Reduced-Complexity Polynomials with Memory Applied to the Linearization of Power Amplifiers with Real-Time Discrete Gain Control

Schuartz, Luis; Santos, Edson Leonardo dos; Leite, Bernardo; Mariano, André; Lima, Eduardo G.

doi:10.1007/s00034-019-01049-6

Cited by 5 publications

(8 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The ascendant method was proposed to memory polynomials models to reduce their amount of parameters [12]. In this work, the ascendant method is employed to simplify the TLP implementation.…”

Section: Ascendant Methodsmentioning

confidence: 99%

“…In what concerns the multimode real-time operation, besides of previously presented characteristics, its AM-AM and AM-PM discontinuous curves are a challenge. Modeling the contribution of each mode in a specific TLP is a possible solution [12]. Thus, the block diagram of the TLP-based model for the multimode PA is shown in Fig.…”

Section: Tlp-based Model For Multimode Pasmentioning

confidence: 99%

“…Since DPD units need inverse models with the least possible complexity to ensure low energy consumption, the ascendant method for neural networks will be suggested and validated. This method was proposed for models based on memory polynomials and focused on selecting the most relevant parameters to be kept active, reducing the number of operations required in the model [12].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Behavioral Modeling of a Multimode Power Amplifier Using Perceptrons with AM-AM and AM-PM Complex-Valued Activation Function

Freire

Schuartz²,

Santos³

et al. 2021

JICS

Self Cite

View full text Add to dashboard Cite

Three-layer perceptron (TLP) is one of the approaches to the behavioral modeling of radio frequency (RF) power amplifiers (PAs) for wireless communication systems. A low-pass equivalent PA behavioral modeling should be able to represent the AM-AM and AM-PM characteristic curves, besides generating only useful contributions. This work shows that adapting a complex-valued activation function toencompass the strict characteristics from the AM-AM and AM-PM curves at the same time fulfilling the bandpass constraint gives better results in terms of network complexity and accuracy. Besides that, a novel TLP-based model for a multimode PA is proposed. This type of PA has characteristics that add difficulty to its inverse modeling. A modification in the ascendant method is also proposed for application with TLP-based PA models. This method simplifies the models in terms of operations needed for implementation. In a scenario of same number of network parameters, a significant reduction in modeling error is achieved when a complex-valued TLP that accounts for both AM-AM and AM-PM is used instead of a complex-valued TLP that accounts only for AM-AM. Such reduction is quantified by up to 5.73 dB (-35.53 dB to -41.26 dB) improvements in normalized mean square error (NMSE) metric. Moreover, when applying the ascendant method a 67% reduction in active parameters is achieved (36 to 12 active parameters), deteriorating the NMSE by less than 0.5 dB.

show abstract

Section: Ascendant Methodsmentioning

confidence: 99%

Section: Tlp-based Model For Multimode Pasmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Behavioral Modeling of a Multimode Power Amplifier Using Perceptrons with AM-AM and AM-PM Complex-Valued Activation Function

Freire

Schuartz²,

Santos³

et al. 2021

JICS

Self Cite

View full text Add to dashboard Cite

show abstract

“…As discussed in (ELAD, 2010), (DAVIES; RILLING; BLUMENSATH, 2012), the greedy pursuit performance guarantees are typically weaker than LASSO, both in terms of variable selection consistency and accuracy. Also note that CS-based techniques has the additional advantage of reducing the DPD running complexity, whereas techniques that carry out a regression basis change, such as the principal component analysis (PCA) in (LOPEZ-BUENO et al, 2018), (GILABERT et al, 2013) imply an increase in DPD running cost (SCHUARTZ et al, 2019). Examples in the literature that have applied LASSO to reduce the dimensionality of PA and DPD models are (ZENTENO et al, 2015), (KEKATOS; GIANNAKIS, 2011), (WISELL; JALDEN; HANDEL, 2008).…”

Section: Sparse Least Squares Estimationmentioning

confidence: 99%

“…Although orthogonalizing the regression matrix Φ simplifies the estimation and model sizing tasks associated to the model in eq. (6.1), in the context of DPD the transformation of the original, observable regression basis, Φ, into an orthogonal basis Θ severely impacts the running cost of the model, since a matrix-to-matrix multiplication is required before the DPD filtering of the input signal (SCHUARTZ et al, 2019).…”

Section: Orthogonalizationmentioning

confidence: 99%

Digital signal processing techniques for the compensation of analog circuit impairments in impedance mismatched radio-frequency transmitters.

