serhannn
Member level 4
Hello, I know that in a CDMA system every user should use mutually orthogonal codes so that data from different user can be separated in the receiver and we can create these codes, for example, using Walsh/Hadamard matrices or PN sequences.
My question is, if a user is transmitting e.g. 100 symbols and there are 10 transmitting users, should I use a different code also for each symbol of every user? In other words, should I have only 10 mutually orthogonal codes considering there are 10 users or should I have 100*10 = 1000 mutually orthogonal codes. The latter option seems to me a little illogical because it would probably be impractical to create so many orthogonal codes. But on the other hand, there can also be cross-correlation among different symbols from the same user, right? So if their codes are not also mutually orthogonal, would this not lead to an interference?
A picture for what I am trying to implement is attached. Basically, this is code matrix for the nth symbol of the ith user and cin is the code vector with spreading gain Gi. Li the number of multipath fingers; apparently here we have 3 fingers. So again, my question would be do we need to generate different cin's for respective symbols of the same user and if yes, how should we proceed with that?
Thanks in advance.
My question is, if a user is transmitting e.g. 100 symbols and there are 10 transmitting users, should I use a different code also for each symbol of every user? In other words, should I have only 10 mutually orthogonal codes considering there are 10 users or should I have 100*10 = 1000 mutually orthogonal codes. The latter option seems to me a little illogical because it would probably be impractical to create so many orthogonal codes. But on the other hand, there can also be cross-correlation among different symbols from the same user, right? So if their codes are not also mutually orthogonal, would this not lead to an interference?
A picture for what I am trying to implement is attached. Basically, this is code matrix for the nth symbol of the ith user and cin is the code vector with spreading gain Gi. Li the number of multipath fingers; apparently here we have 3 fingers. So again, my question would be do we need to generate different cin's for respective symbols of the same user and if yes, how should we proceed with that?
Thanks in advance.
