边缘滤波在GPS多径信道估计中的应用研究

Research on Marginalized filter for GPS multipath channel estimation

在全球定位系统中,多径传播会导致定位估计精度的下降。在某些应用环境下,为了监视定位的完整性,探测多径信号的发生同样重要。文章把GPS多径信道建模近似为一阶马尔科夫模型,提出了一种边缘化的滤波方法,来联合估计由路径状态、路径幅度和路径时间延构成的连续/离散状态空间模型。这一方法使用卡尔曼滤波来解析的估计路径幅度,使用格滤波估计路径状态,最后使用粒子滤波来估计路径时间延迟。另外,使用了数据压缩技术和内插技术来降低计算的复杂度。计算仿真说明了所提出方法的性能。

In the global positioning system, multipath signal propagation can degrade seriously the position estimation. In some environments, in order to monitor the integrity of positioning ,detecting the occurrence of multipath signal is also significant.In this paper, we approximate GPS multipath channel as a first order Markov model.Then we propose to use marginalized filtering to perform joint estimation of the mixed discrete/continuous state space formed by the multipath status, amplitude, and delay. This approach use a Kalman filter to estimate analytically the amplitude of each path,and use a grid-based filter to estimate path activity parameters and finally use a paticle filter to estimate the delay of each path. A computationally efficient procedure for solving the filtering problem is considered using data compression methods and interpolation technique . Simulation results show the effectiveness of the proposed method.