Beyond Nyquist: Efficient Sampling of Sparse Bandlimited Signals
read more
Citations
CoSaMP: Iterative signal recovery from incomplete and inaccurate samples
CoSaMP: iterative signal recovery from incomplete and inaccurate samples
From Theory to Practice: Sub-Nyquist Sampling of Sparse Wideband Analog Signals
Structured Compressed Sensing: From Theory to Applications
Signal Processing With Compressive Measurements
References
Compressed sensing
A wavelet tour of signal processing
Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information
Discrete-Time Signal Processing
Atomic Decomposition by Basis Pursuit
Related Papers (5)
Frequently Asked Questions (11)
Q2. What is the amplitude coefficient of the random demodulator?
In a hardware implementation of the random demodulator, it may be advisable to apply a prewhitening filter to preserve the magnitudes of the amplitude coefficients.
Q3. What is the easiest conclusion to draw from the bounds on SNR performance?
The easiest conclusion to draw from the bounds on SNR performance is that the random demodulator may allow us to acquire high-bandwidth signals that are not accessible with current technologies.
Q4. What was the earliest example of a stable reconstruction of a bandlimited signal?
In the 1960s, Landau demonstrated that stable reconstruction of a bandlimited signal demands a sampling rate no less than the Nyquist rate [4].
Q5. What are the factors that cause the degradations in the SNR?
These degradations are caused by factors such as the nonlinearity of the multiplier and jitter of the pseudorandom modulation signal.
Q6. How many FFTs does it take to recover a sparse signal?
Roughly speaking, it takes several hundred FFTs to recover a sparse signal with Nyquist rate from measurements made by the random demodulator.
Q7. What is the advantage of the random demodulator?
The random demodulator, on the other hand, benefits from the integrator, which effectively lowers the bandwidth of the input into the ADC.
Q8. How many harmonic tones are required to approximate a nonharmonic sinusoid?
If there are multiple nonharmonic sinusoids present, then the number of harmonic tones required to approximate the signal to a certain tolerance scales linearly with the number of nonharmonic sinusoids.
Q9. How does the random demodulator deal with the problem?
The random demodulator deals with the problem by smearing the tone across the entire spectrum so that it leaves a signature that can be detected by a low-rate sampler.
Q10. Why is it preferable to use pseudorandom numbers?
It is preferable to use pseudorandom numbers for several reasons: they are easier to generate; they are easier to store; and their structure can be exploited by digital algorithms.
Q11. How can the authors prove that the demodulator still offers the same guarantees?
The authors have been able to prove that if the chipping sequence consists of -wise independent random variables (for an appropriate value of ), then the demodulator still offers the same guarantees.