The probability distribution of the signal-to-interference plus noise ratio (SINR) is analyzed and approximated by the generalized gamma distribution for mode-dependent loss (MDL)-impaired fiber links. Based on the characteristics of SINR, an improved reduced-search algorithm (IRS) based on multilevel SINR thresholding (IRS-MST) is proposed. IRS-MST dynamically determines the size of the search space for each mode based on multiple SINR thresholds. Furthermore, a compounding function is used to balance the size of the search space and performance. Simulation results show that IRS-MST can achieve near ML performance for different MDLs. Compared to the conventional IRS algorithm, the complexity of IRS-MST is approximately 5.4 times smaller than that of the IRS. INDEX TERMS Mode division multiplexing, mode-dependent loss, MIMO detection, reduced search, SINR.
Mode division multiplexing (MDM) technology enables multiple data signals to be transmitted over the same fiber channel, which is a potential method for increasing the capacity of optical communication systems. Primary impaired factor mode-dependent loss (MDL) poses fundamental capacity limitations, increases the hardware complexity of the receiver, and decreases the performance of the MIMO algorithm. In this work, the alternating direction method of multipliers with the threshold (ADMM-T) detection algorithm is proposed for MDL-impaired fiber links by exploiting the SINR distribution of minimum mean square error (MMSE) and alternating direction method of multipliers (ADMM). Since the post-processing SINR distribution of MMSE has high reliability in the high SINR symbols, the MMSE detected symbols above the pre-defined SINR thresholds are retained as part of the final detection result and the remaining lower-SINR symbols are to be iteratively detected by ADMM after successive interference cancellation (SIC). Computational complexity can be reduced by reserving the high SINR symbols as the final solution by MMSE. Simulation results demonstrate that the proposed ADMM-T detection algorithm outperforms MMSE, ADMM, and some recently proposed detection algorithms. Meanwhile, the computational complexity of the proposed detection algorithm is lower than that of the other detection algorithms.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.