I am working on a signal processing assignment where I need to find out one common time domain signal from 4 observation. The math is like this:

\begin{array}{lcl} y_1(t) & = & a_1(t)*x(t) + b_1(t) \\ y_2(t)& = & a_1(t)*x(t) + b_1(t) \\ . & & \\ . & & \\ y_n(t) & = & a_n(t)*x(t) + b_n(t)\end{array}


$n$ is up to 4 in my case, and $x(t)$ is my signal of interest which I want to extract from observed signals $y_n(t)$.

What is the general direction I should take in doing background study for this kind of application?

Thanks, K

  • $\begingroup$ I would say the autocorrelation could help you $\endgroup$
    – Engine
    Apr 11 '14 at 11:18
  • $\begingroup$ Independant component analysis or Principal component analysis may help you $\endgroup$
    – tobassist
    Apr 11 '14 at 11:56

Because you have a set of linear equations, why don't you set up a linear system (for each data point) and solve the system (using SVD for example) to obtain the solution?

Write in the the matrix form such as $Y=\beta X$ where $\beta$ are known and $X$ is unknown.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.