I am analysing a periodic signal, which contains a periodic component, a trend component and noise component. I am only interested in the periodic component, and in particular, I am only interested in determining the period (1/frequency) of this signal. I tried FFT and autocorrelation, but did not succeed. What is the best way to solve this problem?
Thanks!
1 L A T E S T R E P L I E S (Newest First)
cpyang
Posted - 12/03/2017 : 9:54:19 PM Can you give some details on how FFT didn't work for you? Maybe you can show some images?
Maybe you should subtract the trend component with a linear fit?