In order to cancel the interference arising between the two signals transmitted over the horizontal and the vertical polarizations of the electrical field in pol-mux transmission, the coefficients of the channel matrix must be estimated. However, when the coefficients are fast time-varying, to estimate or, better, to track them becomes challenging, especially if the estimation is non-data-aided. This paper shows that a non-data-aided particle filter can be used to track the time-varying coefficients. Also, in the paper the weights and the states of the particles are used in a forward–backward scheme in order to demodulate the signal and get the probability distribution of the transmitted modulation symbols. Computer simulations are presented for validating the proposed method.

为了消除在pol-mux传输中通过电场的水平和垂直极化传输的两个信号之间产生的干扰，必须估算信道矩阵的系数。但是，当系数快速时变时，估计或更好地跟踪它们变得很困难，尤其是在估计没有数据辅助的情况下。本文表明，可以使用非数据辅助的粒子滤波器来跟踪随时间变化的系数。同样，在本文中，粒子的权重和状态用于前向-后向方案中，以便解调信号并获得传输的调制符号的概率分布。计算机仿真被提出来验证所提出的方法。