Blind Signal Separation Using SIMD Inverse Matrix Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing blind signal separation methods require significant time for calculating inverse matrices, particularly when handling broadband signals, leading to prolonged processing times for signals like speech signals.
Innovation Solution
The use of SIMD commands to accelerate the calculation of inverse matrices by employing tensor data structures and efficient determinant and matrix inversion methods, such as Inverse Matrix Calculation Method 1 and Inverse Matrix Calculation Method 2, reduces the computational time required for blind signal separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inverse matrices are calculated using conventional methods, then accurate parameter estimation is achieved, but processing time becomes excessively long
Solution Approach 1:
The patent segments the inverse matrix calculation process into multiple parallel computational tasks that can be executed simultaneously. By dividing the calculation workload across multiple processing units, the patent achieves both accurate parameter estimation and reduced processing time, directly resolving the contradiction between precision and speed.
Solution Approach 2:
The patent transitions from sequential single-threaded inverse matrix calculation to parallel multi-threaded computation, adding a temporal dimension to the processing. This dimensional change allows multiple calculations to proceed simultaneously rather than sequentially, maintaining accuracy while dramatically reducing overall processing time.
2Manufacturing precision
If the number of discrete time points T increases to handle broadband signals, then signal processing completeness is improved, but calculation time becomes dominant
Solution Approach 1:
The patent segments the calculation across multiple time points into independent parallel tasks. Each time point's inverse matrix calculation can be performed simultaneously with others, allowing complete processing of all T discrete time points without the calculation time becoming dominant, thus resolving the contradiction between completeness and speed.
Solution Approach 2:
The patent performs preliminary organization of calculation tasks for all time points before execution. By pre-processing and structuring the computational workload, the system can efficiently execute parallel calculations across all T time points, maintaining signal processing completeness while preventing calculation time from becoming the bottleneck.
Data Source
AI summary
The signal separation device includes: cross product calculation means receiving an input of an observed signal that is a mixture of a plurality of target signals, and calculating a cross product of the observed signal; model calculation means updating a parameter of a model for estimating the cross product with a predetermined algorithm using an inverse matrix of a matrix that represents an estimate of the cross product; inverse matrix calculation means calculating the inverse matrix of a matrix by a SIMD command when the parameter is updated; and separation means calculating the target signals using a matrix representing an estimate of the cross product, the updated parameter, and the observed signal.