes re-dimensioning the reconstructed signal g(n) to form a refined reconstructed signal g’(n), wherein the refined reconstructed signal g’(n) is defined by a second predefined dimension. The refined reconstructed signal g’(n) is a linear distortion free signal. The second predefined dimension is less than the first predefined dimension.
These and other aspects herein will be better appreciated and understood when considered in conjunction with the following description and the accompanying drawings. It should be understood, however, that the following descriptions are given by way of illustration and not of limitation. Many changes and modifications may be made within the scope of the invention herein without departing from the spirit thereof.

BRIEF DESCRIPTION OF FIGURES
The invention is illustrated in the accompanying drawings, throughout which reference letters indicate corresponding parts in the drawings. The invention herein will be better understood from the following description with reference to the drawings, in which:
FIG. 1 illustrates a general block diagram of a DPD block, according to the prior art.
FIG. 2 illustrates an example graph of an instantaneous power level of a signal with high peak power (Ppeak) compared to its average power (Pavg), according to prior art.
FIG. 3 illustrates an example graph of power efficiency as a function of output power for a power amplifier (PA), according to the prior art.
FIG. 4 illustrates an example pictorial representation of an output of an RF-power amplifier, according to the prior art.
FIG. 5 illustrates a block diagram of an O-RU having an enhanced DPD model performance in an O-RAN, according to the present disclosure.
FIG. 6 illustrates the working of a power amplifier in combination with the pre-distorter training and PPCA unit with DPD models explained in connection with FIG. 5.
FIG. 7 is a flow chart illustrating a method of enhancing DPD model performance in the O-RAN.
FIG. 8 is a flow chart illustrating various operations to perform DPD coefficient extraction and dimension reduction for enhancing the DPD model performance in the O-RAN.
FIG. 9 illustrates an example graph of variation of normalized mean square error (NMSE) with the percentage of missing data (64 Bit).
FIG. 10 illustrates an example graph of variation of NMSE with the percentage of missing data (16 Bit).
FIG. 11 illustrates an example spectrum plot for inverse modelling.
FIG. 12 illustrates an example graph of amplitude-amplitude distortion (AM/AM) with respect to input power.
FIG. 13 illustrates an example graph of amplitude-phase distortion (AM/PM) with respect to input power.

