2D Look-Up Table Segmentation for RF Signal Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital pre-distortion techniques using polynomial models and two-dimensional look-up tables are inadequate for modeling non-linear systems, particularly in situations with discontinuities and are inefficient in terms of silicon and power usage, especially in battery-operated devices.
Innovation Solution
The use of two-dimensional look-up tables with bilinear interpolation to model non-linear systems, where parameters are directly optimized from physical data to minimize mean square error, allowing for reduced size and improved accuracy in digital pre-distortion applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If two-dimensional look-up tables are used to implement non-linear functions, then the ability to capture discontinuities and model complex RF signals is improved, but the silicon area and power consumption increase due to N2 complexity
Solution Approach 1:
The patent segments the complex two-dimensional look-up table into multiple smaller one-dimensional look-up tables. Instead of storing a single large 2D table with N2 complexity, the system divides it into multiple 1D tables with significantly reduced size, while maintaining the ability to model complex non-linear RF signals through coordinated use of these segmented tables.
Solution Approach 2:
The patent transforms the problem from a two-dimensional look-up table structure to a combination of one-dimensional look-up tables. By changing the dimensional approach and using multiple 1D tables instead of a single 2D table, the system reduces silicon area complexity while preserving the modeling capability for discontinuous and complex RF signal characteristics.
2Reliability
If two-dimensional look-up tables are used to implement non-linear functions, then the ability to capture discontinuities and model complex RF signals is improved, but power consumption increases
Solution Approach 1:
The patent segments the complex two-dimensional look-up table into multiple smaller one-dimensional look-up tables. Instead of storing a single large 2D table with N2 complexity, the system divides it into multiple 1D tables with significantly reduced size, while maintaining the ability to model complex non-linear RF signals through coordinated use of these segmented tables.
Solution Approach 2:
The patent uses multiple one-dimensional look-up tables that collectively represent the functionality of a single large two-dimensional table. By creating these smaller 1D table copies and coordinating their output, the system achieves the same modeling accuracy with reduced power consumption from smaller memory access and processing requirements.
3Device complexity
If polynomial models are used for digital pre-distortion, then implementation is simpler, but they fail to capture functions with discontinuities
Solution Approach 1:
The patent introduces look-up tables as an intermediary structure between the simple polynomial approach and the complex continuous function requirement. The look-up tables store pre-computed values that capture discontinuities and non-linearities, serving as a mediator that provides both implementation simplicity (through table lookup) and high modeling accuracy (through stored empirical data).
Data Source
AI summary
Methods and apparatus are provided for modeling of a physical system using two-dimensional look-up tables. A method can include determining coefficients for a piece-wise polynomial function that estimates a non-linear function, storing the coefficients as entries in one or more two-dimensional look up tables in the memory, receiving a complex value input corresponding to an actual input value, identifying two closest entries, a first closest entry and a second closest entry, to the complex input value in a two-dimensional look-up table of the one or more two-dimensional look-up tables, evaluating, using the one or more hardware processors, the piece-wise polynomial function at the complex input value twice, a first evaluation using the coefficients corresponding to the first closest entry and a second evaluation using the coefficients corresponding to the second closest entry, to generate first and second output values, and generating an output value by performing a linear interpolation between the first and second output values.