LTE Sounding Reference Signal Processing for Low-Complexity Channel Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SRS processing in LTE wireless networks faces challenges in accurately estimating channel gain, noise variance, and timing offset, which are crucial for UL MU-MIMO/SIMO and DL eigen-beamforming, due to high computational complexity and interference from multiplexed UEs.
Innovation Solution
A low-complexity time-domain based SRS receiver with group-UE cyclic shift de-multiplexing and noise variance estimation techniques, including per-antenna per-sub-carrier channel estimation, noise removal, and timing offset estimation, to improve channel estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional SRS processing is used to estimate channel gain, noise variance, and timing offset, then channel estimation can be performed, but computational complexity is high and accuracy is degraded due to interference from multiplexed UEs
Solution Approach 1:
The patent segments the SRS signal processing by separating channel estimation, noise variance estimation, and timing offset estimation into distinct processing stages. It also segments the SRS resources into different combs and cyclic shifts, allowing independent processing of each segment to reduce overall computational complexity while maintaining accuracy
Solution Approach 2:
The patent extracts and removes the estimated noise variance component from the channel estimation process. By taking out the noise variance as a separate estimated parameter and subtracting it from the total received signal power, the channel gain estimation becomes more accurate and unbiased, directly addressing the measurement precision issue
2Reliability
If SRS processing is performed for frequency dependent scheduling and link adaptation, then scheduling performance is improved, but interference from multiple multiplexed UEs degrades the estimation accuracy
Solution Approach 1:
The patent implements feedback mechanisms where the estimated channel gain, noise variance, and timing offset are fed back into the scheduling and link adaptation processes. The SINR estimation derived from these parameters provides feedback for optimizing resource allocation and modulation and coding scheme selection, improving overall system reliability
Solution Approach 2:
The patent introduces SINR as an intermediary parameter that mediates between the raw channel estimates and the final scheduling decisions. By computing SINR that accounts for both channel gain and noise variance, it provides a more accurate metric for frequency dependent scheduling and link adaptation, improving reliability while filtering out interference effects
3Measurement precision
If unbiased channel gain estimation is achieved by removing noise variance, then measurement accuracy is improved, but additional computational steps are required
Solution Approach 1:
The patent merges the noise variance estimation and channel gain estimation processes into a unified framework. By estimating noise variance from the same SRS measurements used for channel estimation and combining these estimates through a single computational flow, it achieves unbiased channel gain estimation without proportionally increasing processing steps
Solution Approach 2:
The patent enables the SRS signal itself to serve multiple purposes: it provides both the channel information and the noise statistics needed for unbiased estimation. The same received SRS signals that carry channel gain information also contain noise variance information, allowing the system to self-service by extracting both parameters from the same measurement without requiring separate dedicated signals
Data Source
AI summary
A wireless communication receiver including a serial to parallel converter receiving an radio frequency signal, a fast Fourier transform device connected to said serial to parallel converter converting NFFT corresponding serial signals into a frequency domain; an EZC root sequence unit generating a set of root sequence signals; an element-by-element multiply unit forming a set of products including a product of each of said frequency domain signals from said fast Fourier transform device and a corresponding root sequence signal, an NSRS-length IDFT unit performing a group cyclic-shift de-multiplexing of the products and a discrete Fourier transform unit converting connected cyclic shift de-multiplexing signals back to frequency-domain.