Limits...
Augmented twin-nonlinear two-box behavioral models for multicarrier LTE power amplifiers.

Hammi O - ScientificWorldJournal (2014)

Bottom Line: A novel class of behavioral models is proposed for LTE-driven Doherty power amplifiers with strong memory effects.When strong memory effects are observed, the augmented twin-nonlinear two-box models can improve the normalized mean square error by up to 3 dB for the same number of coefficients when compared to state-of-the-art twin-nonlinear two-box models.Furthermore, the augmented twin-nonlinear two-box models lead to the same performance as previously reported twin-nonlinear two-box models while requiring up to 80% less coefficients.

View Article: PubMed Central - PubMed

Affiliation: Electrical Engineering Department, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia.

ABSTRACT
A novel class of behavioral models is proposed for LTE-driven Doherty power amplifiers with strong memory effects. The proposed models, labeled augmented twin-nonlinear two-box models, are built by cascading a highly nonlinear memoryless function with a mildly nonlinear memory polynomial with cross terms. Experimental validation on gallium nitride based Doherty power amplifiers illustrates the accuracy enhancement and complexity reduction achieved by the proposed models. When strong memory effects are observed, the augmented twin-nonlinear two-box models can improve the normalized mean square error by up to 3 dB for the same number of coefficients when compared to state-of-the-art twin-nonlinear two-box models. Furthermore, the augmented twin-nonlinear two-box models lead to the same performance as previously reported twin-nonlinear two-box models while requiring up to 80% less coefficients.

Show MeSH

Related in: MedlinePlus

NMSE performance of the forward twin-nonlinear two-box model and the augmented forward twin-nonlinear two-box model as function of the nonlinearity order.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3927868&req=5

fig5: NMSE performance of the forward twin-nonlinear two-box model and the augmented forward twin-nonlinear two-box model as function of the nonlinearity order.

Mentions: First, the FTNTB model was identified using the measured data for a wide set of nonlinearity orders and memory depths. The nonlinearity order (N1) was varied from 5 to 15 and the memory depth (M1) was swept from 2 to 10. For each set of coefficients, the performance of the FTNTB model was evaluated in terms of the normalized mean-square error (NMSE). Then, the AFTNTB model was identified for the same ranges of the nonlinearity order and memory depth and for a leading and lagging cross terms order (L) of 1. For each set of coefficients, the performance of the AFTNTB model was evaluated in terms of its NMSE. The results are reported in Figure 5. For clarity reasons, these results include the NMSE values obtained for a memory depth up to 6. Higher values of memory depths were found to result in marginal NMSE improvement and are thus not included in the figure. Figure 5 demonstrates the performance enhancement obtained with the proposed model by including the first order cross terms. Indeed, a 3 dB NMSE improvement is obtained when the AFTNTB model is used. The performances of both the state-of-the-art model and the proposed one improve as the nonlinearity order and/or the memory depth of the model increases. However, the proposed ATNTB model consistently outperforms the conventional one by more than 3 dB when the same nonlinearity order and memory depth are used for both models.


Augmented twin-nonlinear two-box behavioral models for multicarrier LTE power amplifiers.

Hammi O - ScientificWorldJournal (2014)

NMSE performance of the forward twin-nonlinear two-box model and the augmented forward twin-nonlinear two-box model as function of the nonlinearity order.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3927868&req=5

fig5: NMSE performance of the forward twin-nonlinear two-box model and the augmented forward twin-nonlinear two-box model as function of the nonlinearity order.
Mentions: First, the FTNTB model was identified using the measured data for a wide set of nonlinearity orders and memory depths. The nonlinearity order (N1) was varied from 5 to 15 and the memory depth (M1) was swept from 2 to 10. For each set of coefficients, the performance of the FTNTB model was evaluated in terms of the normalized mean-square error (NMSE). Then, the AFTNTB model was identified for the same ranges of the nonlinearity order and memory depth and for a leading and lagging cross terms order (L) of 1. For each set of coefficients, the performance of the AFTNTB model was evaluated in terms of its NMSE. The results are reported in Figure 5. For clarity reasons, these results include the NMSE values obtained for a memory depth up to 6. Higher values of memory depths were found to result in marginal NMSE improvement and are thus not included in the figure. Figure 5 demonstrates the performance enhancement obtained with the proposed model by including the first order cross terms. Indeed, a 3 dB NMSE improvement is obtained when the AFTNTB model is used. The performances of both the state-of-the-art model and the proposed one improve as the nonlinearity order and/or the memory depth of the model increases. However, the proposed ATNTB model consistently outperforms the conventional one by more than 3 dB when the same nonlinearity order and memory depth are used for both models.

Bottom Line: A novel class of behavioral models is proposed for LTE-driven Doherty power amplifiers with strong memory effects.When strong memory effects are observed, the augmented twin-nonlinear two-box models can improve the normalized mean square error by up to 3 dB for the same number of coefficients when compared to state-of-the-art twin-nonlinear two-box models.Furthermore, the augmented twin-nonlinear two-box models lead to the same performance as previously reported twin-nonlinear two-box models while requiring up to 80% less coefficients.

View Article: PubMed Central - PubMed

Affiliation: Electrical Engineering Department, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia.

ABSTRACT
A novel class of behavioral models is proposed for LTE-driven Doherty power amplifiers with strong memory effects. The proposed models, labeled augmented twin-nonlinear two-box models, are built by cascading a highly nonlinear memoryless function with a mildly nonlinear memory polynomial with cross terms. Experimental validation on gallium nitride based Doherty power amplifiers illustrates the accuracy enhancement and complexity reduction achieved by the proposed models. When strong memory effects are observed, the augmented twin-nonlinear two-box models can improve the normalized mean square error by up to 3 dB for the same number of coefficients when compared to state-of-the-art twin-nonlinear two-box models. Furthermore, the augmented twin-nonlinear two-box models lead to the same performance as previously reported twin-nonlinear two-box models while requiring up to 80% less coefficients.

Show MeSH
Related in: MedlinePlus