Complex Polynomial Vector Processor Reduces Signal Non-Linearity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal processing technologies face challenges in addressing non-linearity in mixed signal processing, particularly in high-performance conditions, where devices often operate outside linear ranges, requiring complex and resource-intensive correction methods that are not always effective.
Innovation Solution
A complex polynomial vector processor system that includes a data processing unit and a co-efficient feeder unit, utilizing a normalized cordic transform, polynomial power calculation, and programmable gains to convert high-speed data streams into polar-like format, reducing non-linearity and power consumption by dynamically calculating polynomial coefficients and eliminating the need for shadow memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If post-processing or pre-processing methods are used to correct non-linearity, then non-linearity correction is achieved, but hardware resources are excessively consumed and system complexity increases
Solution Approach 1:
The patent changes the parameter representation from Cartesian coordinates to polar-like format, where the magnitude and phase are processed separately. This parameter transformation enables more efficient correction of non-linearity by working in a coordinate system that naturally separates amplitude and phase components, reducing the computational complexity required for non-linearity compensation.
Solution Approach 2:
The system dynamically adjusts polynomial coefficients based on operating conditions rather than using fixed correction algorithms. The co-efficient feeder unit provides dynamically calculated coefficients that adapt to varying environmental conditions and device characteristics, enabling effective non-linearity correction without requiring excessive hardware resources for multiple fixed algorithms.
2Productivity
If fixed algorithmic solutions are used for non-linearity correction, then real-time processing is achieved, but hardware resource consumption increases significantly
Solution Approach 1:
The correction process is segmented into distinct functional units: a co-efficient feeder unit that dynamically provides polynomial coefficients, a data processing unit that performs polynomial calculations, and separate transformation units (CORDIC) for coordinate conversion. This segmentation allows real-time processing with optimized resource allocation, as each unit performs a specific function efficiently rather than requiring complete duplicate correction algorithms for different scenarios.
Solution Approach 2:
The polynomial-based correction approach serves multiple functions: it corrects non-linearity, adapts to different operating conditions, and works with varying device characteristics all through a single unified algorithmic framework. The same polynomial correction mechanism handles different scenarios by adjusting coefficients rather than requiring separate fixed algorithms, reducing overall hardware resource requirements.
3Measurement precision
If polynomial coefficients are dynamically calculated, then non-linearity correction accuracy is improved, but processing complexity increases
Solution Approach 1:
Polynomial coefficients are pre-calculated and stored in memory before actual signal processing begins. The co-efficient feeder unit retrieves these pre-computed coefficients during operation rather than calculating them in real-time, which maintains high correction accuracy while avoiding the processing complexity of dynamic coefficient calculation during signal processing.
Solution Approach 2:
The patent introduces an intermediary polynomial approximation layer between the non-linear device output and the final linearized signal. Instead of directly inverting complex non-linear relationships, the polynomial serves as an intermediary representation that simplifies the correction process while maintaining accuracy, and the CORDIC transform acts as another intermediary to convert between coordinate systems efficiently.
Data Source
AI summary
A system for reducing non-linearity in mixed signal processing using complex polynomial vector processor 102 is provided. The complex polynomial vector processor 102 includes a data processing unit (104) and a co-efficient feeder unit (106). The data processing unit (104) converts a high-speed data stream into a polar-like format (PL) data and calculates required polynomial powers for the high-speed data stream using the PL format data. The data processing unit (104) includes a multiplier accumulator (MAC) unit (206) that generates processed high-speed data and a delay unit (208) that combines time separated input with the processed high-speed data to generate output data with reduced non-linearity.