Crest Factor Reduction with Running Maximum Peak Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing crest factor reduction (CFR) methods, particularly peak cancellation crest factor reduction (PCCFR) and window CFR, face challenges such as inaccurate peak detection due to low oversampling rates, missed peaks, and the creation of new peaks during pulse cancellation, leading to inefficiencies and increased latency in systems like base stations.
Innovation Solution
A system and method for CFR that includes a peak detector for accurate peak location and magnitude estimation using polynomial fits, combined with a window CFR module to prevent peak misses and over-cancellations, utilizing a running maximum filter and gain correction to refine peak cancellation, thereby improving peak detection and reduction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the oversampling rate is decreased to reduce processing complexity, then device complexity is reduced, but measurement precision of peak magnitude deteriorates
Solution Approach 1:
The patent applies preliminary action by performing polynomial fitting on oversampled peak data before final peak magnitude calculation. This preprocessing step captures the underlying signal trend, allowing accurate peak estimation even when the actual peak occurs between discrete samples. The polynomial model is fitted to neighboring samples and then evaluated at the true peak location to obtain accurate magnitude estimates without requiring high oversampling rates.
2Reliability
If multiple processing stages are used to address missed peaks and create new peaks, then reliability of peak cancellation is improved, but loss of time increases
Solution Approach 1:
The patent implements feedback by using a running maximum filter that continuously monitors the signal and compares current samples with previous maximum values. This feedback mechanism allows the system to detect peaks in real-time as they occur, rather than requiring multiple processing passes. The filter maintains a record of recent maximum values and identifies peaks when the current sample exceeds this running maximum, enabling single-pass processing with high reliability.
Solution Approach 2:
The patent replaces the mechanical iterative pulse cancellation approach with a computational polynomial fitting method. Instead of repeatedly applying cancellation pulses and reprocessing the signal multiple times, the system uses polynomial models to directly calculate and correct peak magnitudes in a single pass. This substitution of computational mathematics for iterative mechanical processing dramatically reduces latency while maintaining or improving peak cancellation accuracy.
Data Source
AI summary
A system for crest factor reduction (CFR) includes a peak detector configured to receive an input signal (xk); a running maximum filter configured to generate a scaling factor based on a window gain (Gk) and a filter length, wherein the window gain (Gk) is based on the input signal (xk) and a threshold value (T); a window CFR gain filter configured to generate a gain correction (Fk) based on the scaling factor and the filter length; a delay configured to delay the input signal (xk) to generate a delayed input signal; a multiplier configured to multiply the gain correction (Fk) by the delayed input signal to obtain a peak correction value; and an adder configured to determine an output signal (yk) based on the peak correction value and the delayed input signal.