Advanced CDMA for Wireless Communications

A New Algorithm for Blind Adaptive Multiuser Detection in Frequency Selective Multipath Fading Channel

“…The optimum value of increases with the number of interfering users or the power level of the received signals. Hence, we can derive the optimum value of as (23) This represents the comparison term on energy to assure energy constraint at each step of gradient algorithm, i.e., if , adaptive algorithm is carried on, otherwise the detector imposes a new equal to (24) The initial condition value strongly influences the convergence of the stochastic gradient algorithm (Appendix C). This algorithm guarantees a large convergence region, but not a good convergence speed.…”

“…Moreover, the conventional detector suffers from a substantial performance loss as the number of the interfering users increases or the signals are received with different power levels (near-far effect). 1 Multiuser detection (MUD) techniques [15], [23], [24] were proposed in order to cope with the near-far effect and nonorthogonal spreading code scenarios: the matched filter outputs and the parameters of the multiple users are used to jointly perform an optimum detection for each individual signal. MUD is able to mitigate the MAI which is the main limit for a CDMA system capacity as bandwidth is for frequency division multiple-access and time division multiple-access systems.…”

“…The optimum multiuser detector [14] consists in a bank of matched filters and a Viterbi decoder whose complexity is exponentially related with the number of users and usually results to be excessive in several practical applications. Hence, research has focused on suboptimal detectors which allows a lower implementation complexity [15], [23]. Recently, some attention has been devoted to adaptive blind MUD approaches which avoid the knowledge of the signature waveforms, the timing and the complex amplitudes of the interfering users.…”

Wideband CDMA network sensitivity function

“…The vector (8d) can be expressed as [39] ( 12) where (13) is a diagonal matrix of transmitted amplitudes C (14) is the matrix of channel coefficient vectors…”

“…For the lower limit we use (37) and the first term becomes (38) For a signal with -and -components, the parameter should be replaced by (39) where is the information in the interfering channel ( or ), and is the cross correlation between the codes used in the -and -channels. For small tracking errors, this term can be replaced by (40) where the notation is further simplified by dropping the subscript .…”