Reduced Complexity MMSE Joint Detection in TDD CDMA Downlink

摘要

Joint detection techniques such as zero-forcing and minimum mean square error (MMSE) employ the Cholesky factorization (CF) to solve a complex system of linear equations. The number of computations in the CF block which is the most computationally intensive part of the detector can be significantly reduced by utilizing the sparseness and similarity characteristics of the covariance matrix combining the effects of the spreading sequences and the channel and the resulting CF matrices. In this paper we propose a technique that leads to further significant reduction in the number of computations in the CF block by deciding on the size of the computational task at real reception time, rather than having them predetermined at design time. Simulations of the technique in the TDD CDMA downlink show that this "reception time" decision is influenced by only a small number of determinable factors such as the span of ISI, number of users and the level of interference power.