利用卡尔曼滤波调整神经网络的参数，不仅可以减少网络的学习周期，而且可以优化网络的结构。
It can not only reduce the network learning cycle, but also optimize the network structure by using Kalman filter to adjust of the parameters of the neural network.
仿真结果表明变周期距离参数化扩展卡尔曼滤波算法能有效解决经典扩展卡尔曼滤波算法在纯方位角目标跟踪时可能出现的滤波发散现象，并能处理声音信号的传输时间延迟问题。
The simulation results indicate that the variable cycle RP-EKF algorithm can resolve the filtering instability of EKF for the bearing-only target tracking and the time-delay problem.
应用推荐