DETAILED DESCRIPTION
In the following detailed description of the invention, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be obvious to a person skilled in the art that the invention may be practiced with or without these specific details. In other instances, well known methods, procedures and components have not been described in details so as not to unnecessarily obscure aspects of the invention.
Furthermore, it will be clear that the invention is not limited to these alternatives only. Numerous modifications, changes, variations, substitutions and equivalents will be apparent to those skilled in the art, without parting from the scope of the invention.
The accompanying drawings are used to help easily understand various technical features and it should be understood that the alternatives presented herein are not limited by the accompanying drawings. As such, the present disclosure should be construed to extend to any alterations, equivalents and substitutes in addition to those which are particularly set out in the accompanying drawings. Although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are generally only used to distinguish one element from another.
The present disclosure provides methods for enhancing Digital Pre-Distortion model performance in an O-RAN. The method includes receiving, by an O-RU, a baseband output signal z(n) by a power amplifier (PA). The baseband output signal z(n) is defined by a first predefined dimension. Further, the method includes comparing, by the O-RU, the baseband output signal z(n) with a baseband input signal x(n) to obtain one or more missing values m(n), wherein the baseband input signal x(n) is received at the PA. Further, the method includes reconstructing, by the O-RU, the baseband output signal z(n) by adding the one or more missing values m(n) to form a reconstructed signal g(n), wherein the reconstructed signal g(n) is defined by the first predefined dimension, wherein the reconstructed signal g(n) is a linear distortion free signal.
The deficiency/distortion in previous techniques (as shown in FIG. 1 to FIG. 4 and background) can be solved by the enhanced/improved DPD model performance in the O-RAN. The distortion is introduced in the power amplifier(s) which implies the variation in the waveform received at the output with respect to the applied input. That is, the unwanted alterations generated during amplification are known as distortion. The DPD model performance is enhanced by using an inverse modelling method with Probabilistic Principal Component Analysis (PPCA).
The PPCA is a technique to estimate principal axes when any data vector has one or more missing values. The PPCA assumes that the values are missing at random through the data set. An expectation-maximization (EM) technique is used for both complete and missing data. The EM technique is an iterative method to find (local) maximum likelihood or Maximum A Posteriori (MAP) estimates of parameters in statistical models, where the model depends on unobserved latent variables. The method can be used to enhance the dimension reduction procedure in the presence of missing data. The method can be used to compensate for the impact of missing data (if any) to improve the performance of DPD. The method is implemented in a baseband unit of O-RU that does not need any command form O-DU and O-CU, which saves power.
In other words, the proposed method is purely a baseband technique. The baseband technique is implemented in a baseband unit to get a tunned signal from the baseband as input to a converter. The proposed method can also be used for FR-2, and O-RU. The proposed method tackles the issue of DC offset and IQ (in-phase (I) and quadrature (Q)) imbalance.
FIG. 5 illustrates a block diagram of an O-RU (500) having an enhanced DPD model performance in an O-RAN, according to the present disclosure. The O-RU (500) converts radio signals sent to and from an antenna (not shown) to a digital signal that is transmitted over a fronthaul (not shown) to a Distributed Unit (DU) (not shown) in the O-RAN. The O-RU (500) comprises a pre-distorter (502), an RF transmitter (504), a PA (506), a 1/G (508), a low noise amplifier (LNA) (510), an RF receiver (512), and a pre-distorter training and PPCA unit (514). A baseband input signal x(n) is fed to the pre-distorter (502) for Digital Pre-Distortion and received at the PA (506). The baseband input signal x(n) is generated from a baseband that contains the current and past samples of the cartesian In-phase and Quadrature (I/Q) components along with the envelope-dependent terms. The baseband input signal is transmitted through a frequency multiplier (not shown) and PA chain. In general, the DPD (Digital Pre-Distortion) is a technique to increase linearity or compensate for non-linearity in the power amplifier (506). The Digital Pre-Distortion applies inverse distortion, using the pre-distorter (502), at the input signal of the PA (506) to cancel the distortion generated by the power amplifier (506). The PA (506) produces a baseband output signal z(n) from the input signal received from the pre-distorter (502) through the RF transmitter (504). The pre-distorter training and PPCA unit (514) receives the baseband output signal z(n) from the PA (506) through the 1/G (508), the LNA (510), and the RF receiver (512), where the baseband output signal z(n) is defined by a first predefined dimension. The 1/G (508) provides a linear output of the baseband output signal z(n) and transmits the same to the LNA (510). The LNA (510) amplifies a very low-power signal without significantly degrading signal-to-noise ratio of the signal received from the 1/G (508) and transmits the same to the pre-distorter training and PPCA unit (514) through the RF receiver (512). Further, the pre-distorter training and PPCA unit (514) compares the baseband output signal z(n) with the baseband input signal x(n) to obtain one or more missing values m(n).
