I know how to check a 1-D system for these conditions, but am confused about translating this to a 2-D system over a lattice. The system $H$ is define as:

$$w[\mathbf x] = H\left\{u[\mathbf x]\right\} = \sum_{\mathbf y \in \mathbf Q} \varphi[\mathbf y] u[\mathbf x - \mathbf y],\quad\text{with}\quad\begin{cases} &\mathbf x \in \Lambda \subset R^2,\\ &\mathbf Q = \left\{\mathbf q \in \Lambda : \lVert\mathbf q\rVert \le \rho\right\},\\ &\quad \rho \in R, \ 0 < \rho < \infty \end{cases}$$

I have mainly done problems with very simple 1-D systems. I know that the summation part represents convolution and that the check for linearity is:

$$H\left\{af_{1}[\mathbf x] + bf_{2}[\mathbf x]\right\} = aH\left\{f_{1}[\mathbf x]\right\} + bH\left\{f_{2}[\mathbf x]\right\}$$

and the check for shift invariance is:

$$w[\mathbf x - \mathbf y] = H\left\{f[\mathbf x-\mathbf y]\right\}$$

However I can't figure out how to actually do these checks on this particular system.


1 Answer 1


Well, actually you only need to insert the system $H$ into the definition of linearity and check, if it holds:

$$\begin{align}H\{af_1[x]+bf_2[x]\}&=\sum_y\phi[y](af_1[x-y]+bf_2[x-y])\\ &= a\sum_y\phi[y]f_1[x-y]+b\sum_y\phi[y]f_2[x-y]\\ &=aH\{f_1[x]\}+bH\{f_2[x]\}\end{align}$$

So, the linearity holds. There is nothing special about the fact that the system is 2-dimensional or similar. You just follow the rules for the calculation, and you end up with the statement of linearity.

A similar thing can be carried out for the shift-invariance.

  • $\begingroup$ Should the second sum have the first function be a function of y? $\endgroup$
    – Austin
    Commented Feb 11, 2017 at 19:48
  • 1
    $\begingroup$ yes, good spot! $\endgroup$ Commented Feb 11, 2017 at 19:53

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.