Frequency-Reconfigurable Digital Filter for Flexible Passband Tuning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital filters face limitations in reconfiguring pass band characteristics, requiring extensive coefficient storage and being unable to freely adjust frequency bands beyond pre-stored settings, which restricts their adaptability and efficiency in applications like mobile communication systems and equalizers.
Innovation Solution
A frequency reconfigurable digital filter is developed using a sampling kernel with a sampling constant and Multi images, allowing for the generation of Multi images and complementary images with constant periods, enabling easy adjustment of pass bands through a closed-form response in the time domain, and an equalizer is designed to utilize these filters for flexible frequency reconfiguration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the finite impulse response filter uses more taps and coefficients to achieve similar amplitude response as infinite impulse response filter, then the filter order is increased, but the hardware complexity including adders and multipliers increases significantly
Solution Approach 1:
The filter is divided into multiple banks, each handling a specific frequency band. Each bank contains sub-filters that process different portions of the spectrum. This segmentation allows the system to achieve high amplitude response accuracy in each band without requiring a single high-order filter with excessive hardware complexity.
Solution Approach 2:
The patent transitions from time-domain filtering to frequency-domain filtering by applying FFT (Fast Fourier Transform). This dimensional change from time to frequency domain enables efficient spectral analysis and filtering, achieving accurate amplitude response with reduced computational complexity through frequency-selective processing.
2Reliability
If the finite impulse response filter coefficients are fixed to guarantee stability and linear phase characteristic, then the filter is reliable, but the pass band cannot be reconfigured
Solution Approach 1:
The filter system implements dynamic reconfigurability where filter coefficients and bank selections can be changed in real-time to adapt to different pass band requirements. The multi-bank structure with selectable sub-filters allows the system to dynamically reconfigure its frequency response while maintaining stability through controlled coefficient transitions.
Solution Approach 2:
The patent changes filter parameters (coefficients, bank selection, sub-filter activation) to reconfigure the pass band. By systematically varying these parameters, the filter can adapt to different frequency bands and applications while maintaining the stability guarantees of FIR filters through proper coefficient design and selection.
3Adaptability or versatility
If conventional filters store multiple sets of coefficients for different pass bands, then frequency reconfiguration is possible within pre-stored bands, but the memory requirement increases and frequency bands are limited to pre-stored settings
Solution Approach 1:
The patent pre-computes and stores only the essential filter coefficients and spectral data needed for filtering operations. The multi-bank structure with shared coefficient storage and selective activation allows the system to provide multiple frequency band configurations without storing complete coefficient sets for every possible band, significantly reducing memory requirements while maintaining reconfiguration capability.
Solution Approach 2:
The filter banks share common computational resources and coefficient storage structures. Multiple frequency bands can be achieved using the same hardware resources through different bank selections and sub-filter combinations, making the system universal and adaptable to various frequency requirements without proportionally increasing memory and hardware resources.
4Measurement precision
If the filter uses a high number of taps to achieve sharp frequency selectivity, then the frequency resolution is improved, but the computational load and processing time increase
Solution Approach 1:
The patent applies FFT-based frequency domain processing to achieve sharp frequency selectivity without requiring long time-domain filters. By transforming the filtering operation to the frequency domain, the system achieves high frequency resolution through efficient spectral analysis, avoiding the computational burden of convolving with long impulse responses in the time domain.
Solution Approach 2:
The frequency spectrum is segmented into multiple bands with dedicated sub-filters in each bank. This segmentation allows each sub-filter to focus on a specific frequency range, achieving sharp frequency selectivity in targeted bands with fewer coefficients than a single high-order filter would require, thereby reducing overall computational load and processing time.
Data Source
Figure 1~3
Figure 4~6
Figure 7~8
AI summary
A frequency reconfigurable digital filter and an equalizer using the same are disclosed. The digital filter includes a sampling kernel storage section configured to store a sampling kernel for performing upsampling of a model filter response scaled by a sampling constant and generating response of Multi images which are repeatedly formed with a constant period; a complementary conversion section configured to generate response of Multi complementary images repeatedly formed with constant period in frequency domain where the Multi images are not generated, the Multi complimentary images having the same characteristic as the Multi images; and an image response operation section configured to operate response of an image corresponding to a selected band among the Multi complementary images and the Multi images. The filter reconfigures frequency to realize various band pass characteristics only through changing very small number of parameter, and a user may change easily band of the filter.