Prediction of state transitions in Rayleigh fading channels
Prediction of state transitions in Rayleigh fading channels
复制标题
DOI：
10.1109/tvt.2007.891421
复制
发表时间：
2007-03-01
影响因子：
6.8
通讯作者：
Chandra, Kavitha
中科院分区：
文献类型：
--
作者：
Sharma, Prachee;Chandra, Kavitha
本文提出了一种信道采样方案，允许强大的估计信道过渡到和退出衰落状态。该采样方案是基于二阶自回归（AR-2）模型的瑞利衰落信道和相应的状态空间表示的衰落和非衰落状态。将衰落信号分割成状态的阈值参数是信噪比和误码率性能度量的函数。采样率对应于占据特定状态的过程的互信息最大化的值，该互信息以通道的两个过去的观测为条件。所提出的采样方案在基于模型的预测算法的性能。接收到的包络的状态被预测为具有98%的准确度。当应用于信道增益的估计时，结合基于卡尔曼滤波器的AR-2预测器的采样方案产生一步预测，准确地跟踪衰落信号。使用估计的信道冲激响应的多径信道的均衡示出了改进的误差性能比传统的递归最小二乘均衡器。
This paper presents a channel-sampling scheme that allows robust estimation of channel transitions into and out of the fade state. The sampling scheme is based upon a second-order autoregressive (AR-2) model for the Rayleigh fading channel and a corresponding state-space representation in terms of fade and nonfade states. The threshold parameter that segments the fading signal into states is derived as a function of signal-to-noise ratio and a bit-error-rate performance metric. The sampling rates correspond to the values at which the mutual information of the process occupying a particular state, conditioned on two past observations of the channel, is maximized. The performance of the proposed sampling scheme in a model-based prediction algorithm is presented. The state of the received envelope is predicted with 98 % accuracy. When applied to the estimation of channel gain, the sampling scheme in conjunction with a Kalman-filter-based AR-2 predictor yields one-step forecasts that accurately track the fading signal. Equalization of multipath channels using the estimated channel impulse response shows improved error performance over traditional recursive least squares equalizers.