The baseband output signal z(n) is compared with the baseband input signal x(n) to obtain the one or more missing values m(n) by estimating one or more model coefficients, wherein the one or more model coefficients are associated with the baseband input signal x(n) and the baseband output signal z(n).
Alternatively, the baseband output signal z(n) is compared with the baseband input signal x(n) to obtain the one or more missing values m(n) by extracting one or more model coefficients from the baseband input signal x(n) for PA modelling, reducing the one or more model coefficients by considering relevant coefficients that are needed for accurately modelling the PA (506), and comparing the reduced one or more model coefficients with a set of baseband input signal x(n) and reference values and controlling adjustable parameters to obtain the one or more missing values m(n), wherein the adjustable parameters are checking normalize mean square error (NMSE) values.
The PA modelling is computed by the baseband output signal z(n) at the RF receiver (512) as z(n)=d_0+∑_(i=0)^M ∑_(q=0)^(N-1) b_iq x(n-i) |x(n-i)|^q +∑_(i=0)^M ∑_(q=0)^(N-1) C_iq x^* (n-i) |x^* (n-i)|^q …(1)
where M is a memory order, N is a non-linearity order of the PA (506), do is a constant term, b_iq,C_iq are the DPD model coefficients, x(n) is the PA input signal (baseband input signal) and z(n) is the PA output signal (baseband output signal).
The equation (1) can be written in matrix form as: z ⃗=W ⃗B ⃗…(2)
where B ⃗ is a vector containing the model-coefficients, W ⃗ is the observation data matrix.
The one or more missing values are estimated using the Probabilistic Principal Component Analysis by determining a correlation matrix using an observation data matrix (W ⃗ ) as … (3), wherein an eigenvalue decomposition is applied as … (4)
where O is the matrix of eigenvectors and ∧ is the diagonal matrix of eigenvalues , where the desired length of the data matrix is determined using the weight of the eigenvalue of data variance, and the principal components are defined as using S eigenvectors having all important data and after discarding the insignificant data.
G= [o_1,o_2, o_3,….o_s ] … (5)
Further, the data is projected to a new low-dimensional space using principal component G.
Hence, the new observation data matrix is:
…… (6)
Hence, Z ⃗=V ⃗C ⃗, Where C ⃗ is the reduced coefficient vector.
Further, the baseband output signal z(n) is received at the RF-receiver (512) as the baseband input signal by computing one or more model coefficients using a least square (LS) technique, wherein the LS technique is defined by normalizing mean square error (NMSE) values for new dimension data, removing coefficient from the new dimension data, comparing original data with a new set of data, and computing missing value by eigenvalues , wherein the reduced data is the new set of data which is obtained by the application of PPCA, wherein the eigenvalues define reduction of the dimensionality of the data. The one or more model coefficients are computed using the least mean square (LMS) technique by comparing the new set of data with the input baseband signal x(n), wherein the LMS technique is used to compute a vector containing the model-coefficients B ⃗=(W ⃗^H W ⃗ ) W ⃗^H z ⃗, where, W is the input observation data matrix.
The pre-distorter training and PPCA unit (514) reconstructs the baseband output signal z(n) by adding one or more missing values m(n) and producing a reconstructed signal g(n). The baseband output signal z(n) is reconstructed by reducing the one or more model coefficients and the one or more missing values m(n) from the baseband output signal z(n) to obtain the reconstructed signal g(n). The reconstructed signal g(n) is defined by the first predefined dimension, wherein the reconstructed signal g(n) is a linear distortion free signal.
Further, the pre-distorter training and PPCA unit (514) re-dimensions the reconstructed signal g(n) to form a refined reconstructed signal g’(n). The refined reconstructed signal g’(n) is defined by a second predefined dimension, wherein the refined reconstructed signal g’(n) is a linear distortion free signal. The second predefined dimension is less than the first predefined dimension.
FIG. 6 illustrates the working of the power amplifier (506) in combination with the pre-distorter training and PPCA unit (514) with DPD models explained in connection with FIG. 5. The functions and operations of the DPD models are already explained in FIG. 5. For the sake of brevity, the operations and functions of the DPD models are not repeated in the patent disclosure.
