An improved averaged stochastic gradient algorithms for blind multiuser detection in DS/CDMA systems

An improved blind adaptive gradient algorithm (IBAG) for code-aided simultaneous suppression of multiple-access interference (MAI) and narrow-band interference (NBI) in a direct-sequence code-division multiple-access (DS/CDMA) system is proposed. The blind adaptive gradient algorithm (BAG) has two types: tracking blind adaptive gradient algorithm (TBAG), which has the tracking ability, and time-invariant blind adaptive gradient algorithm (TIBAG), which is applied in time-invariant circumstances. We prove that the IBAG algorithm has similar tracking properties to, and the same order of computational cost as, the BAG algorithm, and its output signal-to-interference ratio (SIR) rises more quickly than that of the BAG algorithm at the beginning