Hemsi¹

View full text Add to dashboard Cite

Esta tese investiga e propõe novos modelos para a compensação digital de imperfeições de rádio frequência (RFradio frequency) em transmissores sem fio de banda larga, mais especificamente para amplificadores de potência (PApower amplifiers) não-lineares e sujeitos a descasamento de impedância de carga (LMMload mismatch). Tais imperfeições em transmissores de RF, juntamente com o desbalanceamento entre as componentes em fase e quadratura (IQin-phase and quadrature) em moduladores digitais, são responsáveis por degradar o desempenho do transmissor, tanto em termos de pureza espectral, quanto em qualidade da modulação e taxa de erro de bit (BERbit error rate).Diversos cenários práticos em que PAs estão sujeitos a LMM motivam a pesquisa por modelos comportamentais não-lineares e com memória voltados a predistorção digital (DPDdigital pre-distortion) mais avançados, capazes de superar as limitações relatadas na literatura de modelos tradicionais e que sejam menos complexos do que as abordagens existentes para PA LMM.Nesta tese é proposta a aplicação do modelo polinomial Wiener-Hammerstein com realimentação (WHFB -Wiener-Hammerstein with feedback) como modelo comportamental simplificado para DPD no contexto de PAs sujeitos a LMM.A elevada dimensionalidade da estrutura WHFB proposta pode ser reduzida através de técnicas esparsas de estimação, tais como o operador de encolhimento e seleção absoluto mínimo (LASSOleast absolute shrinkage and selection operator) e extensões de LASSO em blocos, que são capazes de diminuir significativamente o número de coeficientes necessário, reduzindo, assim, o comprimento do filtro de DPD e, proporcionalmente, o custo da filtragem.Além disso, extensões de LASSO voltadas a blocos, como Group-LASSO e Sparse-group LASSO, são aplicadas no contexto de dimensionamento do modelo, ou seja, na tarefa de determinar valores adequados para seus parâmetros, o que, tradicionalmente, requer uma busca exaustiva.Os modelos WHFB denso e esparso são validados experimentalmente através de medições em um set-up de testes experimental e também comparados a outros, incluindo modelos baseados na decomposição em fatores paralelos (PARAFACparallel factors) e na expansão de Laguerre, demonstrando, assim, sua capacidade de compensar adequadamente PAs sujeitos a LMM.Finalmente, é proposta uma nova estratégia de redução do modelo de Volterra, que resulta em um polinômio de memória mais flexível e modular, em que os parâmetros são escolhidos independentemente para cada ordem de não linearidade. Esta abordagem flexível é capaz de descrever com precisão uma gama ampla de condições operacionais/ ambientais de PAs.

show abstract

Concurrent Tri-band CMOS Power Amplifier Linearized by 3D Improved Memory Polynomial Digital Predistorter

Schuartz

Silva

Hara

et al. 2020

Circuits Syst Signal Process

View full text Add to dashboard Cite

Reduced-Complexity Polynomials with Memory Applied to the Linearization of Power Amplifiers with Real-Time Discrete Gain Control

Cited by 5 publications

References 37 publications

Behavioral Modeling of a Multimode Power Amplifier Using Perceptrons with AM-AM and AM-PM Complex-Valued Activation Function

Behavioral Modeling of a Multimode Power Amplifier Using Perceptrons with AM-AM and AM-PM Complex-Valued Activation Function

Digital signal processing techniques for the compensation of analog circuit impairments in impedance mismatched radio-frequency transmitters.

Concurrent Tri-band CMOS Power Amplifier Linearized by 3D Improved Memory Polynomial Digital Predistorter

Contact Info

Product

Resources

About