Reduced-rank MMSE multiuser receiver for synchronous CDMA

The conventional matched filter receiver does not perform well for CDMA because of multiple-access interference (MAI) which can create the near-far problem. Many multiuser detection schemes have been developed to the mitigate near-far effect and suppress MAI. However, these receivers usually are too complicated to implement in real-time. A new adaptive algorithm is presented to efficiently implement the minimum mean-square error (MMSE) multiuser receiver for direct-sequence code-division multiple-access (DS-CDMA). This new algorithm generalizes the Wiener filter structure called the multistage Wiener filter. With the multistage representation of the MMSE receiver, a large matrix inversion can be avoided. In addition, the capability of this new algorithm to achieve rank reduction provides the flexibility needed to compromise between the computational complexity and performance. The adaptation formulas of the proposed algorithm are also developed in the paper. Finally, the performance comparison of the proposed algorithm to other existing algorithms is obtained by simulations