home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!emory!ogicse!news.u.washington.edu!stein.u.washington.edu!hlab
- From: thinman@netcom.com (Technically Sweet)
- Newsgroups: sci.virtual-worlds
- Subject: Re: TECH: Filtering head position/orientation data
- Message-ID: <1ib8l9INN3u@shelley.u.washington.edu>
- Date: 3 Jan 93 06:43:33 GMT
- Article-I.D.: shelley.1ib8l9INN3u
- Organization: University of Washington
- Lines: 27
- Approved: cyberoid@milton.u.washington.edu
- NNTP-Posting-Host: stein.u.washington.edu
- Originator: hlab@stein.u.washington.edu
-
-
-
- Diego Montefusco wanted to do dead reckoning:
-
- Everett Carter at the Naval PostGrad School in Monterey, CA, USA,
- posted his Kalman filtering C++ class to comp.dsp awhile back.
-
- Just spouting off:
-
- I always wondered if it would be possible to allow palsied users to
- use a mouse by running the mouse input through a bandpass filter.
-
- The assumption that processing a position input as a signal is
- somewhat dubious. Basic signal processing is biased towards the model
- of sound as a periodic waveform. If your sampling rate is fast enough
- (50hz maybe?) it is a reasonable use of DSP technology, but at
- 10hz-15hz for Polhemi or Power Gloves it makes more sense to adopt an
- underlying model of a person moving around jerkily and predict from
- that. With all-over joint sensors you could predict from a model
- incorporating the rotational abilities of the connected joints...
-
-
- --
-
- Lance Norskog
- thinman@netcom.com
- Data is not information is not knowledge is not wisdom.
-