Parallel Undersampling for High-Frequency Baseband Equalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current high-frequency signal transmission and reception systems face challenges with time-interleaved analog-to-digital converters, which require complex and costly compensation for phase offsets and have high implementation costs due to the need for identical signal processing across multiple converters, and parallel digital-to-analog converters for high-volume data transmission, which also involve costly mixer setups.
Innovation Solution
The method employs parallel filters with different frequency responses for undersampled analog-to-digital conversion, allowing for efficient conversion of high-frequency signals with lower sampling frequencies, eliminating the need for complex phase offset control and reducing the complexity of mixer implementations by mapping spectral components into baseband signal components for equalization and decoupling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If time-interleaved analog-to-digital converters are used for parallel sampling, then the sampling rate can meet the Nyquist criterion for very high-frequency signals, but complex compensation mechanisms are required to correct phase offsets and ensure identical signal processing characteristics
Solution Approach 1:
The patent changes the fundamental parameter of signal processing by using bandpass sampling (undersampling) instead of lowpass sampling. This allows direct sampling of high-frequency signals at lower rates while avoiding the need for complex phase offset compensation. The sampling frequency is specifically chosen to be less than twice the highest signal frequency, creating a controlled aliasing effect that can be easily corrected through digital filtering rather than complex hardware compensation.
Solution Approach 2:
The patent extracts and removes the problematic phase offset compensation requirement entirely from the system. By using bandpass sampling with carefully selected sampling frequencies, the method eliminates the need for complex compensation mechanisms that are inherent in time-interleaved lowpass sampling approaches. The solution takes out the problematic element (phase offset sensitivity) rather than trying to compensate for it.
2Productivity
If multiple parallel analog-to-digital converters are used for high-volume data transmission, then the data rate increases, but all converters must exhibit identical signal processing characteristics which increases implementation cost and complexity
Solution Approach 1:
The patent changes the sampling approach from lowpass to bandpass sampling, which fundamentally alters how parallel converters process signals. With bandpass sampling, each converter can operate independently at lower sampling rates while capturing different portions of the high-frequency spectrum. This eliminates the requirement for identical signal processing characteristics across all parallel converters, as each converter processes a different frequency band with its own optimized parameters.
Solution Approach 2:
The patent applies local quality by allowing different signal processing characteristics for different frequency bands processed by parallel converters. Instead of requiring all converters to be identical, the system enables each converter to be optimized for its specific frequency range, with different filtering and processing parameters applied locally to each converter's output based on its assigned frequency band.
3Speed
If undersampling is used to reduce sampling frequency, then the sampling rate decreases, but the Nyquist criterion is no longer satisfied requiring sophisticated filtering and equalization
Solution Approach 1:
The patent changes the sampling paradigm by intentionally violating the traditional Nyquist criterion through bandpass sampling. Instead of sampling at rates above twice the highest frequency, the system samples at lower rates that create controlled aliasing. The key is that the sampling frequency is carefully selected based on the signal's spectral characteristics, and digital filtering is applied to separate the aliased frequency components, achieving accurate reconstruction at lower sampling rates.
Solution Approach 2:
The patent converts the harmful effect of aliasing (which normally degrades signal quality) into a beneficial feature. By intentionally allowing aliasing through undersampling and then using digital filtering to separate and reconstruct the original frequency components, the system transforms what is traditionally considered a distortion mechanism into a tool for achieving lower sampling rates. The aliasing products become useful intermediate representations that can be easily filtered and reconstructed.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
The method according to the invention and the device according to the invention for receiving at least one high-frequency signal (x(t)) using parallel and undersampled baseband signal processing generates a plurality of filtered signals (y1(t),y2(t), …,y(t)) by parallel filtering of the high-frequency signal (x(t)), wherein each individual filtering procedure is performed by means of a different filter frequency response. For each filtered signal (y1(t),y2(t), …,y(t)), an associated digitized filtered signal (y 1 (n ∙ T A ), y 2 (n ∙T A ),..., y N (n ∙T A ) is subsequently generated by analog-to-digital conversion of the respective filtered signal (y1(t),y2(t), …,y(t)), each analog-to-digital conversion being performed by undersampling. Lastly, the signal components (x 1 (n ∙ T A ),..., x M (n ∙T A ); X 1 (k ∙ ∆f),..,X M (k ∙∆f)) of the high-frequency signal (x(t)) in the digital baseband are determined by equalizing the baseband signal components (l 1 (n ∙ T A ), l 2 (n ∙T A ),..., l N (n ∙T A ); L 1 (k ∙ ∆f), L 2 (k ∙∆f),...,L N (k ∙∆f)) of the associated digitized filtered signals (y1(t),y2(t), …,y(t)). The invention also relates to a complementary method and device for transmitting at least one high-frequency signal (z(t)) using parallel and undersampled baseband signal processing and a system for transmitting at least one high-frequency signal using parallel and undersampled baseband signal processing.