|
发表于 19-10-2006 01:48 PM
|
显示全部楼层
In DSP, there are 2 kinds of digital filters, FIR and IIR. To design these filters correctly, u need to know the "coefficients" of these filters. To calculate the value of coefficients, using MATLAB, u need to do DFT (for FIR) or z-domain (for IIR). z-domain can be reached by Laplace transform.
FIR filters does not have a feedback, meaning, it does not concern changing of values with time. Thus, using DFT to calculate coefficients in the frequency domain is the way to do it.
IIR filters has a feedback, the reason to have a feedback is to reduce the 'taps' of the filters. Lesser taps, smaller design. Since time is a concern (ie. feedback), DFT is not appropriate because calculation is done in frequency domain, not time-domain. Laplace transform provides a z-domain that can calculate coefficients based on value of the past provided by the feedback. Examples of IIR filters designs are Butterworth, Tschebyshev, elliptical, Bessel, adaptive IIR (Kalman) filters.
Laplace transform also has important application in sensing & control engineering like robots, missile trajectory applications because these systems needs feedbacks to operate.
[ 本帖最后由 katami 于 19-10-2006 01:50 PM 编辑 ] |
|