论文部分内容阅读
混沌信号在本质上属于非线性非高斯信号,它在无线传感器网络下的应用还涉及到信号量化问题,这使得混沌信号在此应用环境下的信号盲分离更为棘手.针对此问题,本文在容积卡尔曼粒子滤波的框架下提出一种解决方法.文中首先推导出观测信号的概率密度函数,在量化比特有限的情况下,采用最优量化器,获得最优的量化结果.在此基础上,使用容积卡尔曼滤波器产生粒子滤波中的重要性概率密度函数,融入最新的观测值,提高粒子对系统状态后验概率的逼近,提高信号盲分离的精度.仿真结果表明算法能够有效地分离混合混沌信号,参数估计的精度及其运算量均优于已有的无先导卡尔曼粒子滤波算法,其运行时间为无先导卡尔曼粒子滤波算法的88.77%.
In essence, chaotic signal belongs to non-linear non-Gaussian signal, and its application in wireless sensor networks also involves signal quantization, which makes the signal blind separation of chaotic signal in this application environment more difficult.In this paper, Volume Kalman particle filter framework proposed a solution.In this paper, the probability density function of the observed signal is derived first, the optimal quantization device is used to obtain the optimal quantization results in the case of limited quantization bits .On the basis of this , The volumetric Kalman filter is used to generate the importance probability density function in the particle filter, and the latest observation value is incorporated to improve the approximation of the particle’s posterior probability of the system state and improve the precision of blind signal separation. The simulation results show that the algorithm can effectively separate The hybrid chaotic signal, the accuracy of the parameter estimation and its computation are better than the existing no-pilot-based Kalman particle filter algorithm, and its running time is 88.77% without the pilot Kalman particle filter algorithm.