Low-Complexity Equalization of Orthogonal Signal-Division Multiplexing in Doubly-Selective Channels
Jing Han,Lingling Zhang,Qunfei Zhang,Geert Leus
DOI: https://doi.org/10.1109/tsp.2018.2887191
IF: 4.875
2018-01-01
IEEE Transactions on Signal Processing
Abstract:Orthogonal signal-division multiplexing (OSDM) is a generalized modulation scheme to bridge the gap between orthogonal frequency-division multiplexing (OFDM) and single-carrier frequency-domain equalization. It allows significantly more flexibility in system design; however, over doubly-selective channels, it suffers from a special signal distortion structure, namely inter-vector interference, which is analogous to inter-carrier interference in conventional OFDM. To analyze its effect, in this paper, the complex exponential basis expansion model (CE-BEM) is used to approximate the doubly-selective channel. We show that the composite channel matrix of OSDM systems is cyclically block banded in this case, and the blocks in its main band can be further diagonalized. By exploiting this unique matrix structure, low-complexity block and serial OSDM equalization algorithms are then proposed. These two equalization algorithms are based on block LDL$^H$ factorization and block iterative matrix inversion, respectively, both of which are implemented in a transformed domain to avoid direct inversion of large matrices. In addition, a CE-BEM channel estimation method is designed for OSDM systems, which uses frequency-shifted Chu sequences as pilots to ease the computation. Numerical simulations are finally provided to justify the validity of our channel equalization and estimation algorithms.