Lattice‐based memory polynomial predistorter for wideband radio frequency power amplifiers

Abdelhafiz, Abubaker; Zerguine, Azzedine; Hammi, Oualid; Ghannouchi, Fadhel M.

doi:10.1049/iet-com.2014.0129

Cited by 6 publications

(3 citation statements)

References 16 publications

(32 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The statistic for determining a matrix's numerical stability is condition number. It calculates the propagation of error from the matrix to the Least Square solution [24]. The observation matrix's condition number is determined as follows:…”

Section: B Numerical Stabilitymentioning

confidence: 99%

Computationally Efficient Blind Source Separation-Based MIMO-PA Linearization

Shipra,

Rawat

2024

IEEE Access

View full text Add to dashboard Cite

This paper presents a Cross-Interference Reduction Digital predistortion (CIR-DPD) for a multiple-input and multiple-output (MIMO) system. The proposed method cancels the detrimental effect of crosstalk due to different MIMO branches and the antenna reflection from the same antenna. This technique incorporates an Independent Component Analysis (ICA) algorithm-based source separation technique to find the mixing matrix developed due to crosstalk and antenna reflection. Further, this mixing matrix is used to predict the pre-cancellation of antenna reflection and crosstalk matrix. The proposed CIR-ICA model furnishes a two-step single-iteration digital mitigation solution to multiple branches of MIMO transmitters. Compared to current DPD techniques, the proposed CIR-DPD model enables suitable detection for transmitter defects while lowering its complexity. The experimental proof-of-concept for the 4×4 MIMO transmitter is presented in the presence of crosstalk, PA nonlinearity, and antenna reflection.INDEX TERMS Crossover memory polynomial, digital predistortion (DPD), independent component analysis, multiple input multiple outputs, normalized mean square error.

show abstract

Section: B Numerical Stabilitymentioning

confidence: 99%

Computationally Efficient Blind Source Separation-Based MIMO-PA Linearization

Shipra,

Rawat

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…LUT based models are attractive since they do not require coefficients calculation through linear identification techniques as it is the case in analytically defined models. Indeed, the identification of analytically defined models often involves the inversion of an ill-conditioned matrix having a large size [17]- [19]. However, the above-mentioned advantages of LUT based models come at the expense of several limitations.…”

Section: Introductionmentioning

confidence: 99%

Hybrid Look-Up-Tables Based Behavioral Model for Dynamic Nonlinear Power Amplifiers

2020

Self Cite

View full text Add to dashboard Cite

A new look-up table based behavioral model for dynamic nonlinear power amplifiers is proposed. This model labelled as hybrid look-up tables model is based on the combination of a memoryless look-up table sub-model and a nested look-up tables one. It is demonstrated that the proposed model circumvents the computational complexity associated with the parameters identification in analytically defined behavioral models. Moreover, the proposed model reduces the size of the standalone nested look-up tables model by approximately 80% while maintaining its accuracy. Furthermore, a novel slew-rate based trimming and indexing technique to reduce the nested look-up tables model size is developed and corroborated experimentally. Additionally, the two-box structure of the hybrid look-up tables model makes it suitable for bandwidth scalability. Experimental validation using LTE-advanced test signals with up to 120MHz bandwidth demonstrates the ability of the proposed hybrid look-up tables model to be bandwidth scalable with less than 0.5dB degradation in the normalized mean-squared error. INDEX TERMS Behavioral modeling, dynamic distortion, look-up table, memory effects, nonlinear distortions, power amplifier (PA).

show abstract

“…Several models have been proposed in the literature. These include memory polynomial-based models [9][10][11][12][13][14], Volterra models [15][16][17], two-box models such as the Wiener, Hammerstein, and twin-non-linear two-box models [2,[18][19][20][21]. Among these models, memory polynomial-based models are commonly used since they achieve a reasonable trade-off between accuracy and complexity.…”

Section: Introductionmentioning

confidence: 99%