A new multiuser receiver for synchronous code-division multiple-access (CDMA) systems with error control coding is proposed. The receiver achieves interference cancellation by projecting the undesired users onto the space spanned by the desired users' signal vectors. The detector calculates the least squares (LS) estimate of the interfering users data, that is used to yield an adjusted metric for maximum likelihood sequence estimation (MLSE) for the desired users' sequences. Simulation results indicate that close to optimal performance can be achieved when all but one of the users are projected using only a single user decoder for the desired user. Further, an adaptive receiver structure based on the recursive LS update is presented that is well-suited for DSP implementation due to it's computational efficiency.

It is expected that the available spectrum for the proposed personal communication networks will soon be at a premium as the user population increases, and using multilevel quadrature amplitude modulation (MQAM), an bandwidth efficient transmission method for digital signals, may significantly ease the problem. A new lattice polynomial perceptron (LPP), with faster and less input signal-dependent convergence behavior, is presented and applied to frequency-selective slow fading channel equalization and adjacent-channel interference (ACI) suppression in a 16-QAM system. Computer simulation results are given, which shows that, in a 16-QAM system, the performance of LPP is clearly superior to that of the other structures.