一种改进的EKPF算法在固定单站无源定位中的应用

Application of An Improved EKPF Algorithm in Fixed Single Observer Passive Location

本文针对EKPF算法在固定单站无源定位目标跟踪的应用中运算量大、实时性差的问题,通过对部分粒子进行EKF采样,将EKPF算法进行改进,改进的EKPF算法不仅有效降低了运算量,同时增加了粒子的多样性,使粒子集更能体现概率密度函数的真实分布。Matlab仿真表明,与传统的EKPF算法相比,改进算法在保证滤波性能基本不变的前提下,算法运算量大幅下降。

Extended Kalman particle filter （EKPF） algorithm applied in fixed single observer passive location has disadvantages of high computation load and poor real-time performance. An improved EKPF algorithm is achieved by way of extended kalman filter （EKP） sampling to a portion of particles. The improved EKPF algorithm can not only effectively reduce the computation load, but also increase the diversity of particles, which makes the particles set be capable of showing real distribution of probability density function. Matlab simulation result shows that, comparing with traditional EKPF algorithm, computation load of the improved algorithm got reduced significantly meanwhile it can ensure filtering performance.