Hi,
I wonder if it is possible to fit with convoluted model function, for example, numerical Instrumental Response Function (IRF) and exponential decay.
I found related threads like
https://forums.ni.com/t5/LabVIEW/Nonlinear-curve-fitting-and-convolution/td-p/733833/page/2?profile
https://forums.ni.com/t5/LabVIEW/Curve-Fitting-with-Convolution/td-p/493836?profile
But, I still don't understand detailed procedure.
In principle, observed signal is descrete convolution between IRF and exponential decay in time-domain. And equation of descrete convolution is here,
I(t)=∑(L(t'-t shift)×F(t-tj)×Δtj), j=1,.....,t
I(t) is intensity of observed signal at the time t (measured), L(t'-t shift) is IRF function (measured), F(t-tj) is original signal that is having parameter I want.
In this case F(t) = A*exp(-t/τ), and I want to extract the fitting parameter A and tau.
However, nonlinear curve fit.vi (Levenberg-Marquardt algorithm) only requires model function (reference vi or equation), time t (X), signal I(t) (Y), initial parameters.
Because the measured IRF data remains unused, I connected IRF array to data (top of fit.vi). But, it seems to be not right way.
So, I can't find the solution to fit convoluted data.
Please advise the new user.
Thanks.
