home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.models.rc
- Path: sparky!uunet!ftpbox!mothost!white!rtsg.mot.com!svoboda
- From: svoboda@rtsg.mot.com (David Svoboda)
- Subject: Re: Radio dreams
- Message-ID: <1992Sep12.041137.9338@rtsg.mot.com>
- Sender: news@rtsg.mot.com
- Nntp-Posting-Host: guppie44
- Organization: Motorola Inc., Cellular Infrastructure Group
- References: <1992Sep9.150729.4381@informix.com> <1992Sep9.190035.20060@usceast.cs.scarolina.edu>
- Date: Sat, 12 Sep 1992 04:11:37 GMT
- Lines: 46
-
- In article <1992Sep9.190035.20060@usceast.cs.scarolina.edu> hutto@smtc.engr.scarolina.edu writes:
- |
- | Actually, what you are asking for is best done using a set of techniques
- | collectively known as "spread spectrum". Each transmitter/receiver pair
- | uses a unique pseudo-random sequence of numbers to "hop" frequency every,
- | let's say, 10ms. As long as they stay in sync, the receiver "hears" the
- | transmitter pretty much continuously (except for a tiny interval where
- | the frequency is settling on its new value). However, as long as all
- | transmitter/receiver pairs us sequences that are sufficiently random
- | (actual they just have to be sufficiently independant of each other),
- | the worst thing likely to happen is a very occasional missed 10ms frame
- | when two of the pseudo-random sequences "cross" on the same frequency
- | for one frame.
-
- What you very nicely described is known as "frequency hopping" or "direct
- sequence" spread spectrum. That method is fairly old technology, having
- been used in some form or another since WWII. (Quiz: who knows what famous
- movie actress has the basic patents on frequency hopping?)
-
- A rather more elegant method of accomplishing spread spectrum is to take
- your binary signal sequence and xor it with a long binary key which has a
- very high data rate compared with your signal. The resultant signal is
- modulated, but the high data rate of the key causes the power to "splatter"
- about the spectrum. Then, the receiver can cross correlate the key with
- the incoming wideband signal, and when a high correlation is found, the
- signal can be recovered. As long as the key of another "interfering"
- transmission is orthogonal to that of your receiver (by design), it
- will not correlate, and will simply look like low-power noise. To a
- narrow band receiver, the total power of the spread-spectrum transmitter
- at a given frequency is down in the thermal noise. And the spread
- spectrum receiver doesn't see the narrowband transmissions because
- their energy is concentrated in a narrow frequency region, and therefore
- will not greatly affect the overall correlation.
-
- If you make the key sufficiently long, say, 10e10 symbols, then repeat,
- you can have a staggering number of mutually orthogonal codes, and
- therefore non-interfering radios. Synchronization between transmitter
- and receiver is critical (the longer you have to make your search
- window for correlation, the more processing power it takes), but that
- might be as simple as plugging your transmitter and receiver together
- while charging.
-
- This spread spectrum method is what is planned for cellular radio,
- though rather more sophisticated in execution.
-
- Dave Svoboda, Palatine, IL
-