Blind Interference Cancellation via Covariance Matrix Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing interference cancellation techniques in CDMA wireless communication systems, such as LMMSE and NLIC receivers, face high computational complexity and require accurate estimation of interference signals, which is challenging especially in WCDMA downlink with limited knowledge of spreading factors and inter-channel interferences.
Innovation Solution
The method employs blind interference cancellation using second-order statistics to estimate and remove dominant interference contributions from the received covariance matrix, reducing computational complexity and processing delay by iteratively canceling weaker interferences, and utilizing subspace techniques to project interference on maximum eigenvalues and eigenvectors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LMMSE based interference cancellation receivers estimate the covariance matrix and solve large matrix inversion, then interference cancellation performance is improved, but computational complexity increases
Solution Approach 1:
The patent segments the interference cancellation process into two distinct stages: a training phase where the covariance matrix is estimated and stored, and a data phase where the pre-computed covariance matrix is reused for interference cancellation. This segmentation allows the computationally intensive matrix inversion to be performed only once during training, rather than repeatedly during data processing, thereby resolving the contradiction between achieving good interference cancellation performance and maintaining low computational complexity during operation.
2Reliability
If NLIC receivers estimate interference items and subtract them from received signal, then interference cancellation is achieved, but processing delay increases
Solution Approach 1:
The patent performs preliminary action by estimating and storing the covariance matrix during a training phase before actual data reception. The pre-computed covariance matrix and its inverse are then directly applied during the data phase for rapid interference cancellation without requiring real-time matrix inversion. This preliminary computation eliminates the time-consuming matrix operations during critical data processing, thereby resolving the contradiction between achieving effective interference cancellation and minimizing processing delay.
3Measurement precision
If interference covariance matrix is estimated with measured signal samples, then interference estimation is improved, but error in power ratio estimation increases
Solution Approach 1:
The patent employs feedback by using the estimated covariance matrix to compute the inverse matrix, which is then applied to the received signal to cancel interference. The effectiveness of this cancellation provides feedback on the accuracy of the covariance matrix estimation. By iteratively refining the covariance matrix estimation based on the interference cancellation results and using the pre-computed inverse matrix to minimize estimation errors, the system resolves the contradiction between improving interference covariance matrix estimation and maintaining accurate power ratio estimation.
Data Source
Figure 1~2
Figure 3~4
Figure 5A
AI summary
Aspects of the present invention include methods, systems, and computer-readable medium for canceling interference in wireless communication. The method includes receiving wireless CDMA communication signals using one or more antennas at least from a first entity via a first communication channel and a second entity via a second communication channel, determining a set of known characteristics associated with the first entity, the first set of characteristics comprising a first signal strength, a first synchronization information, and an first channel identification information, and determining an aggregate signal matrix based on signals received from at least the first entity and the second entity. The method further includes determining a covariance matrix associated with the aggregate signal value, determining a reference signal matrix based on the set of known characteristics, calculating an interference matrix by subtracting the reference signal matrix from the covariance matrix, and removing the interference estimation from the communication signals.