3
$\begingroup$

Problem

Given the compound system below, with the input $x(t)=\operatorname{sinc(t)}$, the output of A is $y(t)=\operatorname{sinc(2t)}$ and the output of B is $z(t)=\operatorname{sinc(t)}$, determine which of the A, B systems are Linear time invariant (LTI).
enter image description here

  • a) A
  • b) B
  • c) Both
  • d) Neither

My approach:

Using two facts below (if I got them right):

  • A and B are inverse of each other (because the overall output is equal to the input)

  • If an LTI system is reversible, then the reverse is also LTI.

after a quick scanning I crossed a, b as the answered. I concluded that either both systems are LTI or neither of them are.

For the next step, using the given information I derived the system A:

\begin{align} x(t)&=\frac {\sin(\pi t)}{\pi t}\\ y(t)&=\frac {\sin(2\pi t)}{\pi t}=\frac{2\sin(\pi t)\cos(\pi t)}{\pi t}=\frac{2\sin(\pi t)\sin(\pi t+\frac{\pi}{2})}{\pi t}=2\pi t x(t)x\left(t+\frac{\pi}{2}\right) \end{align}

now with the $t$ as the coefficient of the system we can say $y$ is not time invariant and therfore not LTI. so I chose d as the correct answer.
But the text book I'm reading proves that b is the answer, using another method.

Questions:

So my questions are:

  • Is my approach/facts flawed?
  • If so, what is the correct approach to solve this problem.
  • OR possibly did the book get it wrong?

Any help would be highly apprecieted!

$\endgroup$
2
  • 1
    $\begingroup$ What's quite flawed about your approach is the idea that a single such simple pair of input and output signals would give you conclusive information about whether a system is LTI. Really, you can never prove something is LTI by simply looking at a given input and output, only disprove that a system is LTI. (Though in practice, if you compare the power spectrums for many different signals and notice that the output for each frequence is always same same frequency-wise multiple of the same frequency in the input, this is a good indication that the system may well be LTI). $\endgroup$ Mar 20, 2017 at 15:57
  • $\begingroup$ @Amen could you say the name of the book you are learning from? $\endgroup$ Mar 22, 2017 at 13:44

1 Answer 1

4
$\begingroup$

This is how I would approach the problem.

System $A$ cannot be LTI because its output has frequencies not present in its input. The input spectrum is a brickwall from $f=-1/2$ to $f=1/2$, but the output spectrum is a brickwall from $f=-1$ to $f=1$.

It's impossible to arrive at a conclusion for system B. It could be an ideal LTI low-pass filter, but it could also be something else (for example, it could be a signal generator).

Note that we cannot say that A and B are inverses of each other. That statement requires $x(t)=B[A[x(t)]]$ to be true for all $x(t)$. Here, we only know that is true for one specific input.

$\endgroup$
4
  • $\begingroup$ Can we use my two facts to conclude that system B is not LTI as well? $\endgroup$
    – Amen
    Mar 19, 2017 at 16:03
  • $\begingroup$ by the way the book exactly used your approach for the system A and reversed this logic to state B is LTI. $\endgroup$
    – Amen
    Mar 19, 2017 at 16:04
  • $\begingroup$ I don't think so -- B could be LTI. However, given the information we have, there's no way to tell. $\endgroup$
    – MBaz
    Mar 19, 2017 at 16:04
  • $\begingroup$ @Amen, please see my edit. We can't be sure A and B are inverses. $\endgroup$
    – MBaz
    Mar 19, 2017 at 16:08

Your Answer

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

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