1
$\begingroup$

I am trying to find a general way to find system stability.I have applied these methods and struck in confusion
For eg:
If
$$y(t)= \int_{-\infty}^{t} x(\tau) \sin(4\tau) d\tau$$
then find whether the system is stable or unstable?

Process (1):
If $$y(t) = \int_{-\infty}^{t} x(\tau)\sin(4\tau) d\tau$$
then $$h(t) = \int_{-\infty}^{t} \delta(\tau) \sin(4\tau) d\tau$$
$$\implies h(t) = \int_{-\infty}^{\infty} \delta(\tau)\sin(4\tau) u(t-\tau) d\tau$$
$$\implies h(t) = [\sin(4\tau)u(t-\tau)]|_{\tau=0}$$ $$\implies h(t)=0$$.

so system is stable as $$\int_{-\infty}^{\infty} |h(t)| dt= 0 < \infty$$ ;

Process (2):
If $$y(t)= \int_{-\infty}^{t} x(\tau)\sin(4\tau) d\tau$$
then when $x(t)=\sin(4t)$
$$y(t)= \int_{-\infty}^{t} \sin(4\tau)\sin(4\tau) d\tau = \infty $$ so the system produces unbounded output $y(t)$ for bounded input $x(t)= \sin(4t)$
Hence the system is unstable

Thus actually what will be the system stability and which process is right and which is wrong and how it is wrong?? please explain....

$\endgroup$
  • $\begingroup$ This is totally unreadable, please use Latex formatting for the formulas. $\endgroup$ – Matt L. Feb 7 '18 at 13:32
  • $\begingroup$ i believe that you are trying to convey convolution, and if you are, it is not correct. $\endgroup$ – Stanley Pawlukiewicz Feb 7 '18 at 16:44
  • $\begingroup$ i have tried my best to edit this post so that it is clear to you ;please now remove the closed from my post @Peter K $\endgroup$ – Suresh Feb 15 '18 at 13:57
  • 1
    $\begingroup$ @user9198116 Re-opened, but please learn more MathJax! :-) $\endgroup$ – Peter K. Feb 15 '18 at 15:03
1
$\begingroup$

Your first method assumes that the system is entirely characterized by its response to the input $x(t)=\delta(t)$. This would only be the case if the system were linear and time-invariant (LTI). In that case, the input-output relation would have the form

$$y(t)=\int_{-\infty}^{\infty}x(\tau)h(t-\tau)d\tau\tag{1}$$

where $h(t)$ is the system's impulse response. However, the given input-output relation is not of the form $(1)$, and hence the system is not LTI. Consequently, the system's response to $x(t)=\delta(t)$ does not say much about the system's behavior in general. That's why your first method of determining the stability is not valid.

The second method is valid. You found a bounded input signal for which the output is unbounded, and in this way you've shown that the system is not BIBO-stable.

$\endgroup$

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.