Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I just finished reading the fundamentals chapter about signals (linearity,causality,memory and time invariance). I wanted to solve some exercises and I found this one.

We have a signal with output described by $$y(t) = \int_t^{t+1}x(\tau-a)\;d\tau ,a\in R$$

  1. find if the system has memory, is stable, time invariant and linear
  2. which values of $a$ make the system causal

I am trying to understand things here. The system has memory if $a<0$ and memoryless $a>0$? I don't know from where to start to find the others.


To prove linearity I did this

$$y(t)=F\{x(t)\}$$ $$F\{c_1x_1(t)+c_2x_2(t)\}$$ $$\int_t^{t+1}c_1x_1(\tau -a)+c_2x_2(\tau -a)d\tau $$ $$=\int_t^{t+1}c_1x_1(\tau -a)d\tau +\int_t^{t+1}c_2x_2(\tau -a)d\tau$$ $$=c_1y_1(t) + c_2y_2(t)$$

time invariance

$$F\{x(t-t_0)\}$$ $$=\int_t^{t+1}x(\tau -a-t_0)d\tau $$ $$=\int_{t-t_0}^{t+1-t_0}x(u -a)du $$ so $$=\int_{t-t_0}^{t+1-t_0}x(\tau -a)d\tau $$

which means it is time invariant

Am I right?

share|cite|improve this question
Have you been already told about Fourier transforms and frequency response of the system? – Ilya Mar 22 '12 at 11:00
I know about frequency response but not Fourier. I am not there yet. – john_paker Mar 22 '12 at 11:03
Maybe you should ask the moderators to move this question to the signal processing site dsp.SE. – Dilip Sarwate Mar 22 '12 at 11:24
@DilipSarwate It is not about dsp. – john_paker Mar 22 '12 at 11:28
Did I say that your question was about dsp (digital signal processing)? I recommended moving the question to the signal processing site dsp.SE where discussions deal with signal processing in general, not just dsp. – Dilip Sarwate Mar 22 '12 at 11:36
up vote 0 down vote accepted

First, note that the output signal at time $t$ is formed by integrating the values of the input over a unit length interval: $[t-a,t+1-a]$ (a continuous average).

Then, it should be evident that the system is linear (integration is, like the sum, a linear operator), and time-invariant (the relation of input-output does not change over time). And it's certainly not memoryless (the output at time $t$ does depend on other values besides the input at time $t$). To be causal, the output at time $t$ should depend only on "past" values of the input. For example, if $a=0$ or $a=0.5$, it 's NOT casual.

All this is rather informal, if you need to proof these facts formally, this is a start, bu t you should work this out in detail.

If you have learned about the input response $h(t)$, you can also find out that the system is indeed described by a $h(t)$, which is non-zero (actually, 1) over the interval $[a-1,a]$. Once you see this, (and you are supposed to see this quickly when you have some training in signals) you can answer immediately all the original questions.

share|cite|improve this answer
I edited my question. I think I proved linearity. I am struggling to prove the others. – john_paker Mar 22 '12 at 12:06
About causality. Does this means that it is causal only for a = 1? – john_paker Mar 22 '12 at 12:43
@john_paker: rather, it's casual for $a \ge 1$ – leonbloy Mar 22 '12 at 13:41
Thanks for your help. I now understand what is going on. – john_paker Mar 22 '12 at 18:58

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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