Convolution Integrals

From Qwiki

Jump to: navigation, search

Convolution Integrals are used extensively for solving differential equations in terms of impulse responses. The convolution of two functions f1(t) and f2(t), denoted by * , is usually defined as

\left(f_1 * f_2\right)(\tau) = \int_{-\infty}^\infty f_1(t) f_2(\tau-t) dt.

Convolutions are commutatitve

 \left(f_1 * f_2\right)(\tau) =  \left(f_2 * f_1\right)(\tau)

and associative
 \left(f_1 * \left(f_2 * f_3\right)\right)(\tau) =  \left(\left(f_1 * f_2 \right)*f_3\right)(\tau).

An alternative, and sometimes very useful, way to write the convolution integral is explicitly symmetric

\left(f_1 * f_2\right)(\tau) = \int_{-\infty}^\infty dt_1 \int_{-\infty}^\infty dt_2 f_1(t_1) f_2(t_2) \delta\left(\tau-t_1-t_2\right) .

In this form, the multiconvolution is a straightforward generalization of the simple convolution

 (f_1*\cdots *f_n)(\tau) = \int dt_1\cdots \int dt_n f_1(t_1) \cdots f_n(t_n) \delta(\tau-t_1-\cdots - t_n) .
Personal tools