correlation with barker seqeunce in 802.11b

[bizoo]

802.11b barker

hi
i am simulation 802.11 WLAN and in this study i am focussing right now only on header which is 128 bit and goes through DBPSK modulation and then i uses DSSS by using 11 bit barker code

now i am testing this small system by correlating the output i sent on the channel with 11 bit barker code, for time being i have no noise and i also assume perfect synchronization so i know when the packet start, as a result when i corerlate 11 bits of data on the channel with 11 bits of barker code, i get my original data back

but the probelm is that if i want to see a string of the retrieved bits e.g. 22 bits(chips) of data from the channel correlated with 2 barker codes (11+11=22) should give me back my two origianl bits and show it on the graph (matlab),

but problem is that i often get more than two peaks (3 in case of 22 chips, 5 peaks in 33 chips, 13 in 99 chips)and so on and so forth

can some one explain this phenomenon, 11 bits(chits) from channel vs 11 bits of barker is ok, this correlation shows one peak which accounts for one origianl bit,. but the problem arises if tend to see more than one bit on graph using correlation

waiting for prompt reply

br

 

[mahdithdn]

Hi,

Barker codes are optimal in pulse radars, i mean discrete data transmission. so the correlation is aperiodic. maybe you should exploit another spread spectrum sequence like m-sequence.
in other hand if other peaks in correlation are local peaks so they are trivial.
if these are not your problem, attach your scripts.

best regards

 

[bizoo]

well thanks for replying
first is that i have to use only barker sequence because its there in the standard,


secondly, i could not get the part in which You say things about trivial peaks,
which ones are trivial and which ones not?

let me reiterate the fact that when i correlate 11 chips of incoming data from the channel with 11 bits of barker code ( one sequence), i get my original bit back, ie i will get a peak at 0 and its amplitude suggests that whether i transmitted 1 or zero

but whenever i try to correlate my signal with more than one seuqence of barker code (eg 22 bits of barker code=2 barker sequences ) i get more than two peaks, and the bits i get also are not the ones i trasnmitt

i have attached a figure which shows that i correlated 22 chips of my signal with 2 barker sequences . i should have got 2 peaks but i sometimes get 3 :S as in this case

 

[mahdithdn]

Please send your m-file of generating 2 sequences and calculating correlation it at the receiver.
the problem is dizzy here.