Identification Algorithm - Sinusoidal Fitting -
Also this project is a little piece of my MS thesys.
In this module the problem rised is the identification of a sinusoidal signal sampled by sampler or a DSO (my case)
The desired objects are amplitude, mean value, frequency and phase of sampled signal.
This algorithm is complex because the are several problems in the frequency and phase detection. The use of FFT is usefull until the frequency is not much high then a pitch-detection algorithm is developed.
Below a flow-diagram of the module of "Sinusoidal Fitting"
How you can see above, the module starts computing ICs (initial conditions).
The module is formed by a nested loop (Least Square + Newton-Raphson 4x4). In this loop the function calculate the error functions, its gradient and the Jacobian of this gradient (since gradient of error functions are zero). Therefore the delta step required for error calcolus. At this point the termination criteria establish if go on with another step or stop the loop.
Thank you for reading