Multi-dimensional matched filter identification technique for channel equalization deployed in spatial diversity receivers

Abstract

This paper proposes a multi-dimensional matched filtering technique for spatial diversity receivers. The coefficients of the multi-dimensional matched filter are identified by making use of an adaptive filter, the update of which doesn't require the transmission of any training symbols within the transmitted data stream. Therefore the use of the proposed technique improves the data rate efficiency. Furthermore, it is well known that implementing multi-dimensional matched filtering is essential for equalization purposes to obtain the optimum error rate performance from spatial diversity receivers. For that reason the technique is designed not only to identify the unknown matched filter but also to simultaneously lead to the equalization of the channel too. In order to update the adaptive filter, the Constant Modulus Algorithm (CMA) is utilized, which is an implementation convenient algorithm. Therefore the proposed technique is not computationally complex in comparison to those identification algorithms proposed for spatial diversity receivers. Simulations are provided to present the equalization performance of the novel technique.