Clock recovery in symbol-rate DSP
In one aspect, the disclosure relates to a method of clock recovery. The method includes sampling one or more signals at a symbol rate to generate a set of samples; zero-padding one or more samples in the set of samples; filtering one or more zero-padded signals with a filter to generate a set of filtered samples; and modifying the set of filtered samples by applying a non-linear operator to generate an output comprising a spectral correlation.
1. A method of clock recovery comprising:
sampling one or more signals at a symbol rate to generate a set of samples;
zero-padding one or more samples in the set of samples;
filtering the one or more zero-padded samples with a filter to generate a set of filtered samples; and
modifying the set of filtered samples by applying a non-linear operator to generate an output comprising a spectral correlation.
2. The method of claim 1 , wherein the spectral correlation does not change for the one or more signals at the symbol rate when a signal-to-noise ratio varies based on increasing and decreasing noise levels.
3. The method of claim 1 , wherein a spectral correlation of frequency components spaced at the symbol rate is absent from the one or more signals sampled at the symbol rate.
4. The method of claim 1 , wherein the filter is a low pass filter.
5. The method of claim 1 further comprising recovering a clock of at least one of the one or more signals from the set of samples by computing a spectral line.
6. The method of claim 5 , wherein the zero-padding and filtering steps are performed such that one sample per symbol is recovered.
7. The method of claim 6 , further comprising estimating one or more-a clock phase errors over a clock period with regard to the recovered clock.
8. The method of claim 7 , wherein the clock phase errors is ∈, wherein
ϵ
=
1
2
π
arg
(
〈
∑
k
=
0
N
-
1
X
k
″
X
k
-
N
/
2
″
*
〉
)
.
9. The method of claim 5 wherein the clock may be recovered when sampling is performed at one sample per symbol.
10. The method of claim 1 , wherein the spectral correlation is X k X k-N/2 * , wherein N is a DFT size, wherein X k =DFT(x n ), wherein x n is N-length sampled data.
11. The method of claim 1 , wherein the non-linear operator is (x n ′) 2 , wherein x n ′ is the set of filtered samples.
12. The method of claim 1 , wherein the non-linear operator is an exponential operator.
13. The method of claim 1 , wherein the non-linear operator is configured to modify the set of filtered samples such that spectral correlation does not vanish when signals are sampled at the symbol rate.
14. The method of claim 1 wherein the sampling comprises sampling a continuous-time signal x a ′.
15. The method of claim 1 , wherein applying the non-linear operator does not cause aliasing.