Optimization of iterative DFE-IDMA detection for multipath fading channels

In this paper, a new receiver structure is proposed for uplink interleaved division multiple access (IDMA) on multipath fading channels. In a conventional IDMA rake detector, jointly removing inter symbol interference (ISI) and multiple access interference (MAI) is achieved by exchanging extrinsic log-likelihood ratio (LLR) chips between an elementary signal estimator (ESE) and a posteriori probability decoders (DEC)s in a turbo-like manner. The same exchanging principle takes place between a centralized decision feedback equalizer (CDFE) and the DECs in CDFE-IDMA systems. However, the proposed IDMA receiver can be considered a mixture between the two mentioned receivers as it employs an adaptive DFE for each user to remove the ISI effects while MAI is eliminated by a parallel interference canceller (PIC). The PIC utilizes the same ESE principles in removing MAI effects from received symbols which is more efficient than cross-over filters in a CDFE-IDMA system. Moreover, the ISI abstraction in DFE is more active in removing ISI than ESE. Simulation results demonstrate the superiority of DFE-IDMA over the other systems' performances found in the literature for both optimal and adaptive algorithms (AA).