Dimension Reduction Method Based on Partial Matched Filters for Multi-User Detection in Direct Sequence Spread Spectrum Communication System

For multi-user detection in code division multiple access communication systems, the conventional dimension reduction approaches are subspace tracking methods. These methods may entail complicated computation and require iterative processes to ensure effectiveness. To circumvent the high-complexity issues in blind multi-user detection methods with large code lengths, this paper proposes a fast dimension reduction method based on partial matched filters. It reduces the signal dimension by several pre-calculated matrices, only utilizing the prior information of the spreading code of the target user. This method offers quick and flexible adjustment through various partial matched filter group designs, trading off between interference suppression performance and computational complexity.