Low-complexity implementation of CDMA downlink equalization

Beginning with the equalizer algorithm of Ghauri and Slock (1998) we investigate several reduced complexity implementations. We assume that the receiver performs channel estimation (based on the continuously available pilot), and this estimate is used to compute the required equalizer. This computation requires calculation of a matrix inverse, and this is the focus of the complexity reduction. We propose three algorithms: 1) taking advantage of the block Toeplitz structure, we apply a generalized-Levinson algorithm. 2) the equalizer is approximated by M parallel equalizers, utilizing the polyphase structure. 3) the equalizer solution is approximated iteratively via the Gauss-Seidel technique. We compare the performance of these approaches, and estimate complexity in terms of the number of complex multiplications.