read.me

Adaptive digital Filters in C++

		Misc Notes for the Adaptive digital Filters in C++


		The Postscript documentation on the digital filters is not
		up to date, refer to the source code for the most accurate
		information.  This package includes only the ADAPTIVE filters
		described in that manual (LMS and LSL filters).


		The data file 'pogo.dat' is a sample of acousting signals from
		a sinking oceanographic instrument.  A plot of the data will
		reveal 3 or 4 pulses arriving at the hydrophone.

		The data file 'drift.dat' is the same data, with a synthetic
		time varying bias added to the data.  Here a simple linear
		ramp from times 400 to 800 was used.  The LMS filter can be
		applied to this signal to adaptively adjust for the varying
		bias level of the signal, and remove it.

		lmstest:

			lmstest -evb1.0e-7 drift.dat out.dat

		will apply the LMS filter to drift.dat and send the error output
		result to out.dat.  If you compare out.dat with drift.dat
		you will see the ramp got removed from the data, so a simple
		threshold could be used to determine the time of the signal
		pulses.


			lmstest --		will give a list of options
						for the program


		Two notes on future updates:

		I have not yet written a test case for the LSL filter.


		I am currently working on other adaptive filters, beginning
		with the square-root Recursive Least-Squares filter.

	
		both of these will be included in future updates of this package
		when they are done.



		The full reference for Cowan & Grant:

		Cowan, C.F.N., and P.M. Grant, (eds) 1985; Adaptive Filters,
		Prentice-Hall, Englewood Cliffs N.J., 308 pages



 Everett (Skip) Carter          Phone: [phone] FAX:  [phone]
 Taygeta Scientific Inc.        INTERNET: [e-mail]
 100 Dolores St, Suite 186      UUCP:     ...!uunet!taygeta!skip
 Carmel, CA. 93923              WWW: http://www.taygeta.com/skip.html