Upgrading Behavioral Models for the Design of Digital Predistorters

Area of Science:

• Electrical Engineering
• Signal Processing
• Communications Engineering

Background:

• Power amplifier (PA) behavioral models and digital predistortion (DPD) are often truncated due to computational constraints.
• A priori pruned model structures may omit significant terms, limiting performance.

Purpose of the Study:

• To present a strategy for upgrading incomplete PA models.
• To improve the performance of DPD systems by enhancing model accuracy.

Main Methods:

• A general procedure to augment model structures by incorporating significant nonlinear and memory depth terms.
• Utilizing the sparse nature of the problem for successive search and addition of terms.
• Investigating the approach on commercial class AB and class J PAs.

Main Results:

• Demonstrated capability to augment incomplete PA models effectively.
• Significant improvement in the linearization capabilities of DPDs was observed.
• The method successfully identified and incorporated crucial nonlinear and memory terms.

Conclusions:

• The proposed upgrading procedure enhances PA behavioral models.
• This leads to superior linearization performance in DPD applications.
• The strategy offers a viable solution for improving PA modeling with limited resources.