慢变无线信道下的序列检测
Sequence estimation under the fiat fading wireless channel
王东昱 1武梦龙 2刘文楷 2杨大成3
作者信息
- 1. 北方工业大学,信息工程学院,北京,100041;北京邮电大学,无线通信中心,北京,100876
- 2. 北方工业大学,信息工程学院,北京,100041
- 3. 北京邮电大学,无线通信中心,北京,100876
- 折叠
摘要
本文对多径信道序列检测的VA(Viterbi Algorithm)算法进行了改进,建立了节点度量的递推方程,改进后的VA算法适用于多径信道先验信息已知的序列检测.其中信道先验信息由健壮型RLM(Recursive Least M-estimate)算法及Kalman滤波器在训练序列阶段联合估计.仿真结果显示:在训练阶段对信道进行联合估计时,对相比于传统的RLS算法,RLM算法有效地提高了抵抗大脉冲干扰的能力,加快了待估的参数的收敛速度.而在实际接收阶段,新的度量递推算法在比特误码性能上接近采用全序列比较欧氏距离的最佳算法,但是复杂度却很低.
Abstract
This paper modifies the Viterbi algorithm used in the multi-path channels, and gives the recursive metric equation. The modified VA algorithm can be used in the MLSE(Maximum Likelihood Sequence Estimate) receiver when the channel information is known a prior. The channel information can be estimated simultaneously by using the robust RLM (Recursive Least M-estimate) and Kaiman filters during the training period. Simulation results show that the coupled channel estimator is less sensitive to consecutive impulse noise and has better convergence ability than conventional algorithms during the training stage. In the real decoding stage, the modified VA is near to the optimal algorithm, however, its complexity is lower than the optimal algorithm.
关键词
VA算法/累积度量/分枝转移度量/AR模型/RLM算法/Kalman滤波器Key words
VA algorithm/cumulative metric/branch transition metric/AR model/RLM algorithm/Kalman filter引用本文复制引用
出版年
2008
NETL
NSTL
维普
万方数据