Hello Lowfwer's
Suppression of noise only works well when its behavior is known
at the place where the suppression acts. A suppression of pulse
type noise by a DSP will work even in the very narrow band case
when the impulse response of the system including the antenna is
known. The DSP would correlate the incoming signal with some
typical noise patterns and subtract this noise correctly in
amplitude and phase without cutting off the signal at that time.
The typical noise patterns could be identified by the DSP itself.
The problem of the narrow band case is that the characteristic
steep edges of pulses are filtered out. So the differentiation
between gaussian noise, pulses, and signal is difficult.
A parallel running broadband receiver would help to locate the
pulses.
Designing such a noise blanker specially for LF-type noise seems
to me a very useful and interesting task. Unfortunately I have
null time to do such things on my DSP's before september.
Is someone interested?
My course on Digital Signal Processing ist available from:
http://tech-www.informatik.uni-hamburg.de/Lehre/dsp99/
It is entirely based on MATLAB (Student's Edition)
(and Motorola DSP56002EVM).
The file dsp.zip contains the whole material.
Chapters 13...16 are still in progress.
73 de Klaus, DJ5HG
|
