OFDM Decoder Phase Noise Suppression With Iterative TD Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional phase noise suppression methods for orthogonal frequency division multiplexing (OFDM) systems face challenges such as high computational complexity, significant pilot overhead, and processing delays, particularly when dealing with large modulation constellations like 256-QAM and higher, which deteriorate system performance and spectral efficiency.
Innovation Solution
A unified framework combining frequency domain (FD) and time domain (TD) phase noise suppression techniques, utilizing an iterative process with linear interpolation in the time domain (LI-TD) to refine phase noise estimation, reducing reliance on matrix inversions and maintaining low pilot symbol overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional frequency domain phase noise suppression methods are used, then phase noise compensation is achieved, but computational complexity increases due to matrix inversions
Solution Approach 1:
The patent extracts and removes the computationally intensive matrix inversion operation from the phase noise suppression algorithm. By using an alternative estimation approach that avoids matrix inversions, the method maintains phase noise compensation performance while significantly reducing computational complexity at the receiver.
Solution Approach 2:
The patent changes the computational parameters of the phase noise estimation process by switching from frequency domain matrix-based methods to a time domain approach using linear interpolation. This parameter change transforms the computational burden from complex matrix operations to simpler arithmetic operations.
2Measurement precision
If high pilot overhead is used for phase noise estimation, then estimation accuracy improves, but spectral efficiency decreases
Solution Approach 1:
The patent enables the system to use existing pilot symbols already present in the OFDM signal structure for phase noise estimation, rather than requiring additional dedicated pilot symbols. The linear interpolation method allows accurate phase noise estimation to be obtained from the available pilots without increasing overhead, making the system self-sufficient with standard pilot configurations.
3Reliability
If complex matrix inversion methods are used for phase noise suppression, then suppression performance improves, but processing delay increases
Solution Approach 1:
The patent removes the computationally intensive matrix inversion step from the processing chain, directly reducing the processing time required for phase noise suppression. The alternative method using linear interpolation in the time domain executes much faster while maintaining suppression effectiveness.
Solution Approach 2:
The patent replaces expensive, time-consuming matrix inversion operations with cheaper, faster arithmetic operations. The simplified algorithm uses basic mathematical operations that can be executed rapidly, reducing the time penalty associated with phase noise compensation.
4Ease of manufacture
If inexpensive local oscillators are used, then cost is reduced, but phase noise increases
Solution Approach 1:
The patent converts the harmful phase noise introduced by inexpensive local oscillators into a manageable parameter through digital signal processing. By using linear interpolation to estimate and compensate for the phase noise in the time domain, the system transforms the disadvantage of using low-cost oscillators into an opportunity to demonstrate effective digital compensation techniques.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Phase noise (PN) is suppressed in an OFDM signal. In an initial iteration, an estimation of the PN samples in an OFDM signal vector is produced and the OFDM signal is demodulated using the initial estimation of the PN to generate constellation symbols for the initial iteration. In an additional iteration, an Inverse Fast Fourier Transform of constellation symbols generated in a preceding iteration is calculated to reconstruct preceding samples of the transmitted signal vector. A PN effect on the reconstructed samples is estimated. A next estimation of the PN in a next signal vector is produced based on the estimated PN effect on the reconstructed samples. The next signal vector is demodulated using the next estimation of the PN to generate constellation symbols for the additional iteration. A predetermined maximum number of additional iterations may be used.