Successive interference cancellation using stage dependent decision functions

In this paper, we introduce the concept of stage-dependent decision functions. The latter exploit the improvement of the soft-decisions´ reliability with increasing number of stages to take more hard-decisions and thus cancels more noise from stage to stage. We prove that if the stage-dependent thresholds can be optimized at each stage then the nonlinear SIC can approach the single user bound. Simulation results show significant BER improvement compared to the SIC using stage-independent decision functions.