EKF、UKF、PF混沌同步性能分析

针对离散混沌系统同步问题,分析扩展卡尔曼滤波(extended Kalman filter,EKF)、无损卡尔曼滤波(unscented Kalman filter,UKF)和粒子滤波(particle filter,PF)这3种同步算法的同步性能。混沌系统的非线性程度及噪声状态分布会影响算法同步性能,根据非线性程度对混沌系统进行分类,在此基础上比较3种算法在高斯和非高斯噪声干扰下的同步性能。引入克拉美罗界(Cramér-Rao lower bound,CRLB)作为高斯噪声下同步误差的下限标准,并分析了3种同步算法的时间复杂度。仿真结果表明,在同步Ⅰ型高斯系统时EKF性能最优,在同步Ⅱ型高斯系统时UKF性能最优,当系统受非高斯噪声影响时,PF算法精度最高。

Chaotic synchronization performance analysis of EKF,UKF and PF

Aiming at the synchronization problem of discrete chaotic system,the synchronization performances of extended Kalman filter (EKF),unscented Kalman filter (UKF) and particle filter (PF) were analyzed. Considering that the synchronization performance is affected by the nonlinear degree of chaotic systems and the distribution of noise states,the chaotic system was classified according to the degree of nonlinearity,and synchronization performances of these algorithms were compared under the Gaussian and non-Gaussian noise interference based on the classification. The CRLB was introduced as lower limit standard of synchronization error,and the time complexity of the three synchronization algorithms was also analyzed. The simulation results demonstrate that the EKF is optimal in the synchronization of type I Gaussian system,and the UKF is optimal when synchronizing type II Gaussian system is used. When the system is interfered by non-Gaussian noise,the PF algorithm shows the highest accuracy.