The discrete-time signal $x_n$ is a sum of $K$ discrete-time complex damped exponential: \begin{equation} x_n = \sum_{k=1}^{K} \underbrace{(a_k e^{j\phi_k})}_{c_k} {\underbrace{e^{\{(- \alpha_k + j2\pi f_k )\Delta t\}}}_{z_k}}^n, \quad n = 0,1,...,N-1 \end{equation}

where $a_k$ - amplitudes, $\phi_k$ - phases, $\alpha_k$ - damping factors and $f_k$ - frequencies are the parameters of the model and $n$ is the time index and $\Delta t$ is the time interval.

\begin{equation} \label{sig_modf} x_n = \sum_{k=1}^{K} c_k z_k^n, \quad n = 0,1,...,N-1 \end{equation}

where $c_k$ is the complex amplitude including the phase of the $k$-th damped exponential and $z_k$ is the $k$-th signal pole.

Such a signal will satisfy the below recurrence relation: $x_n = p_1 x_{n-1} + p_2 x_{n-2} + ... + p_K x_{n-K} \tag 1$

For this system, I would like to know what is the state-space and Transfer-function model. I know that the companion matrix given by the coefficients $p$'s in the first row and 1's in the sub-diagonal is the state matrix $A$.

$A = \begin{bmatrix} -p_1 & -p_2 & -p_3 & \dots & -p_{K-1} & -p_K \\ 1 & 0 & 0 & \dots & 0 & 0\\ 0 & 1 & 0 & \dots & 0 & 0\\ 0 & 0 & 1 & \dots & 0 & 0\\ \vdots & \vdots & \vdots & \ddots & \vdots & \vdots\\ 0 & 0 & 0 & \dots & 1 & 0 \\ \end{bmatrix} \tag 2$

Is this correct and are my B, C and D matrices 0? Also, what is the TF model for this system?

Given the signal sequence $x_n$, is it possible to find the state-space and transfer-function model using MATLAB lsim?

  • $\begingroup$ Hi: googe for "AR(p) model in state space form" and a lot of things will come up. I didn't know which one to send or which one was most useful so I'll leave that to you. $\endgroup$
    – mark leeds
    Sep 27, 2021 at 14:59
  • $\begingroup$ "are my B, C and D matrices 0?" I don't know -- does your system have any inputs you didn't mention? Does it have any outputs you didn't mention? A transfer function for a system without either input or output is boring; it is just $H(z) = \emptyset$. $\endgroup$
    – TimWescott
    Sep 27, 2021 at 15:54
  • $\begingroup$ Thank you for your comments. I have a signal sequence that is computed from a mathematical model - sum of damped complex exponentials. Thus, they satisfy the recurrence relation given in (1). This is the only information I have. I was also thinking the tf = 0, but I wanted to make sure, whether it is feasible. $\endgroup$
    – Neuling
    Sep 27, 2021 at 16:45
  • $\begingroup$ Presumably you have the initial state of the system, or some measurements of the states, and you're interested in some output or the states -- if so, edit your question to say so, and detail what you have and what you're interested in. $\endgroup$
    – TimWescott
    Sep 27, 2021 at 17:40
  • $\begingroup$ I provided the background. I do not have information, other than the signal sequence $x_n$ and the mathematical model to generate it. I am interested in knowing the state space model of this system. My ultimate goal is to estimate the parameters $f,a,\alpha,\phi$ by determining the eigenvalues, which are the signal poles. $\endgroup$
    – Neuling
    Sep 27, 2021 at 18:07

1 Answer 1


Since you don't have any inputs, $x_n$ is only a function of past values of itself, your B an D should indeed be zero. You don't have your output defined either. You could pick the most recent $x_n$ as output, but it could also be something else.

A transfer functions describe an input output relation. However, you don't have inputs, so also no input output relation thus no transfer function.

  • $\begingroup$ Could you also comment on the C matrix? is it also zero? $\endgroup$
    – Neuling
    Sep 27, 2021 at 16:47
  • $\begingroup$ @Neuling look at my comments regarding the output of the system. Using your current description there is also no definition of what the C matrix should be. $\endgroup$
    – fibonatic
    Sep 27, 2021 at 17:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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