Differential Covariance Feedback for Wireless Beamforming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing closed-loop transmission feedback methods in wireless communication systems are not robust enough to handle poor channel conditions and are prone to high feedback errors, with the quality of the covariance matrix not improving over time.
Innovation Solution
The differential-error-resistant covariance (DERC) feedback method calculates a covariance matrix at time t based on a previous quantized matrix and a forgetting factor, creating a DERC feedback message that is transmitted with pilots, allowing the base unit to detect and use the DERC values to compute an accurate covariance matrix estimate for beamforming, thereby resisting feedback errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional closed-loop transmission feedback methods are used, then the system can operate with simple feedback mechanisms, but the feedback accuracy deteriorates in poor channel conditions
Solution Approach 1:
The patent implements differential feedback where the mobile station feeds back only the change (difference) between the current covariance matrix and the previous quantized covariance matrix, rather than the full covariance matrix. This differential feedback mechanism reduces feedback overhead and improves reliability in poor channel conditions while maintaining measurement precision through the use of a forgetting factor to combine historical and current information.
Solution Approach 2:
The patent changes the feedback parameter from transmitting the complete covariance matrix to transmitting only the differential update values. By parameterizing the feedback as differences rather than absolute values, the system achieves better reliability in noisy channels. Additionally, the forgetting factor parameter is introduced to weight the contribution of past covariance information, allowing the system to adapt to changing channel conditions while maintaining accuracy.
2Loss of information
If quantization is applied to reduce feedback overhead, then the feedback message size decreases, but the feedback error increases
Solution Approach 1:
The patent segments the covariance matrix feedback into two parts: the previously quantized covariance matrix (stored at both base station and mobile station) and the current differential update values. Only the differential part needs to be transmitted, which significantly reduces the amount of information that requires quantization and transmission. This segmentation approach minimizes information loss while maintaining reliability.
Solution Approach 2:
The system performs preliminary quantization and storage of the covariance matrix at both the base station and mobile station before feedback transmission. By pre-quantizing and storing the baseline covariance matrix, the system prepares the reference data needed for differential encoding, reducing the quantization burden on the feedback channel and improving error resistance.
3Measurement precision
If the covariance matrix is updated frequently to track channel changes, then the channel tracking accuracy improves, but the feedback overhead increases
Solution Approach 1:
The patent implements differential feedback where only the changes in the covariance matrix are transmitted rather than the complete matrix. This allows frequent updates to track channel variations while keeping feedback data volume low. The differential encoding efficiently represents small changes that occur in rapidly varying channels.
Solution Approach 2:
The system dynamically adapts the covariance matrix update process by combining historical quantized values with current measurements through a forgetting factor. This dynamic approach allows the system to track channel changes effectively while maintaining efficient feedback rates, as the differential updates naturally adapt to the rate of channel variation.
Data Source
AI summary
A method and apparatus for providing channel feedback is provided herein. During operation a covariance matrix at time t (R) is calculated as a function of a received downlink signal. Matrix Ct is also calculated and is based on a previous quantized covariance matrix (Rqt−1), the covariance matrix (R) at time t, and a forgetting factor (γ) that is applied to Rqt−1. The Ct is then used to create a DERC feedback message (signal or waveform) and may be transmitted with pilots on a proper feedback channel to a base unit. The base unit receives the feedback (Ct) as a DERC waveform on a proper feedback channel. The base unit uses non-coherent or coherent detection to detect the DERC values send by the remote unit and uses the DERC values with a previous quantized covariance matrix estimate, a forgetting factor, and a weighting value to compute a covariance matrix estimate to use for beamforming. The base unit then uses the covariance matrix estimate to determine appropriate channel beamforming weights, and instructs transmit beamforming circuitry to use the appropriate weights.