Digital Receiver Polyphase Filter Bank for Narrowband Frequency-Slice Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing communication systems face challenges in efficiently processing broadband signals, particularly in extracting multiple frequency-slices at arbitrary sampling rates, which requires efficient resampling and filtering methods to reduce processing complexity and power consumption.

Innovation Solution

A digital receiver architecture that employs a Polyphase filter bank with interpolated sub-filters, derived from a prototype Low-Pass Filter, allows for selective filtering and down-conversion of broadband signals into multiple frequency-slices at a lower sampling rate, using a memory buffer and cyclic shifting to align samples for efficient IFFT processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional broadband signal processing methods are used, then signal processing can be performed, but processing complexity and power consumption increase

Engineering Contradiction:
Improvesignal processing efficiencyVSAvoidprocessing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides the broadband signal into multiple narrowband frequency-slices using polyphase filter banks. This segmentation allows each slice to be processed independently at lower sampling rates, reducing overall processing complexity while maintaining efficient signal extraction. The filter bank structure segments the frequency spectrum into manageable portions that can be handled by simpler processing stages.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts specific frequency-slices from the broadband signal at arbitrary sampling rates using selective filtering and down-conversion. By taking out only the required frequency portions rather than processing the entire broadband signal, the system reduces processing complexity and power consumption while maintaining the ability to extract multiple frequency slices simultaneously.

Inventive Principle:
Principle #2Taking out (Extraction)

2Adaptability or versatility

If arbitrary sampling rates are used for frequency-slice extraction, then flexible signal processing is achieved, but resampling and filtering complexity increases

Engineering Contradiction:
Improvesampling rate flexibilityVSAvoidresampling and filtering complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent employs a polyphase filter bank structure that can handle arbitrary sampling rate conversions through a unified framework. The filter bank is designed to work with any sampling rate ratio, providing universal functionality for resampling and frequency-slice extraction. This multi-functional approach allows the same structure to adapt to different sampling rates without requiring separate dedicated circuits for each rate, thereby reducing overall complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent changes the sampling rate parameter dynamically to achieve arbitrary output rates for frequency-slice extraction. By using rational sampling rate ratios and adjusting the decimation factor accordingly, the system achieves flexible sampling rate adaptation. The filter bank coefficients and structure are configured based on the desired sampling rate ratio, allowing parameter-driven adaptation without structural redesign.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If multiple frequency-slices are extracted simultaneously, then processing efficiency improves, but filter bank complexity increases

Engineering Contradiction:
Improvefrequency-slice extraction efficiencyVSAvoidfilter bank complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent merges multiple filtering operations into a single polyphase filter bank structure that simultaneously extracts multiple frequency-slices. Instead of using separate filter banks for each frequency slice, the system combines them into one unified structure that processes all slices in parallel through shared computational resources. This merging reduces overall filter bank complexity while maintaining the ability to extract multiple frequency slices simultaneously.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent transforms the frequency-domain extraction problem into a time-domain polyphase decomposition problem. By using polyphase filter banks, the system extracts multiple frequency-slices simultaneously through time-domain filtering and decimation operations. This dimensional transformation allows efficient simultaneous extraction of multiple frequency slices by exploiting the structure of the polyphase components, reducing the complexity compared to traditional frequency-domain approaches.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10009203B1Small-footprint digital receiver for extracting narrowband frequency-slices
Publication Date: 2018.06.26 CAPACICOM LTD

AI summary

A digital receiver includes a memory buffer and circuitry. The memory buffer stores samples of an input signal having a sampling rate Fsi. The circuitry includes P filter-families that each includes N filters derived from a prototype Low-Pass Filter (LPF) whose stopband frequency depends on an output sampling rate Fso=α·Fsi. The circuitry is configured to set a sampling time according to Fso, to select, based on the sampling time, multiple filter-families out of the P filter-families, and to construct, based on the N filters in each of the selected filter-families, N interpolated filters that are each aligned to the sampling time. The circuitry is further configured to calculate N filtered samples by applying the N interpolated filters to the samples in the memory buffer, and to generate a sample for an output signal of a frequency-slice of the input signal, by digitally down-converting the N filtered samples.