Signal Feature Extraction Using Extended Impulse Response Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In filter bank analysis, overlapping bands of band-pass filters cause signal leakage to adjacent filters, leading to redundant analysis results and difficulty in analyzing feature patterns of input signals.
Innovation Solution
A signal feature extraction apparatus that generates an extended impulse response matrix and calculates expansion coefficients using the input signal vector and the extended impulse response matrix to reduce signal leakage between adjacent band-pass filters and in the time axis direction, allowing for flexible handling of diverse feature patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If conventional filter bank analysis with overlapping bands is used, then frequency coverage is improved, but signal leakage to adjacent filters occurs causing redundant analysis results
Solution Approach 1:
The patent segments the signal representation by using multiple impulse response vectors (including time-shifted versions) as basis functions. This segmentation allows the signal to be decomposed into distinct components, each associated with specific time shifts and frequency bands, thereby reducing leakage between adjacent filters while maintaining comprehensive frequency coverage.
Solution Approach 2:
The patent extends the analysis from a single impulse response per filter to multiple impulse responses with different time shifts (k values from -2L to 2L). This adds a time-shift dimension to the filter bank analysis, creating an extended impulse response matrix that captures signal features more precisely and reduces leakage to adjacent filters.
2Measurement precision
If multiple band-pass filters with overlapping bands are applied, then frequency resolution is improved, but analysis complexity increases due to redundant outputs
Solution Approach 1:
The patent extracts only the relevant expansion coefficients from the full decomposition result. By selecting specific coefficients corresponding to particular time shifts and frequency bands, the method obtains the necessary frequency resolution information without processing all redundant outputs, thereby reducing analysis complexity while maintaining frequency resolution.
Solution Approach 2:
The patent changes the representation parameters by using an extended impulse response matrix with multiple time-shifted versions of each filter's impulse response. This parameter change allows the system to achieve better frequency resolution through a more structured decomposition, reducing the effective complexity of analyzing overlapping filter outputs.
3Productivity
If standard impulse response vectors are used, then computational simplicity is maintained, but signal feature extraction accuracy deteriorates due to time-axis leakage
Solution Approach 1:
The patent performs preliminary action by pre-defining the extended impulse response matrix with multiple time-shifted impulse response vectors before signal analysis. This pre-computation of the basis functions with various time shifts enables accurate feature extraction without increasing the complexity of the actual signal processing, as the temporal distribution characteristics are already embedded in the basis functions.
Data Source
AI summary
The purpose of the present invention is to reduce leakage of an output signal between band-pass filters and in the time axis direction. A signal input part 1 vectorizes an input signal x(n). An extended impulse response matrix generation part 2 generates an extended impulse response matrix He in which impulse response vectors using impulse response sequences of band-pass filters as elements are extended in the time axis direction. An expansion coefficient calculation part 3 calculates an expansion coefficient vector {circumflex over ( )}y(n) using an input signal vectors x(n) and the extended impulse response matrix He. A signal output part 4 outputs at least one of expansion coefficients corresponding to a center vector of the extended impulse response matrix He of the expansion coefficient vector {circumflex over ( )}y(n).