Problem in writing the Verilog HDLcode for EMD algorithm?

sir, I am trying to do following -

I am giving a signal x(t) as input which has 1000 samples.(i have taken one signal and samples it using the matlab and got the points in form of integars)

Now i have extracted the local maxima and minima of that signal.

By using local maxima i have calculated the upper envelope by using linear interpolation and same i have done with local minima to calculate the lower envelope and then i have taken the average of the upper and lower envelope and have got a signal m(t)

now i have to calculate a new signal h(t) which will be

h(t) = x(t) - m(t)

I have to repeat all above three times, means now again h(t) will be the input signal,
x(t) = h(t)

How to repeat the steps? I have written the code separately for each step. Should i use FSm? If i am going to use FSMs then how can i instantiate the modules in FSMs? Please suggest. I ma in big trouble

[accessed Nov 9, 2015].

That looks like a software flow diagram. Not much use in hardware land.
In the hardware, these 1000 samples will probably be streamed in, so you might as well process the data as it is streamed in. There will probably be some buffering, FIFOs etc.

I suggest you DRAW the circuit on paper BEFORE you write any HDL. HDL is hardware description language - if you dont know what the circuit should be, how can you write the code.