At step 602, the pre-distorter training and PPCA unit (514) acquires the baseband complex waveforms under the appropriate drive signal. In step 604, the pre-distorter training and PPCA unit (514) performs the delay estimation and compensation for the baseband complex waveforms under the appropriate drive signal. The delay estimation is a technique for estimating the delay between two received signals which are originated from the same transmitter. At step 606, the pre-distorter training and PPCA unit (514) identifies the DPD model based on the delay estimation and compensation. At step 608, the pre-distorter training and PPCA unit (514) validates and applies the DPD model in the O-RU (500).
FIG. 7 is a flow chart (700) illustrating a method of enhancing the DPD model performance in the O-RAN. The operations (702-708) are handled by the pre-distorter training and PPCA unit (514). For the sake of brevity, the operations and functions of the pre-distorter training and PPCA unit (514) are not repeated in the patent disclosure.
At step 702, the method includes receiving the baseband output signal z(n) by the power amplifier (PA) (506). The baseband output signal z(n) is defined by the first predefined dimension. At step 704, the method includes comparing the baseband output signal z(n) with the baseband input signal x(n) to obtain the one or more missing values m(n), wherein the baseband input signal x(n) is received at the PA (506). In step 706, the method includes reconstructing the baseband output signal z(n) by adding the one or more missing values m(n) to form a reconstructed signal g(n), wherein the reconstructed signal g(n) is defined by the first predefined dimension, wherein the reconstructed signal g(n) is a linear distortion free signal. In step 708, the method includes re-dimensioning the reconstructed signal g(n) to form a refined reconstructed signal g’(n). The refined reconstructed signal g’(n) is defined by the second predefined dimension, wherein the refined reconstructed signal g’(n) is a linear distortion free signal.
Advantageously, the proposed method enhances the dimension reduction procedure under the presence of missing data by using the inverse modelling method with PPCA and compensating for the impact of missing data, if any to improve the performance of DPD. The proposed method can be used for power saving and hardware imperfections removal for FR-1 radio. The proposed method can be implemented in the O-RU only and does not need any command from other units (like O-DU, O-CU) to save power. The FR-1 defines bands in the sub-6 GHz spectrum (although 7125 MHz is the maximum) and the FR-2 defines bands in the mmWave spectrum. Because of the higher carrier frequencies in FR-2, it has higher maximum bandwidth. Bandwidths include 5-100 MHz (FR1) and 50/100/200/400 MHz (FR-2). The proposed method can be configured as per the need for Multiple-input/multiple-output (MIMO) and massive MIMO system. The proposed method can also be useful for GAN based RF power amplifiers having Long term memory (LTM) effect.
FIG. 8 is a flow chart (800) illustrating various operations to perform DPD coefficient extraction and dimension reduction for enhancing the DPD model performance in the O-RAN. At step 802, the pre-distorter training and PPCA unit (514) receives the output signal in the baseband and delay compensation. At step 804, the pre-distorter training and PPCA unit (514) determines the DPD models and extraction and reduction of coefficients. In step 806, the pre-distorter training and PPCA unit (514) compares the determined DPD models and extraction and reduction of coefficients with the set of reference values and controlling the adjustable parameters, wherein the adjustable parameters are checking normalized mean square error (NMSE) values.
FIG. 9 illustrates an example graph (900) of variation of NMSE with the percentage of missing data (64 Bit). This graph shows the performance of inverse modelling using the conventional methods and proposed method where we can see that it outperforms other methods in the case of a large dataset missing. For example, it provides good performance if there more than 20% of data is missing in the worst-case scenario.
FIG. 10 illustrates an example graph (1000) of variation of NMSE with the percentage of missing data (16 Bits). We can see a similar performance as mentioned above in the case of lower-bit implementation since the proposed method performs good.
FIG. 11 illustrates an example plot (1100) of the modelling spectrum using the proposed method where we can see that the modeled and measured spectrum is following and the error between the two is minimized.
FIG. 12 illustrates an example graph (1200) which shows the AM/AM (Amplitude modulation to Amplitude modulation) vs. input power which shows that the inverse modelling can model the nonlinearity of the PA.
FIG. 13 illustrates an example graph (1300) which shows the AM/PM (Amplitude modulation to Phase modulation) vs. input power which shows that the inverse modelling can model the nonlinearity of the PA.

