The following document contains transcribed notes from Professor Gershenfeld's
class lecture on 12/1/03
Shannon's Equation:
C=Df log (1+s/n)
where C is the channel capacity, f is the frequency, s is the energy or strength
of the signal, and n is the noise or the energy put out by the item/medium the
signal is being transmitted through.
The error rate while sending data through a channel can be zero (up to a certain
point) with error correction. Even though a process is statistical does not
mean that it is unreliable.
Complexity in Communications -
- Modulation - making your signal look different from anything else
- AM (amplitude modulation), FM (frequency modulation), etc.
- AM is easy, but natural noise is AM
- AM, FM are also not too good because the modulation is one dimensional
(are only modulating one thing)
- Ideally, you want to increase the dimensionality of your signal modulation
- Codes generated are similar to the packing of spheres where the 'volume'
of a 'sphere' is the the volume in which a signal can be received without
purturbation/noise/etc. (I'm not too clear on this point)
Interdevice networking
- Internet zero - internet for everyday controllers
- 802.11b or ethernet complexity comes from the bit that is being transfered
being smaller than the conduit it is transfered through. This leads to
relativistic effects? This why impedence matching is very important (otherwise
reflections will occur upon the transfer of the information from one medium
to another which is not good.)
- When communicating in something the size of a building, 1Mb/s data transfer
rates are possible with bigger bits.
- This concept leads to some interesting devices/communication protocols
that may be used in class. More to come on this later.