A blind adaptive receiver for interference suppression and multipath reception in long-code DS-CDMA

This paper examines a blind adaptive implementation of a recently proposed linear minimum mean square error (LMMSE) receiver for code-division multiple-access (CDMA) systems with aperiodic spreading sequences in multipath channels. The receiver has been previously shown to perform multiple-access interference (MAI) suppression and multipath diversity combining. The adaptive implementation is based on a fractionally spaced equalizer (FSE) whose taps are updated by the leaky constant modulus algorithm (LCMA) in the cold start and the decision-directed least-mean-square (DD-LMS) algorithm when the channel eye is opened. The LCMA adds a quadratic constraint, defined as the squared norm of the FSE weight vector, to the constant modulus (CM) cost function. Simulation results show that the LCMA with a uniform initialization strategy (all taps equally set to a small non-zero value) acquires all paths associated with the desired user, suppresses MAI, and opens the channel eye for the DD-LMS algorithm to converge to the proximity of the MMSE solution.