Missing Data (%) Modelling NMSE
64 bit 8 bit
MP MP-PCA Present Invention MP MP-PCA Present Invention
10 -14.04 -13.78 -38.44 -13.50 -13.77 -36.76
15 -12.50 -12.30 -35.21 -11.99 -12.30 -34.52
20 -11.37 -11.19 -32.52 -10.99 -11.18 -32.19
25 -10.45 -10.26 -24.71 -9.97 -10.26 -24.65
30 -9.70 -9.49 -21.54 -9.22 -9.50 -21.52
35 -9.06 -8.82 -19.14 -8.56 -8.83 -19.12
Table 1

Missing Data (%) Model Coefficients

MP MP-PCA Present Invention
10 31 12 12
15 31 12 12
20 31 12 12
25 31 12 12
30 31 12 12
35 31 12 12
Table 2

Missing Data (%) Condition Number
64 bit 8 bit
MP MP-PCA Present Invention MP MP-PCA Present Invention
10 245.83 18.02 1364.49 93.30 18.02 1364.5
15 209.34 15.41 882.66 73.75 15.41 882.66
20 175.73 13.53 734.85 62.38 13.52 734.85
25 159.99 12.77 531.424 56.99 12.25 531.424
30 145.63 11.35 386.48 51.52 11.35 386.48
35 133.59 10.70 271.17 46.99 10.69 271.17
Table 3

Missing Data (%) Dispersion Coefficients
64 bit 8 bit
MP MP-PCA Present Invention MP MP-PCA Present Invention
10 2659.86 1578.74 17.10 12.42 Inf 15.84
15 2556.07 124.98 19.9132 13.84 75.11 20.81
20 5604.96 229.57 15.41 5.78 Inf 16.77
25 6248.33 783.76 5.62 4.01 Inf 5.52
30 1670.33 201.69 45.90 5.30 Inf 40.25
35 1090.83 558.43 7.68 7.61 Inf 7.45
Table 4

