This preview shows page 1. Sign up to view the full content.
Unformatted text preview: h the autocorrelation
properties of Barker codes. EE102A:Signal Processing and Linear Systems I; Win 09-10, Pauly 16 Pseudo-Noise Codes
Pseudo-noise codes approximate the ideal sharp auto-correlation, orthogonal
codes. These are deterministic codes that approximate the characteristics
of a noise sequence.
They are ±1 with probability of 1/2, and have the run lengths of −1’s and
1’s of a random sequence. = At zero shift, the product of the two is a contant ”1”, and the autocorrelation
is ”N”. At other shifts, the product of each interval is just is likely to be
+1 as −1, so the autocorrelation is small. EE102A:Signal Processing and Linear Systems I; Win 09-10, Pauly 17 They are not completely orthogonal, although
T T φn(t)φk (t)dt 0 should be small (again, ±1 equally likely, so the integral will be small).
Other channels appear as noise-like interference.
If we correlate the received signal with φn(t) we get a large signal for the
nth channel, with a peak at the delay for the nth channel.
Adding users looks likes like an increased background noise level, which
softly degrades performance. EE102A:Signal Processing and Linear Systems I; Win 09-10, Pauly 18 y(t ) (!1 y)(t ) (!2 y)(t ) EE102A:Signal Processing and Linear Systems I; Win 09-10, Pauly 19 Summary
Orthogonal codes allow users to share a channel.
Hadamard codes work well when the channels can be synchronized, such as
when the basestation is talking to multiple handsets.
For unsynchronized channels, we want both orthogonality, and good
autocorrelations. Psuedo-random codes are a good approximation.
We can continue to add users by handing out more codes, with a soft
degradation of performance. Other users look like an increased noise level. EE102A:Signal Processing and Linear Systems I; Win 09-10, Pauly 20...
View Full Document
- Fall '13
- Signal Processing