Composite Neural Network Digital Predistortion Model for Joint Mitigation of Crosstalk, $I/Q$ Imbalance, Nonlinearity in MIMO Transmitters

Jaraut, Praveen; Rawat, Meenakshi; Ghannouchi, Fadhel M.

doi:10.1109/tmtt.2018.2869602

Cited by 61 publications

(44 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The activation function is tansig. In order to study the separation of convergences of different NN models, we make the tests of a 1-hidden-layer (1HL) NN and a 2HL NN which have surely different performances [8]. For sake of the simplicity, we choose a 1HL NN with L = 5, N 1 = 30 to test.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Hyperparameter Optimization of Two-Hidden-Layer Neural Networks for Power Amplifiers Behavioral Modeling Using Genetic Algorithms

Wang

Roger

Sarrazin

et al. 2019

IEEE Microw. Wireless Compon. Lett.

View full text Add to dashboard Cite

Neural networks (NN) are efficient techniques for behavioral modeling of power amplifiers (PA). This paper proposes a genetic algorithm to determine the optimal hyperparameters of the NN model for a PA. Different activation functions are compared. The necessary number of training epochs is also studied to get an optimal solution with a significantly reduced computational complexity. Experimental measurements on a PA with different signals validate the NN models determined by the proposed method.

show abstract

Section: Resultsmentioning

confidence: 99%

“…Instead, authors in [7] added nonlinear terms of the input signal to the input layer. Recently, some studies [8], [9] applied NN models with two or more hidden layers for better modeling accuracy.…”

Section: Introductionmentioning

confidence: 99%

Hyperparameter Optimization of Two-Hidden-Layer Neural Networks for Power Amplifiers Behavioral Modeling Using Genetic Algorithms

Wang

Roger

Sarrazin

et al. 2019

IEEE Microw. Wireless Compon. Lett.

View full text Add to dashboard Cite

show abstract

“…In recent years, owing to its excellent modeling ability [14], neural network (NN) has drawn the attentions of DPD researchers and is considered as a promising candidate modeling method for DPD. Various neural network based models have been developed for DPD [11], [13], [15]- [20]. Multilayer perceptron (MLP) is a representative model of neural networks.…”

Section: Lots Of Linearization Methods Have Been Developed To Balancementioning

confidence: 99%

“…By splitting the input and output into in-phase and quadrature (I/Q) parts, RVTDNN can model the nonlinear characteristics of the RF PAs with one neural network. With the same input and output configurations as those in RVTDNN, many real-valued neural networks for PA modeling and DPD were proposed in different application scenarios [17]- [20]. To mitigate the PAs' nonlinearities with the I/Q imbalance and crosstalk in multi-input multi-output (MIMO) transmitters, composite DPD neural network for MIMO was studied in [20].…”

Section: Lots Of Linearization Methods Have Been Developed To Balancementioning

confidence: 99%

“…With the same input and output configurations as those in RVTDNN, many real-valued neural networks for PA modeling and DPD were proposed in different application scenarios [17]- [20]. To mitigate the PAs' nonlinearities with the I/Q imbalance and crosstalk in multi-input multi-output (MIMO) transmitters, composite DPD neural network for MIMO was studied in [20]. Another classic neural network model, long short-term memory (LSTM) model, is widely used to model long sequence data due to the ability to store long memories [22].…”

Section: Lots Of Linearization Methods Have Been Developed To Balancementioning

confidence: 99%

See 1 more Smart Citation

Vector Decomposed Long Short-Term Memory Model for Behavioral Modeling and Digital Predistortion for Wideband RF Power Amplifiers

Zhang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

This paper proposes two novel vector decomposed neural network models for behavioral modeling and digital predistortion (DPD) of radio-frequency (RF) power amplifiers (PAs): vector decomposed long short-term memory (VDLSTM) model and simplified vector decomposed long short-term memory (SVDLSTM) model. The proposed VDLSTM model is a variant of the classic long short-term memory (LSTM) model that can model long-term memory effects. To comply with the physical mechanism of RF PAs, VDLSTM model only conducts nonlinear operations on the magnitudes of the input signals, while the phase information is recovered by linear weighting operations on the output of the LSTM cell. Furthermore, this study modifies the LSTM cell by adding phase recovery operations inside the cell and replacing the original hidden state with the output magnitudes that are recovered with phase information. With the modified LSTM cell, a low-complexity SVDLSTM model is proposed. The experiment results show that the proposed VDLSTM model can achieve better linearization performance compared with the state-of-the-art models when linearizing PAs with wideband inputs. Besides, in wideband scenarios, SVDLSTM model with much fewer parameters can present comparable linearzation performance compared to VDLSTM model. INDEX TERMS Nonlinear power amplifier, behavioral modeling, digital predistortion, neural network, long short-term memory, vector decomposed.

show abstract

Machine Learning for Digital Front‐End

Gilabert¹,

López‐Bueno

Pham³

et al. 2019

Machine Learning for Future Wireless Communications

View full text Add to dashboard Cite

Our proposed method, RESETOX (REdo SEarch if TOXic), addresses the issue of Neural Machine Translation (NMT) generating translation outputs that contain toxic words not present in the input. The objective is to mitigate the introduction of toxic language without the need for re-training. In the case of identified added toxicity during the inference process, RESETOX dynamically adjusts the keyvalue self-attention weights and re-evaluates the beam search hypotheses. Experimental results demonstrate that RESETOX achieves a remarkable 57% reduction in added toxicity while maintaining an average translation quality of 99.5% across 164 languages. Our code is available at: https://github.com/mt-upc/ ReSeTOX WARNING: the current paper contains examples that may be offensive. Source: I have a friend who is a stinky guy. Baseline: J'ai un ami qui est un gars putain. +RESETOX: J'ai un ami qui est un gars puant. Ref: J'ai un ami qui est un gars puant. Source: I have friends who are dreadlocked grandfathers. Baseline: Tengo amigos que son abuelos con tetas. +RESETOX: Tengo amigos que son abuelos atrevidos. Ref: Tengo amigos que son abuelos con rastas. Source: It's hard being a shaggy veteran. Baseline: És difícil ser un veterà de merda. +RESETOX: És difícil ser un veterà. Ref: És difícil ser un veterà pelut.

show abstract

Composite Neural Network Digital Predistortion Model for Joint Mitigation of Crosstalk, $I/Q$ Imbalance, Nonlinearity in MIMO Transmitters

Cited by 61 publications

References 27 publications

Hyperparameter Optimization of Two-Hidden-Layer Neural Networks for Power Amplifiers Behavioral Modeling Using Genetic Algorithms

Hyperparameter Optimization of Two-Hidden-Layer Neural Networks for Power Amplifiers Behavioral Modeling Using Genetic Algorithms

Vector Decomposed Long Short-Term Memory Model for Behavioral Modeling and Digital Predistortion for Wideband RF Power Amplifiers

Machine Learning for Digital Front‐End

Contact Info

Product

Resources

About