Table 1, Table 2, Table 3 and Table 4 indicate pre-distorter performance results (i.e., inverse modelling results) for comparison among a Memory Polynomial (MP), the Memory Polynomial Principal Component Analysis (MP-PCA), and the proposed method.
FIG. 11 illustrates an example spectrum plot (1100) for inverse modelling.
FIG. 12 illustrates an example graph (1200) of amplitude-amplitude distortion (AM/AM) with respect to input power.
FIG. 13 illustrates an example graph (1300) of amplitude-phase distortion (AM/PM) with respect to input power.
The various actions, acts, blocks, steps, or the like in the flow chart (700 and 800) may be performed in the order presented, in a different order or simultaneously. Further, in some implementations, some of the actions, acts, blocks, steps, or the like may be omitted, added, modified, skipped, or the like without departing from the scope of the invention.
The embodiments disclosed herein can be implemented using at least one software program running on at least one hardware device and performing network management functions to control the elements.
It will be apparent to those skilled in the art that other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention. While the foregoing written description of the invention enables one of ordinary skill to make and use what is considered presently to be the best mode thereof, those of ordinary skill will understand and appreciate the existence of variations, combinations, and equivalents of the specific embodiment, method, and examples herein. The invention should therefore not be limited by the above-described embodiment, method, and examples, but by all embodiments and methods within the scope of the invention. It is intended that the specification and examples be considered as exemplary, with the true scope of the invention being indicated by the claims.
The methods and processes described herein may have fewer or additional steps or states and the steps or states may be performed in a different order. Not all steps or states need to be reached. The methods and processes described herein may be embodied in, and fully or partially automated via, software code modules executed by one or more general purpose computers. The code modules may be stored in any type of computer-readable medium or other computer storage device. Some or all of the methods may alternatively be embodied in whole or in part in specialized computer hardware.
The results of the disclosed methods may be stored in any type of computer data repository, such as relational databases and flat file systems that use volatile and/or non-volatile memory (e.g., magnetic disk storage, optical storage, EEPROM and/or solid state RAM).
The various illustrative logical blocks, modules, routines, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. The described functionality can be implemented in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the disclosure.
Moreover, the various illustrative logical blocks and modules described in connection with the embodiments disclosed herein can be implemented or performed by a machine, such as a general purpose processor device, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components or any combination thereof designed to perform the functions described herein. A general-purpose processor device can be a microprocessor, but in the alternative, the processor device can be a controller, microcontroller, or state machine, combinations of the same, or the like. A processor device can include electrical circuitry configured to process computer-executable instructions. In another embodiment, a processor device includes an FPGA or other programmable device that performs logic operations without processing computer-executable instructions. A processor device can also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. Although described herein primarily with respect to digital technology, a processor device may also include primarily analog components. A computing environment can include any type of computer system, including, but not limited to, a computer system based on a microprocessor, a mainframe computer, a digital signal processor, a portable computing device, a device controller, or a computational engine within an appliance, to name a few.
The elements of a method, process, routine, or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor device, or in a combination of the two. A software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of a non-transitory computer-readable storage medium. An exemplary storage medium can be coupled to the processor device such that the processor device can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor device. The processor device and the storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In the alternative, the processor device and the storage medium can reside as discrete components in a user terminal.
Conditional language used herein, such as, among others, "can," "may," "might," "may," “e.g.,” and the like, unless specifically stated otherwise, or otherwise understood within the context as used, is generally intended to convey that certain alternatives include, while other alternatives do not include, certain features, elements and/or steps. Thus, such conditional language is not generally intended to imply that features, elements and/or steps are in any way required for one or more alternatives or that one or more alternatives necessarily include logic for deciding, with or without other input or prompting, whether these features, elements and/or steps are included or are to be performed in any particular alternative. The terms “comprising,” “including,” “having,” and the like are synonymous and are used inclusively, in an open-ended fashion, and do not exclude additional elements, features, acts, operations, and so forth. Also, the term “or” is used in its inclusive sense (and not in its exclusive sense) so that when used, for example, to connect a list of elements, the term “or” means one, some, or all of the elements in the list.
Disjunctive language such as the phrase “at least one of X, Y, Z,” unless specifically stated otherwise, is otherwise understood with the context as used in general to present that an item, term, etc., may be either X, Y, or Z, or any combination thereof (e.g., X, Y, and/or Z). Thus, such disjunctive language is not generally intended to, and should not, imply that certain alternatives require at least one of X, at least one of Y, or at least one of Z to each be present.
While the detailed description has shown, described, and pointed out novel features as applied to various alternatives, it can be understood that various omissions, substitutions, and changes in the form and details of the devices or algorithms illustrated can be made without departing from the scope of the disclosure. As can be recognized, certain alternatives described herein can be embodied within a form that does not provide all of the features and benefits set forth herein, as some features can be used or practiced separately from others.
, Claims:CLAIMS

We Claim:
A method of enhancing Digital Pre-Distortion (DPD) model performance in an Open Radio Access Network (O-RAN), the method comprises:
receiving, by an Open Radio Unit (O-RU) (500), a baseband output signal z(n) by a power amplifier (PA) (506), wherein the baseband output signal z(n) is defined by a first predefined dimension;
comparing, by the O-RU (500), the baseband output signal z(n) with a baseband input signal x(n) to obtain one or more missing values m(n), wherein the baseband input signal x(n) is received at the PA (506); and
reconstructing, by the O-RU (500), the baseband output signal z(n) by adding the one or more missing values m(n) to form a reconstructed signal g(n), wherein the reconstructed signal g(n) is defined by the first predefined dimension, wherein the reconstructed signal g(n) is a linear distortion free signal.

The method as claimed in claim 1, wherein the method comprises re-dimensioning, by the O-RU (500), the reconstructed signal g(n) to form a refined reconstructed signal g’(n), wherein the refined reconstructed signal g’(n) is defined by a second predefined dimension, wherein the refined reconstructed signal g’(n) is a linear distortion free signal.

The method as claimed in claim 2, wherein the second predefined dimension is less than the first predefined dimension.

The method as claimed in claim 1, wherein comparing the baseband output signal z(n) with the baseband input signal x(n) to obtain the one or more missing values m(n) comprises estimating one or more model coefficients, wherein the one or more model coefficients are associated with the baseband input signal x(n) and the baseband output signal z(n).
The method as claimed in claim 1, wherein reconstructing the baseband output signal z(n) comprises reducing one or more model coefficients and the one or more missing values m(n) from the baseband output signal z(n) to obtain the reconstructed signal g(n).

