Block Equalization Systolic Array for Multipath Channel Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current block equalization techniques in wireless communications face challenges in efficiently handling multi-path and fading channel conditions, particularly in wideband waveforms, requiring significant computational complexity and knowledge of the channel for effective coefficient updates.
Innovation Solution
A wireless communications device employing a demodulator systolic array with modules for channel estimation, autocorrelation matrix generation, and channel-matched filtering, utilizing techniques like LU decomposition and back-substitution to estimate unknown symbol portions based on known symbol portions, thereby reducing computational overhead and maintaining symmetry for efficient block equalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If symbol-based equalization is used to track and compensate for time-varying multipath channel, then equalization performance is improved, but computing overhead increases considerably due to symbol-by-symbol coefficient updates
Solution Approach 1:
The patent segments the continuous coefficient update process into discrete block-based updates. Instead of updating equalizer coefficients for every symbol, the system processes blocks of symbols together, maintaining tracking performance while reducing the frequency of computationally intensive updates. This segmentation approach allows the system to achieve good equalization performance without the prohibitive computing overhead of symbol-by-symbol updates.
Solution Approach 2:
The patent performs preliminary channel estimation and coefficient calculation for a block of symbols before actual data processing. By pre-computing the equalizer coefficients for the entire block based on pilot symbols or channel estimates, the system avoids repeated calculations during symbol processing, thereby reducing real-time computing overhead while maintaining equalization effectiveness.
2Device complexity
If block equalization is used to decrease complexity for coefficient updates, then computing overhead is reduced, but channel knowledge is required which generally requires channel estimation
Solution Approach 1:
The patent performs channel estimation preliminarily using embedded pilot symbols or known sequences within the transmitted signal. By estimating the channel characteristics before data detection and using these estimates for block equalization, the system obtains the necessary channel knowledge without adding complex real-time measurement mechanisms. The preliminary channel estimation enables subsequent block-based coefficient updates with reduced computing overhead.
Solution Approach 2:
The patent introduces pilot symbols or known reference sequences as intermediaries between the transmitter and receiver. These intermediary elements carry channel information that facilitates estimation without requiring direct measurement of the unknown data symbols. The pilot symbols act as a bridge, providing the necessary channel knowledge for block equalization while keeping the system complexity manageable.
3Measurement precision
If Gaussian elimination is used to solve linear equations in block equalization, then solution accuracy is improved, but computational complexity increases significantly
Solution Approach 1:
The patent transforms the original linear system into a form suitable for back-substitution by applying triangular decomposition (such as LU or Cholesky decomposition). This parameter transformation changes the matrix structure from a general dense matrix to a triangular matrix, enabling efficient solution through back-substitution while maintaining numerical accuracy. The transformation preserves the solution accuracy of Gaussian elimination but reduces the computational complexity from O(n³) to O(n²).
Data Source
AI summary
A wireless communications device may include a wireless receiver receiving signals having alternating known and unknown symbol portions over a channel, and a demodulator systolic array. The demodulator systolic array may include a channel estimation module generating respective channel estimates for each unknown symbol portion based upon the known symbol portions. An autocorrelation module may generate autocorrelation matrices based upon the channel estimates. A channel match filter module may generate respective channel matching coefficients for the unknown symbol portions, and a factorization module may divide the autocorrelation matrices into respective upper and lower autocorrelation matrices. A transformation module may transform the channel matching coefficients into upper and lower channel matching coefficients. A back-substitution module may determine the unknown symbol portions by estimating the unknown symbol portions based upon autocorrelation matrices and channel matching coefficients, lower autocorrelation matrices and channel matching coefficients, and combining the estimates.