基于粒子滤波的混沌系统参数估计和滤波方法

摘要

混沌系统的参数估计是混沌系统控制和同步的前提.鉴于混沌系统具有初值敏感性、不能长期预测等特点,提出了一种基于粒子滤波(PF)的混沌系统参数估计和滤波方法,并将其用于Lorenz混沌系统的参数估计和滤波,在叠加噪声情况下对混沌系统进行仿真分析.结果表明,文中提出的滤波方法在估计偏差方面优于基于扩展卡尔曼滤波(EKF)的混沌系统参数估计和滤波方法,对混沌系统的参数估计和滤波是一种有效的方法.%The parameter estimation of chaotic system is a premise of system control and synchronization. In view of chaotic system's characteristics, such as sensitivity to initial condition, long-term unpredictability and so on, a filter applying to chaotic system was proposed based on chaotic system state space theory and particle filter (PF) theory. In a superimposed noise conditions, the parameter estimation and filtering of Lorenz chaotic system were simulated and analyzed.The simulation results show the proposed filtering algorithm is better than a chaotic system parameter estimation and filtering method based on extended Kalman filter (EKF) in bias estimates, and is an effective method for estimating the parameters of chaotic system and filter.