基于改进截断正态概率密度模型的防发散多尺度卡尔曼滤波

Anti-divergent Multi-scale Kalman Filtering Based on Improved Cutting-off Normal Probability Density Model

借鉴对机动频率建模的思想,对机动目标的截断正态概率密度模型进行改进,提出了一种改进的截断正态概率密度模型(ICN模型),并将防发散算法与多尺度估计理论结合,用克服滤波器发散的S(k)算法对基本的多尺度估计进行机动检测和相应改进,得到单模型单传感器系统的防发散多尺度估计算法,旨在减小噪声影响,减小机动目标模型不准确带来的跟踪误差,改善跟踪效果。

An improved cutting-off normal probability density model（ ICN model） is proposed,referred to the idea of modeling of maneuvering frequency,and the cutting-off normal probability density model of maneuvering target been improved.On the basis of this,the anti-divergent multi-scale estimation algorithm of single model and sensor system is received in order to reduce the influence of noise and the tracking error of maneuvering target model inaccuracy and improve the tracking performance,the anti-divergent algorithm and the multi-scale estimation theory been combined and S（ k） algorithm for overcoming the filter＇ divergence been used to make maneuvering detection and the corresponding improvement of the basic multiscale estimation.