The method as claimed in claim 1, wherein comparing the baseband output signal z(n) with the baseband input signal x(n) to obtain the one or more missing values m(n) comprises:
extracting one or more model coefficients from the baseband input signal x(n) for PA modelling;
reducing the one or more model coefficients by considering relevant coefficients that are needed for accurately modelling the PA (506); and
comparing the reduced one or more model coefficients with a set of baseband input signal x(n) and reference values and controlling adjustable parameters to obtain the one or more missing values m(n), wherein the adjustable parameters are checking normalized mean square error (NMSE) values.

The method as claimed in claim 1, wherein the baseband output signal z(n) is received at a radio frequency (RF)-receiver (512) as a baseband input signal by:
computing one or more model coefficients using a least square (LS) technique, wherein the LS technique is defined by normalized mean square error (NMSE) values for new dimension data;
removing coefficient from the new dimension data;
comparing original data with a new set of data; and
computing missing values by eigenvalues, wherein the eigenvalues define the reduction of the dimensionality of the data.

The method as claimed in claim 6, wherein the PA modelling is computed by:
the baseband output signal z(n) at a radio frequency (RF) receiver (512):

z(n)=d_0+∑_(i=0)^M ∑_(q=0)^(N-1) b_iq x(n-i) |x(n-i)|^q +∑_(i=0)^M ∑_(q=0)^(N-1) C_iq x^* (n-i) |x^* (n-i)|^q
where M is memory order, N is a non-linearity order of PA, do is a constant term, b_iq,C_iq are the DPD model coefficients, x(n) is the baseband input signal and z(n) is the baseband output signal, wherein the equation is written in matrix form as: z ⃗=W ⃗B ⃗
where B ⃗ is a vector containing the model-coefficients, W ⃗ is the observation data matrix.

The method as claimed in claim 7, wherein the one or more model coefficients are computed using the least mean square (LMS) technique by comparing the new set of data with the input baseband signal x(n), wherein the LMS technique is used to compute a vector containing the model-coefficients B ⃗=(W ⃗^H W ⃗ ) W ⃗^H z ⃗, where, W is the input observation data matrix.

The method as claimed in claim 1, wherein the one or more missing values are estimated using a probabilistic principal component analysis by determining a correlation matrix using an observation data matrix (W ⃗) as , wherein an eigenvalue decomposition is applied as
where O is the matrix of eigenvectors and ∧ is the diagonal matrix of eigenvalues , wherein the desired length of the data matrix is determined using the weight of the eigenvalue of data variance, and the principal components are defined as using S eigenvectors having all important data and after discarding the insignificant data.
G= [o_1,o_2, o_3,….o_s ]

The method as claimed in claim 1, wherein the DPD model performance is used for digital compensation of radio frequency (RF) power amplifier nonlinearities in the O-RU (500).

An Open Radio Unit (O-RU) (500) with improved Digital Pre-Distortion (DPD) model performance in an Open Radio Access Network (O-RAN), the O-RU (500) comprises:
a radio frequency (RF) transmitter (504);
a radio frequency (RF) receiver (512); and
a pre-distorter training and PPCA unit (514), coupled with the RF transmitter (504) and the RF receiver (512), configured to:
receive a baseband output signal z(n) by a power amplifier (PA) (506), wherein the baseband output signal z(n) is defined by a first predefined dimension;
compare the baseband output signal z(n) with a baseband input signal x(n) to obtain one or more missing values m(n), wherein the baseband input signal x(n) is received at the PA (506); and
reconstruct the baseband output signal z(n) by adding the one or more missing values m(n) to form a reconstructed signal g(n), wherein the reconstructed signal g(n) is defined by the first predefined dimension, wherein the reconstructed signal g(n) is a linear distortion free signal.