基于UKF滤波的测向定位算法及性能分析

基于传统的The Unscented Kalman Filter(UKF)滤波算法,提出了一种新的改良后的UKF滤波算法.该算法直接选用动态模型中的状态变量,并增加一个再抽样过程,使得计算量更小,实现更简单.将改良后的UKF滤波算法应用到单站无源定位的测向法中,得到优于The extended Kalman Filter(EKF)滤波的定位效果,大大提高了定位精度.计算机仿真实验表明,应用该改进后的UKF算法比以往EKF类算法在滤波性能上有明显的提高.但是由于测向法自身的局限性,即使应用改良后的UKF滤波,当方位角逐渐增大时,测向法的定位效果仍然会严重恶化.因此,所提出的新UKF算法适合在非线性估计等问题中广泛应用,但对于单站无源定位,则应探究更优越的定位算法来代替测向法.

Performance Analysis of Bearings -Only with Modified UKF

Based on the traditional Unscented Kalman Filter(UKF)algorithm,a new enhanced UKF algorithm is presented.The method of the enhanced UKF algorithm directly uses state variables in dynamic model,and adds a re- sampling process,thus making the realization more easier.Comparing to the Extended Kalman Filter(EKF),the lo- cation effect is better when the proposed algorithm is applied to the single passive location.The simulation experi- ments show the superiority of the enhanced UKF including the low-complexity and high accuracy.However,because of the limitation of Bearings-Only,even if using the enhanced UKF to BO,the location effect is also deteriorated when the azimuth angle increases.Therefore the modified UKF proposed in this paper is fit to deal with nonlinear problems,and a more sophisticated method for single passive location is under consideration as a future research.