Beamformer Filter Coefficient Optimization Using Phase Regularization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional LCMV beamformer design techniques fail to consider the relationship of filter coefficients between adjacent frequency bins, leading to instability and poor quality beamformers.
Innovation Solution
A filter coefficient optimization method that incorporates a regularization term to account for the phase difference between adjacent frequency bins, using cost functions and optimization problems to determine optimal filter coefficients that stabilize the beamformer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional LCMV beamformer design techniques are used to optimize filter coefficients, then the target sound emphasis performance is improved, but the beamformer stability deteriorates due to ignoring the relationship between adjacent frequency bins
Solution Approach 1:
The patent combines the individual frequency bin optimization with a regularization term that merges information from adjacent frequency bins. The cost function integrates both the LCMV objective for each frequency bin and a regularization component that considers the relationship between adjacent bins, thereby achieving both performance optimization and stability.
Solution Approach 2:
The patent introduces a regularization parameter that controls the influence of adjacent frequency bins on the current frequency bin's filter coefficient optimization. By adjusting this parameter, the system balances between achieving optimal target sound emphasis and maintaining beamformer stability across frequency bins.
2Device complexity
If filter coefficients are optimized independently for each frequency bin, then the optimization complexity is reduced, but the beamformer quality deteriorates
Solution Approach 1:
The patent applies partial action by considering only the relationship with adjacent frequency bins rather than all frequency bins, and by using a regularization term that provides a controlled amount of coupling between bins. This approach improves beamformer quality without fully coupling all frequency bins, thus maintaining reasonable optimization complexity.
3Stability of the object's composition
If a regularization term considering phase difference between adjacent frequency bins is added, then the beamformer stability is improved, but the computational load increases
Solution Approach 1:
The regularization term focuses on local relationships between adjacent frequency bins rather than global relationships across all bins. This local approach improves beamformer stability while keeping the computational load manageable by only considering immediate neighbors in the frequency domain.
Data Source
AI summary
Provided is a filter coefficient optimization technology that makes it possible to design a stable beamformer having a good quality by considering the relationship of a filter coefficient between adjacent frequency bins. A filter coefficient optimization apparatus includes an optimization unit that calculates an optimum value of a filter coefficient w={w1, . . . , wF} (wf is a filter coefficient of a frequency bin f) of a beamformer that emphasizes sound (target sound) from D sound source, af,d being an array manifold vector in the frequency bin f corresponding to a sound wave that comes from an angular direction θd in which a sound source d exists, the sound wave being a plane wave, the optimization unit calculating the optimum value based on an optimization problem of a cost function defined using a sum of a sum of a cost function LMV_f(wf) and a predetermined regularization term, under a predetermined constraint condition, the predetermined regularization term being defined using a difference in phase between adjacent frequency bins relevant to a response wfHaf,d of the beamformer in the frequency bin f for the angular direction θd.