Methods and apparatus for monitoring occupancy of wideband GHz spectrum and sensing and decoding respective frequency components of time-varying signals using sub-nyquist criterion signal sampling

摘要

Methods and apparatus for monitoring wideband GHz spectrum for wireless communication, and sensing and decoding respective frequency components of a time-varying signal corresponding to the monitored spectrum. Concepts relating to sparse Fast Fourier Transform (sFFT) techniques facilitate identification of one or more frequency components of a sparsely occupied spectrum by sub-sampling the signal corresponding to the monitored spectrum at a sampling rate below the Nyquist criterion. The disclosed methods and apparatus may be implemented using conventional relatively low-power wireless receivers and using off-the-shelf relatively inexpensive low-speed and low-power analog-to-digital converters (ADCs) typically employed in WiFi devices or cellular phones, in tandem with unique processing techniques based on sFFTs and sub-Nyquist criterion sampling, and have demonstrated efficacy even in scenarios where the monitored spectrum is not sparse.

主权项

1. An apparatus to sense one or more frequency components of a time-varying signal having a frequency bandwidth of interest BW and a Nyquist sampling criteria of N samples in a sampling time T, wherein N=T×BW, the apparatus comprising:
a first analog-to-digital converter (ADC) to uniformly sub-sample the time-varying signal at a first time t and at a first sampling rate of (BW/p1) samples/second, wherein p1 is less than N so as to provide a first set of B1=N/p1 samples of the time-varying signal; a second ADC to uniformly sub-sample the time-varying signal at a second time t+τ1 and the first sampling rate (BW/p1) so as to provide a second set of B1 time-shifted samples of the time-varying signal, wherein τ1 is invertible modulo N; and at least one processor communicatively coupled to the first ADC and the second ADC to sense the one or more frequency components of the time-varying signal by:
A) computing a first B1-point Fast Fourier Transform (FFT) for the first set of B1 samples of the time-varying signal so as to determine a magnitude value Xf1 for at least one frequency component of the one or more frequency components of the time-varying signal;B) computing a second B1-point FFT for the second set of B1 time-shifted samples of the time-varying signal; andC) comparing the first B1-point FFT and the second B1-point FFT so as to determine a position f1 in the bandwidth of interest BW of the at least one frequency component.