Dynamic Volterra Kernel Selection for Digital Pre-Distortion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing digital pre-distortion technologies face challenges in efficiently adapting to dynamic changes in modulation schemes, power amplifier output levels, and multi-carrier configurations due to the complexity of Volterra kernels, leading to increased convergence time and performance limitations.

Innovation Solution

A dynamic kernels selection structure/algorithm that models the RF power amplifier's non-linear system using input and output data to select the minimum required Volterra kernels for optimal performance and fastest convergence, continuously updating the Volterra kernels set based on current operating conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If predetermined Volterra kernels are used to support all modulation standards and power amplifier configurations, then the system can handle diverse operating conditions, but the device complexity and computational load increase exponentially

Engineering Contradiction:
Improvesupport for all modulation standards and configurationsVSAvoidnumber of Volterra kernels
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic selection of Volterra kernels based on current operating conditions (modulation scheme, power level, carrier configuration). Instead of using a fixed large set of kernels for all scenarios, the system adaptively determines which kernels are needed for the current state, reducing computational complexity while maintaining versatility across different modulation standards and configurations.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If a large number of predetermined Volterra kernels are used to model all power amplifier operating conditions, then the modeling accuracy is improved, but the adaptation convergence time is significantly increased

Engineering Contradiction:
Improvepower amplifier modeling accuracyVSAvoidadaptation convergence time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts and selects only the dominant Volterra kernels that are most relevant to the current power amplifier operating conditions. By identifying and removing unnecessary kernels from the computational set, the system maintains accurate modeling of PA non-linearities while significantly reducing the number of computations required for adaptation convergence.

Inventive Principle:
Principle #2Taking out (Extraction)

3Speed

If orthogonal polynomials are used to express Volterra kernels, then the converging speed is improved, but redundancy is introduced due to higher order forms containing lower order forms

Engineering Contradiction:
Improveconverging speedVSAvoidredundancy in kernel forms
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The patent removes redundant kernel forms from the orthogonal polynomial expansion by selectively including only the necessary orders and types of Volterra kernels for the current operating conditions. This eliminates the redundancy where higher order forms contain lower order forms, while preserving the fast convergence benefits of orthogonal polynomials for the essential kernel components.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS8804872B1Dynamic determination of volterra kernels for digital pre-distortion
Publication Date: 2014.08.12 TEXAS INSTRUMENTS INC
  • US8804872B1 patent drawing
  • US8804872B1 patent drawing
  • US8804872B1 patent drawing

AI summary

A method of dynamically calculating and updating the Volterra kernels used by the Digital Pre Distortion engine based on output power, input signal bandwidth, multicarrier configuration, frequency response and power amplifier temperature. A dominant Volterra kernels searching DSP engine based on innovation bases with minimum RMS error selection is implemented to continuously update the Volterra kernels set used in DPD to model the power amplifier non linear behaviors.