Soft-Decision Aided Probabilistic Data Association Based Detection for Mode Division Multiplexing Transmission With Mode-Dependent Loss

Few-mode fiber (FMF) based mode division multiplexing (MDM) transmission system has been considered as a potential candidate for future backbone network, unfortunately, the mode-dependent loss (MDL) in MDM system remains to be a tough issue in high speed transmission application. Recently, the optimal maximum-likelihood (ML) detection has demonstrated its potential ability and efficiency in suppressing the capacity-limiting MDL. However, this ML detector employed at the receiver of MDM system achieves the optimal performance at the sacrifice of the exponential growth in complexity. Inspired by the fact that probabilistic data association (PDA) is able to offer near-ML performance without introducing the enormous computational complexity, in this paper, a soft-decision aided PDA detector is proposed for MDM system detection, in which the probabilities of the potential candidate symbols are iteratively estimated by using the approximation of the Bayesian theorem. In addition, only those high reliability symbols are detected to save the computational effort. To validate the proposed scheme, both weakly and strongly coupled modes have been considered. Simulation results show that the proposed PDA based detector can achieve a flexible trade-off between the BER performance and the computational complexity when compared with the optimal ML detection in FMF-based MDM system.