提出了一种基于因子图-消息传递算法的联合信道估计单载波频域均衡(SC-FDE)接收机。针对现有联合信道估计接收机中存在的复杂度和性能之间的矛盾,利用因子图拉伸方法对现有模型进行拉伸变换。变换之后的因子图模型可以将现有模型中的观测节点分割为几个节点的组合,并依据分割后节点的自身特点采用更合适的消息传递规则,进而避免了各种消息传递规则的固有缺陷。仿真结果表明,相比已有联合接收机算法,所提出的基于因子图拉伸和联合期望传播-平均场(BP-MF)算法在提升性能的情况下能够显著降低复杂度。
Based on message passing algorithm,a joint channel estimation and decoding receiver for single carrier with frequency e-qualization(SC-FDE) system has been proposed in this paper.Solving the contradiction of complexity and performance in state-of-the-art receivers,this work exploits the stretch method to decompose the observation node into several simple factor nodes while keeping its original function.Then,different message passing algorithms are adopted to handle factor nodes according to their re-spective characteristic.Compare to the referred receivers,the proposed message-passing-based receiver not only improves the per-formance,but also reduces the complexity significantly.