On-Demand Real-Time Optimizable Dynamic Model Sizing for Digital Predistortion of Broadband RF Power Amplifiers

Li, Yue; Zhu, Anding

doi:10.1109/tmtt.2020.2982165

Cited by 23 publications

(8 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DMS is a model structure adaptation algorithm proposed in [ 28 ], which can well address the challenges of adaptive model pruning problems while achieving good pruning performance. The algorithm starts from a given model structure and iteratively searches for a new model suitable for the current PA condition.…”

Section: Feature Selection Techniquesmentioning

confidence: 99%

Comparison of Feature Selection Techniques for Power Amplifier Behavioral Modeling and Digital Predistortion Linearization

Barry

Becerra

et al. 2021

Sensors

View full text Add to dashboard Cite

The power amplifier (PA) is the most critical subsystem in terms of linearity and power efficiency. Digital predistortion (DPD) is commonly used to mitigate nonlinearities while the PA operates at levels close to saturation, where the device presents its highest power efficiency. Since the DPD is generally based on Volterra series models, its number of coefficients is high, producing ill-conditioned and over-fitted estimations. Recently, a plethora of techniques have been independently proposed for reducing their dimensionality. This paper is devoted to presenting a fair benchmark of the most relevant order reduction techniques present in the literature categorized by the following: (i) greedy pursuits, including Orthogonal Matching Pursuit (OMP), Doubly Orthogonal Matching Pursuit (DOMP), Subspace Pursuit (SP) and Random Forest (RF); (ii) regularization techniques, including ridge regression and least absolute shrinkage and selection operator (LASSO); (iii) heuristic local search methods, including hill climbing (HC) and dynamic model sizing (DMS); and (iv) global probabilistic optimization algorithms, including simulated annealing (SA), genetic algorithms (GA) and adaptive Lipschitz optimization (adaLIPO). The comparison is carried out with modeling and linearization performance and in terms of runtime. The results show that greedy pursuits, particularly the DOMP, provide the best trade-off between execution time and linearization robustness against dimensionality reduction.

show abstract

Section: Feature Selection Techniquesmentioning

confidence: 99%

Comparison of Feature Selection Techniques for Power Amplifier Behavioral Modeling and Digital Predistortion Linearization

Barry

Becerra

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Notwithstanding the notable performance of these approaches, a general concern exists on techniques to upgrade and optimize a given model. This interest has motivated recent publications which compare model pruning or model growing techniques [ 23 , 24 ]. In [ 24 ], the model growth is made taking into account only the initial set of the GMP regressors, but it could be necessary the upgrade with regressors not included in the complete GMP model.…”

Section: Introductionmentioning

confidence: 99%

“…This interest has motivated recent publications which compare model pruning or model growing techniques [ 23 , 24 ]. In [ 24 ], the model growth is made taking into account only the initial set of the GMP regressors, but it could be necessary the upgrade with regressors not included in the complete GMP model. Unfortunately, no results have been published for an HC algorithm applied to a general FV model because the regressor set size is unsuitably large.…”

Section: Introductionmentioning

confidence: 99%

“…The cited works evidence the need for a procedure to upgrade sub-optimal models in multiple scenarios. In the first one, which can be denoted as intra-model scenario, the search of new active regressors is circumscribed to the same model, e.g., the GMP in [ 22 , 23 , 24 ]. In a second inter-model scenario, the search is extended to a new model with a richer set of regressors so that the optimal model will benefit from a boost provided by the regressors of the second model.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Upgrading Behavioral Models for the Design of Digital Predistorters

Crespo‐Cadenas

Madero‐Ayora

Becerra

2021

Sensors

View full text Add to dashboard Cite

This work presents a strategy to upgrade models for power amplifier (PA) behavioral modeling and digital predistortion (DPD). These incomplete structures are the consequence of nonlinear order and memory depth model truncation with the purpose of reducing the demand of the limited computational resources available in standard processors. On the other hand, the alternative use of model structures pruned a priori does not guarantee that every significant term is included. To improve the limited performance of an incomplete model, a general procedure to augment its structure by incorporating significant terms is demonstrated. The sparse nature of the problem allows a successive search incorporating additional terms with higher nonlinear order and memory depth. This approach is investigated in the modeling and linearization of a commercial class AB PA operating at a compression point of about 6 dB, and a class J PA operating near saturation. Results highlight the capabilities of this upgrading procedure in the improvement of linearization capabilities of DPDs.

show abstract

“…But the above techniques are all based on the Volterra series theory, and its greatest disadvantage is that under the condition of a broadband system, the number of model terms is exponentially proportional to the nonlinear order and memory depth, which will increase the model complexity of broadband systems. 14 Recently, some artificial neural network (ANN) models have also been used to compensate for transmitter impairments such as bidirectional long short-term memory (BiLSTM) neural network, radial basis function (RBF) neural network and time delay neural network (TDNN). [15][16][17] However, all these proposed ANN models are only compensating for PA nonlinearity and memory effects, without considering the correction of the imperfect characteristics of the modulator and the RF frontend.…”

Section: Introductionmentioning

confidence: 99%

An improved nonlinear smooth twin support vector regression based‐behavioral model for joint compensation of frequency‐dependent transmitter nonlinearities

Cai

Yao

et al. 2021

Int J RF Microw Comput Aided Eng

View full text Add to dashboard Cite

In this article, an improved nonlinear smooth twin support vector regression (NSTSVR) model is proposed for the modeling and compensating of the transmitter nonlinearities jointly. The proposed model is an improved version of the twin support vector regression (TSVR) model by introducing a smooth function to replace the loss function of TSVR, which can change the dual space solution to the original space solution and speed up the solving solution. In addition, in order to solve the problem of long training time for large sample data in traditional SVR or TSVR model, the new algorithm further adopts the model pruning techniques, such as deleting the kernel matrix and finding sparse diagonal matrices, to reduce the size of the Hessian matrix in the fast Newton iteration process. To verify the performance of the proposed model, two transmitters based on single-device gallium nitride (GaN) PA with IQ imbalance and GaN Doherty PA with modulator imperfections are used for experimental verification and analysis. The experimental results show that the proposed model is superior to the conventional support vector regression and TSVR machine learning models in terms of modeling effect and linearization ability. Furthermore, the proposed model can achieve the improved compensation performance for transmitter impairments compared with some popular Volterra series-based I/Q imbalance models.

show abstract

On-Demand Real-Time Optimizable Dynamic Model Sizing for Digital Predistortion of Broadband RF Power Amplifiers

Cited by 23 publications

References 30 publications

Comparison of Feature Selection Techniques for Power Amplifier Behavioral Modeling and Digital Predistortion Linearization

Comparison of Feature Selection Techniques for Power Amplifier Behavioral Modeling and Digital Predistortion Linearization

Upgrading Behavioral Models for the Design of Digital Predistorters

An improved nonlinear smooth twin support vector regression based‐behavioral model for joint compensation of frequency‐dependent transmitter nonlinearities

Contact Info

Product

Resources

About